All notable changes to this project will be documented in this file.
The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.
The release procedure (prompt template and step-by-step instructions) lives in docs/RELEASE.md.
The entries below hold the details; what a 1.4.0 user changes:
- A rule naming
promptKey: file-body-javanamesfile-body-oop: profile v3 has nofile-body-javaany more. - Configuration values that read differently now: a
{in anexcludes,pathGloborfileNamesglob opens an alternative ({a,b}) and an unclosed one fails the run (a glob has no escape for a literal brace;?matches it);$${name}writes a literal${name}(it was left as written); a key defined twice inaiDefinitions,promptDefinitionsorfactDefinitionsfails the run instead of the last one winning. - A link to a directory below a subtree or in the output directory is no longer walked (on
Windows a junction, which 1.4.0 entered): name a linked directory whose files you want, and whose
target no subtree holds, as a subtree of its own; with
pruneOrphansthe summaries 1.4.0 wrote below it are deleted otherwise. - The plugin's
Gis srcmorph's version, no longer the project's: a reader or a script that took the header'sGfor the project version readsprojectNameand the build instead. - The first run rewrites every summary, one model call each: a summary of 1.4.0 records no
Mand noK, the project'sCis built differently, and every file prompt of profile v3 changed (the secrets rule). To keep the old summaries instead, set all fiveregenerateOnswitches tofalseand keep them there: a field is recorded only when its summary is written, so a run with the switches on later still finds it missing and rewrites then.force: truerewrites everything at once. - Java callers of the library (the CLI and the plugin need nothing):
AiMdHeader.withHashestakes(prompt, facts, model, children); thePackageIndexerand bothProjectIndexerconstructors take aRegenerateOnafterforce; theSourceFileIndexerconstructor takesfileNamesandfileInterpreters(afterfileExtensions), aRegenerateOn(afterforce) andcontentHeadChars(last);SourceFileIndexer.indexFiletakes amodelHash(last);AiMdHeaderSupport, nowAiMdRewritePolicy, takes the five switches as booleans (without arguments, all on).GenerationTimeEstimator(wasAiGenerationTimeEstimator) is linear:REQUEST_OVERHEAD_MS,PREFILL_MS_PER_TOKENandDECODE_MS_PER_TOKENreplace the four CPU coefficients, andestimateDecodeMillistakes only the output tokens.IncompleteAnswerException.Reasonhas a fourth value,ABORTED.AiGenerationUsage.addisGenerationUsage.withCall(one call's usage).GenerationProviderhastakeLoadNanos(). - Java callers of the library, the package restructure (the configuration file, the CLI's
commands and the plugin's goals and properties stay as they are; see the entry under Changed):
fix the imports of the moved classes and rename
LlamaCppJniConfigtoProviderConfig,LlamaCppJniConfigFactorytoProviderConfigFactory,AiFieldGenerationSupporttoSummaryGenerator.GenerationRequesttakes three arguments (no header);SummaryGenerator.generatetakes four (no header).AiConditionEvaluator.CONTENT_HEAD_CHARSisSrcMorphConfiguration.DEFAULT_CONTENT_HEAD_CHARS,AiFactExtractor.READS_PER_CHARandREAD_ALLOWANCEareReadBudget's, andOpenAiCompatibleAiGenerationProvider.withoutUserInfo(text)isUrlUserInfo.mask(text). A plugin subclass extendsAbstractSrcMorphMojoinstead ofAbstractAiIndexMojo. - Java callers of the library, the null-safety review (the configuration file, the CLI and the
plugin stay as they are):
RoutingRule.getPromptKey()/getAiDefinitionKey()(wasAiFieldGenerationConfig) are@Nullable(a skip rule has none); a rule that routes gives them throughrequirePromptKey()/requireAiDefinitionKey().SrcMorphConfiguration.getBaseDirectory()is@Nullableuntil set, and an engine without one throws aSrcMorphExceptioninstead of aNullPointerException.ProviderConfig.builder()takes no argument:.modelPath(path)sets the model file, andmodelPath()isnullrather than""for a model without one (a remote model); the llama.cpp JNI provider refuses such a configuration in its constructor. ThegetKey()ofModelDefinition,PromptDefinitionandFactDefinitionandFactCounter.getLabel()/getPattern()are@Nullable, and a definition without a key throws anIllegalArgumentExceptioninstead of aNullPointerException. The overview constructor ofProjectIndexertakes oneProjectOverview(provider, rule, prompt and model lookups) instead of the four values; the four-argumentSourceFileIndexer.classifytakes both lookups non-null (the two-argument one skips the window check).GenerationProviderFactory.effectiveProvidertakes a non-null run provider and returns a non-null key, andPromptTemplateLoader.load(List)is gone (load(List, Collection)remains). - A
sizecondition measures the source as srcmorph reads it: the bytes the checksum is taken over,\nline endings and no byte order mark (a UTF-16 file: the UTF-8 length of its text), not the bytes on disk. On a CRLF checkout, or for a file with a byte order mark, a file near a tier's bound can reach another rule than in 1.4.0. The run'sminFileSizeBytes/maxFileSizeBytesband stays on the bytes on disk: it filters during the walk, before a file is read. Java callers:SourceLines.normalize(byte[])is not public any more (SourceLines.checksumBytesandSourceLines.sizeare). - A
baseUrlwith user info (https://user:key@host/v1), a query or a fragment fails the check. srcmorph never sent the user info (HttpURLConnectiondoes not authenticate with it), so such a model ran in 1.4.0 against a server without a key; put the key in the variableapiKeyEnvnames and the path alone inbaseUrl. - The self-test's gpt-oss rules map-reduce a file over the window (
onOversize: mapReduceon every rule, the default kinds too), so a configuration copied from it no longer fails a dense file just below a size tier. - Every configuration file renames three keys:
fieldGenerationsisrules(undersrcMorph),packageFieldGenerationsispackageRulesandprojectFieldGenerationsisprojectRules(the loader refuses the old names as unknown keys). The beans' accessors follow (getRules,getPackageRules,getProjectRules; the CLI'sCConfigurationfields too). - Java callers of the window and the parts:
InputWindowCalculator.maxInputChars(config)takes one argument and is the budget of the whole rendered prompt (it took the template out a second time);SummaryGenerator.parts()istakeParts(), which also starts the next file's answers anew. - A configuration that sets
pluginVersionnames itgeneratorVersion(it is srcmorph's own version, the header'sG, in the CLI and the plugin alike); the run record'spluginVersionisgeneratorVersiontoo, and so are the bean's getter and setter. - Java callers of the library, the naming review (
docs/code-review-2026-11-05.mdsection 2):AiGenerationTimingsisTimedGeneration;AiPromptPreparationSupport.getBasePromptLengthisPromptPreparer.basePromptLength;RemoteModel.UNKNOWN_CONTEXT_LENGTHis-1(it was0).GgufModelInfohas no sentinels (null-safety finding S10): it is made byparsed(name, contextLength)orunreadable(failure),modelName()andfailure()arenullwhere they were"",contextLength()is anOptionalLong, andUNKNOWN_CONTEXT_LENGTHis gone. The CLI'sMain.logConfigurationTransformationislogConfigurationEcho,configurationToYAMLisconfigurationToYaml, andreadString,fromJson,fromYaml,configurationToJsonandconfigurationToYamlare package-private.AiMdHeader's accessors are named (the file keeps its letters):h()isheaderVersion(),c()checksum(),d()sourceDate(),t()generatedAt(),g()generatorVersion(),a()aiVersion(),x()nodeType(),p()promptHash(),n()factsHash(),m()modelHash()andk()childrenHash().AiIndexPlan.EntryisIndexPlan.Route, andSourceFileIndexer.recordedHashesreturns aRecordedHashes(promptHash(),factsHash()) instead of aMap.Entry.AiGenerationConfigis gone:ModelRegistry.definition(wasAiModelDefinitionSupport.getConfig) returns a copy of theModelDefinitionwith its model path resolved,ProviderConfigFactorybuilds aProviderConfigfrom it, and theDEFAULT_*constants areModelDefinition's (maxInputCharswas never configurable:DEFAULT_MAX_INPUT_CHARSis the fallback withoutcharsPerToken). TheAiprefix is gone but on theAiMd*classes of the.ai.mdformat, and the*Supportnames that hid a role say it:AiChecksumSupportChecksumSupport,AiGeneratorVersionGeneratorVersion,AiGlobGlob,AiPathSupportSourcePaths,AiProfileFileSupportProfileFileSupport,AiShebangShebang,AiSourceExcludeFilterSourceExcludeFilter,AiSourceLinesSourceLines,AiTimeSupportTimeSupport,AiCalibrationCalibration,AiConditionRuleCondition,AiConditionGroupConditionGroup,AiFactCounterFactCounter,AiFactDefinitionFactDefinition,AiFieldGenerationConfigRoutingRule,AiIncompleteStrategyIncompleteStrategy,AiModelDefinitionModelDefinition,AiModelDefinitionSupportModelRegistry,AiOversizeStrategyOversizeStrategy,AiPromptDefinitionPromptDefinition,AiRangeConditionRangeCondition,AiRegenerateOnRegenerateOn,AiFactDefinitionLoaderFactDefinitionLoader,AiFactDefinitionSupportFactGroups,AiFactExtractorFactExtractor,AiConditionEvaluatorConditionEvaluator,AiFieldGenerationSelectorRuleSelector,AiFileContextFileContext,AiGenerationRequestGenerationRequest,AiPreparedPromptPreparedPrompt,AiPromptPreparationSupportPromptPreparer,AiPromptSupportPromptCatalog,AiPromptTemplateLoaderPromptTemplateLoader,AiMdHeaderSupportAiMdRewritePolicy,AiCompletionParserCompletionParser,AiContextOverflowExceptionContextOverflowException,AiGenerationProviderGenerationProvider,AiGenerationProviderFactoryGenerationProviderFactory,AiGenerationUsageGenerationUsage,AiIncompleteAnswerExceptionIncompleteAnswerException,LlamaCppJniAiGenerationProviderLlamaCppJniGenerationProvider,MockAiGenerationProviderMockGenerationProvider,OpenAiCompatibleAiGenerationProviderOpenAiCompatibleGenerationProvider,AiCalibrationMeasurementCalibrationMeasurement,AiCalibrationRunnerCalibrationRunner,AiHeaderHashesHeaderHashes,AiIndexOutcomeIndexOutcome,AiIndexPlanIndexPlan,AiIndexTallyIndexTally,AiDeterministicSummaryDeterministicSummary,AiGenerationResultSummaryResult,AiGenerationTimeEstimatorGenerationTimeEstimator,AiInputWindowCalculatorInputWindowCalculator,AiMapReduceSupportMapReduceMath,AiOverflowBackoffOverflowBackoff,AiRouteEstimatorRouteEstimator,AiSourceChunkerSourceChunker,AiTokenWindowTokenWindow,AiProgressBarProgressBar.ModelRegistry.getConfigisdefinition,AiMdRewritePolicy.shouldWriteisshouldRewrite,ProviderConfigFactory.fromGenerationConfigisfrom. - Module-path consumers of the library see
configandengineonly:module-info.javaexports those two (andsupportto the CLI alone), no longerdocument,facts,indexer,indexer.generation,prompt,providerandrouting. The engines' results name none of their types:CalibrationReportis read through its three renderings (its constructor,measurements()andModelMeasurementare package-private),ModelListReportthroughrender()(its constructor,listings()andListinglikewise), andRunRecord, which no engine returns, is package-private;CoreArchitectureTest#exportedApiNamesNoInternalTypekeeps it that way. On the classpath (the CLI, the plugin, a Java 8 consumer) nothing changes but these visibilities. - A script that reads srcmorph's log lines: the file phase names a file by its path below the base directory (it was the absolute path), and the plan's model header and its per-file estimate line have a new form (the entries under Changed).
- The plan, the run record and the progress list files of several subtrees in one path order
(
src/extra/...beforesrc/main/...), no longer subtree by subtree in configured order. - Plan times change: lower for map-reduce files whose rounds end in a single partial, lower for
multi-byte (CJK) sources, and 0 s for a chunked file whose window leaves less than 1000 chars for a
chunk. Such a file fails the plan and
check(the "Over window" section saysFAILS BUILD: N chars for a chunk), as anonOversize=failfile does; the run used to fail it file by file. On a model whose provider counts the prompt's tokens (llamacpp-jni) the room is only estimated before the run, which sizes the window by the count, so there the plan warns (handled; WARNING: N chars for a chunk by the estimate) and the run decides. Java callers:SourceFileIndexer.classifywith the window check takes the run'sgenerationProvideras a fifth argument. A context overflow of a prompt of 1000 chars or fewer is not retried; the retry warning's "(A -> B chars)" names the refused prompt's length as A. - Java callers of the generation classes:
RouteEstimator.estimateRouteSeconds(...)has nofileSizeBytesparameter,MapReduceMath.sourceBudgetis gone;SourceFileIndexer.sortByPathandSourceLines.cutIndexare new. .jspxand.tagxtakefactsKey: jspx-facts(jsp-facts with XML comments blanked); a copied configuration splits its jsp rule the same way. A configuration that copied thesecretsrule takes its new form (the self-test's), which keeps CI workflows, Helm templates and Android settings.- A
baseUrlthat holds an@anywhere fails the check (user info before the host, a password with a/, an@in the path), and a 3xx answer is an error: srcmorph follows no redirect, sobaseUrlnames the final server. reasoningEffortdefaults to empty, no longerlow(breaking): a model definition withoutreasoningEffortsends no effort, neither areasoning_effortargument to a local chat template nor anything to a remote server, so the model's own default applies (gpt-oss:medium). A gpt-oss definition that relied on the old default namesreasoningEffort: low, as the self-test, its CLI twin and the Luna example do;models.jsonsets it incommon, so the model card's definitions run as measured.- An
openai-compatiblerequest carriestop_kand the repetition penalty only when the definition sets them (breaking):topKother than40(0is sent now and switches top-k off),repeatPenaltyother than1.0. Neither is an OpenAI chat-completions field; llama-server applies the same defaults, and OpenRouter now applies the model's own. - A source named
packageorprojectfails the plan (breaking): it was summarized into thepackage.ai.mdof its directory (project.ai.mdof the output root), the file the package (project) phase writes; the name is compared in any case on every file system (Packagetoo), since a case-insensitive one would put both into one file. Exclude it or take it with a skip rule. A source namedprojectbelow the base root is summarized intoproject.ai.mdof its directory, and the package phase now reads that file as its directory's content (it left out everyproject.ai.md, not only the output root's). - Summaries of thirteen kinds are written again: the fact counters of Java, OpenAPI, AsyncAPI,
batch, SQL, GitHub Actions, Ansible, Kotlin, Scala, Groovy and JSP changed (see Fixed), and with
them the facts hash
Nof their summaries; Java EE (java-ee-facts) and JSP documents (jspx-facts) include the Java and JSP counters, so theirs changed too. - A large file of an XML kind stays with its kind's rule: the self-test's
big-xmlandhuge-xmltiers no longer take apersistence.xml, MyBatis mapper, WSDL, Facelets page or the other XML kinds (see Changed); a copied configuration takes the tiers' new conditions. - A copied secrets rule takes the new properties pattern of the self-test (see Changed): the
old one let a value that only starts with
${through and skipped a flag or a log level. - A benchmark script that runs
run.pyorscore.pyon an output tree of its own passes--meta <dir>(and score.py--scores <dir>): without them both now refuse an--outother than the card's. run.py's--chars-per-tokentakes a whole number of at least 1. - A fraction in a whole-number key fails the run (breaking):
charsPerToken: 2.5,contextSize: 4096.0and the like were cut to an integer without a word; write the whole number.
-
New key
srcMorph.pruneOrphans: the plan lists the orphaned summaries, and the switch deletes them. Nothing deleted the summary of a deleted, renamed or newly excluded source (or of one a skip rule now takes), and the package and project phases kept reading it. Thegenerateplan now lists every file summary (X: file, read from the header block) below the output directory that no file of the plan is summarized into, and thepackage.ai.md(X: package) of a directory left without any summary and sub-package (a deleted or renamed package, which the package phase neither rewrites nor removes), in a section "Orphaned summaries (N)" with a warning.pruneOrphans: true(defaultfalse) deletes them, each with a log line, before a run that generates; aplanOnlyrun and a check delete nothing. Kept, with a warning: a summary whose header cannot be read, what the walk of the output directory cannot read, and every summary below a source file or directory the source walk could not read. Leave it off while narrowingsubtreesfor one run: the summaries of the subtrees left out are orphans of that run too. A summary that stays or that the run writes keeps thepackage.ai.mdof every directory above it; a missing default subtree (src/main/java) and a script whose#!line cannot be read keep their summaries; any other plan that routes fewer files than before (an exclude, a skip rule, a narrowedfileExtensions) makes the summaries of the rest orphans. On a case-insensitive file system that compares paths by case (macOS), the summary of a source renamed only in case is an orphan once and is written again. -
partsDirectorykeeps the map-reduce answers of each file: every map answer and every reduce batch's answer of a file read in chunks, in<partsDirectory>/<path below the base>.parts.md, so a measurement can see where a summary loses what a part said (TODO row 32). Unset, nothing is kept. -
A rule for
package.json:file-body-opswithnpm-facts. The counters count the entries ofscripts, the four dependency objects andengines, the strings ofworkspaces, and a VS Code extension's contributed commands and language model tools;file-body-opshas a line for an npm manifest (scripts, dependencies, engines, workspaces, an extension's commands, chat participants and tools), so every ops file's prompt hash changes once.run.pyroutespackage.jsonthere. Measured on the manifest of a VS Code extension (internal, three runs each): of its 45 names gpt-oss-20b named 45, 28 and 34 (with the fallback prompt 45, 13, 28 and 15; with the rule but the ops prompt as it was 21, 21 and 28, without the extension's 17 tools), GPT-6 Luna 45 every time (43 to 45). -
The fact counters are measured linear on Java 8 as well: the
linearity-jdk8job ofpublish.ymlrunsFactPatternLinearity, the measure ofProfileFactPatternsLinearityTestwritten in Java 8, on a Java 8 JDK, since the counters' time limit counts the reads of a regex engine that differs between JDKs. It gates both publish jobs. -
$${name}writes a literal${name}in a configuration value, whether a property of that name exists or not: an inline prompt can show a shell variable ($${HOME}) that-DHOME=...or a Maven property would otherwise replace. In a path it still fails the load, as an unreplaced reference does. -
TypeScript takes
file-body-oop: a rule for.ts,.mtsand.ctsinrun.pyand in the Luna example (no counters of its own). Measured on 2026-11-05 on a TypeScript repository, internally: with the oop prompt both models named more of the exported API than with the fallback, gpt-oss-20b clearly more, also of the private helpers and the tests. -
examples/copilot-api-gateway/srcmorph-luna.yaml: GPT-6 Luna through the GitHub Copilot API Gateway for every file, package and the project, with the whole profile v3: a rule per kind of source with its prompt and exact fact counters, andonOversize: mapReduceon every rule, the package and project rules too: an input larger than the window is read chunk by chunk, never sampled. It takes every extension of the self-test profile and leaves outtarget/,.git/,.idea/,.vscode/,.gradle/,node_modules/and.claude/(agent worktrees, a second copy of the sources).ExamplesConfigBindingTestpins the routing, the facts, the strategy and the chunk prompts. It routes every kind profile v3 knows (below): Java with Jakarta EE imports or annotations written with their package (contentMatches) to the EE counters, Kotlin, Groovy and Scala to the shared OOP prompt, the EE descriptors and mappings by name (fileNames), MyBatis mappers and integration routes by their DTD or namespace, BPMN, DMN, WSDL, JSP, Facelets, OpenAPI and AsyncAPI by their first key, Protobuf, Avro and GraphQL, GitHub Actions by path, Kubernetes bykind:, Compose files by name, Ansible playbooks and roles, Terraform, Camel's YAML DSL by its- route:/- from:, Dockerfiles and Jenkinsfiles by name, Gradle build scripts of both DSLs and.propertiesfiles to the ops prompt, PowerShell and batch files to the fallback prompt with their counters, and scripts without an extension by their#!line (fileInterpreters).ExamplesConfigBindingTestroutes one file of each. -
Benchmark: the enterprise corpus (
docs/ai-index-benchmark/corpus-enterprise/), seventeen files of those kinds around one shop backoffice, withmodelcard/facts-enterprise.json.run.pyroutes the kinds its corpora hold as the Luna example does (not AsyncAPI, Ansible, Camel YAML, PowerShell, batch or scripts by their#!line, which none of them has) and takes--java-prompt, so the Java prompt and the shared OOP prompt run side by side on the card corpus;score.pyreportssecretsLeaked, the summaries that copy a value the facts mark as a secret. -
ProfileFactPatternsLinearityTestmeasures that every pattern of every facts file of the profile reads linearly, the[ignore]lines included, over 121 adversarial lines, each also repeated within one line (unclosed openers and tags, heredoc markers, unbalanced parentheses, runs of one character). It counts the characters the matcher reads instead of timing it: four times the input may read at most six times as many (linear reads four times, quadratic sixteen), so the result is the same on any machine. A facts file added to the profile is measured without a change to the test. Against the facts files of 1.4.0 it reports six quadratic patterns, all fixed above. -
Files without an extension can be indexed and routed (
fileNames, undersrcMorphand as a condition leaf): globs over the file name without its directory, matched in any case (Dockerfile,Jenkinsfile,Dockerfile.*); a name with a/fails the run in both places (it could never match: usepathGlob). Up to now onlyfileExtensionsselected a file, so a Dockerfile or a Jenkinsfile could not be indexed at all. -
Routing by content (
contentMatches, a condition leaf): a Java regular expression found in the firstcontentHeadCharscharacters of the file (undersrcMorph, 16,384 by default, at least 256 so theinterpreterscondition sees a whole#!line; raise it for a marker behind a long license header: the head is never read past the file's end), so a Kubernetes manifest, an OpenAPI document and a Camel route, all.yaml, take different rules ((?m)^apiVersion:). Planning reads each candidate once, whole; an invalid expression fails the check. The pattern is compiled once and has the read budget of a fact pattern: one that backtracks past it does not match, with a warning, so it cannot hang the plan. -
Scripts without an extension can be indexed and routed by their
#!line (fileInterpretersundersrcMorph,interpretersas a condition leaf):[bash, sh, perl]selects abin/deploythat starts with#!/usr/bin/env bash. The interpreter is the program the line names, or whatenvruns (options and assignments skipped); a name matches its versions too (pythonmatchespython3.12; a name ending with a digit only from a dot on:python3.1does not matchpython3.12), and the arguments ofenv's options (GNU-u,-C,-a, BSD-P,-L,-U) and its-Scommand line are read asenvreads them. The line is read only for a file the subtrees, excludes and size band let through; one that cannot be read is skipped with a warning, the run goes on.support.Shebangholds the rule; an interpreter given with its directory (/bin/bash) fails the run. -
Fact counters can count before the blanking and count names (
[raw]and[distinct]in a facts file,rawanddistinctin a configuration): a raw counter counts the source as written, so aTODOin a comment counts next to counters that skip comments; a distinct counter counts the different texts it matched, its first group where that took part (twelve@Inject Clockare one type). A summary written before keeps its facts hash; the markers enter it only where set. -
Every fact pattern has a time limit: it may read 2,000 characters per character of the source plus 1,000,000, counted by the matcher's reads, so the limit is the same on every machine. A pattern of a user's own that backtracks past it is stopped: its counter is left out of the facts line, an ignore counter blanks nothing, and a warning names it. The profile's patterns read far less (
profiles/README.mdgives the measured maximum); the linearity test asserts that none reaches the limit. -
Fact counters for the kinds that had a prompt and none:
scala-facts,proto-facts,avro-facts,graphql-facts,asyncapi-facts,terraform-facts(OpenTofu and Terragrunt too),ansible-facts,compose-factsandcamel-yaml-facts;dmn-factsfor DMN decisions (inputs, outputs, rules, hit policies other than UNIQUE, the decision requirements);gradle-factsandproperties-facts, andpowershell-factsandbatch-factsfor the fallback prompt.ContractFactPatternsTest,OpsFactPatternsTest,JvmFactPatternsTestand (DMN)EnterpriseXmlFactPatternsTestpin each on a sample of its kind with a commented-out or quoted declaration that must not count. The YAML ones blank the body of a block scalar from the indentation of its first line, as YAML reads it, and keep its key (ansible.builtin.shell: |counts). -
{a,b}alternatives in every glob (excludes,pathGlob,fileNames), nested too:**/*.{yml,yaml},**/{target,build}/**. The syntax lives once insupport.Glob; a glob that opens a{it does not close fails the run, naming it. -
Profile v3 covers the kinds of an enterprise Java code base, with four prompts and fifteen fact files, so one generic prompt serves every language of a kind:
file-body-oop: one prompt for the class-based languages (Java, Kotlin, Groovy, Scala); it replacesfile-body-java(below), with their facts per language:kotlin-facts(types, companion objects, data and sealed types, functions, extension and suspend functions, constructor properties, nested comments and string templates blanked),groovy-facts(types, typed anddefmethods, Spock features and blocks, closures, GStrings and slashy strings blanked) andjava-ee-facts(java-factsplus the EJB, JPA, CDI, JAX-RS, JAX-WS, Servlet, JTA and security annotations).file-body-xmldefines Jakarta EE descriptors, JPA and Hibernate mappings, MyBatis mappers, Camel, Spring Integration, Mule and Synapse routes, BPMN and DMN, and WSDL, each counted by its own file:jee-facts,persistence-facts,mybatis-facts(which reads the SQL in CDATA),integration-facts,bpmn-facts(the diagram blanked) andwsdl-facts.file-body-viewfor JSP, tag files and Facelets (jsp-facts,facelets-facts): the data a page reads, the actions it sends, its template and its server code.file-body-contractfor OpenAPI, AsyncAPI, JSON Schema, Avro, Protobuf and GraphQL (openapi-facts, YAML or JSON, block scalars blanked) andfile-body-opsfor CI pipelines, Dockerfiles, Compose, Kubernetes, Helm, Ansible and Terraform (k8s-facts,actions-facts,dockerfile-facts).
JvmFactPatternsTest,EnterpriseXmlFactPatternsTest,ViewFactPatternsTestandOpsFactPatternsTestpin every counter on a sample of its kind with its traps; the linearity test measures every pattern. -
A facts file can include another (
[include]<TAB><key>): its counters,[ignore]lines included, at the place of the line, sojee-factsisxml-factsplus the descriptor counters and copies none. A cycle fails the run naming the chain; a label two of the files use fails as before. -
Secrets stay out of the index. Every file prompt names a password, private key, token, API key or credentials in a URL without copying it (
DB_PASSWORD=***): the shell, XML and fallback prompts asked for "every value", so a<password>of a data source or a token in a script went into the summary, and the default output directorysrc/site/aiis published with the Maven site. The self-test,examples/build-your-own-jar/srcmorph.yamland the Luna example skip the files that hold credentials (id: secrets:.env,*.tfvars,*.tfstate, Maven'ssettings.xml, keys,*secret*and*credential*configuration files, and a.propertiesfile whose head gives a password, secret, token or API key a value that is not a${...}reference, such as agradle.propertieswithsigning.passwordor anapplication.propertieswithspring.datasource.password), so they never reach a model;ExamplesConfigBindingTestpins which names it skips and that no source file is among them. Of the 20 properties files of the 2026-11-07 runs two are skipped that way (keystore passwords of a TomEE WS-Security example). -
The CLI runs on the module path in CI: the
smoke-jarjob plans the demo tree withjava -p ... -m net.ladenthin.srcmorph.cli/..., the only run that reads themodule-info.javafiles (what the core exports, what the CLI opens to Jackson). -
modelcard/correlate.pycomputes the model card's correlation of an index's size with its search scores fromscores/(Pearson and Spearman, the card's 32 runs by default).
-
logback-classic 1.6.7 (test scope),
actions/setup-java6.3.0 andanthropics/claude-code-actionv1.0.276, in step with the sibling repositories. -
- Dependabot: one grouped pull request per ecosystem a week, on Saturday.
.github/dependabot.ymlis the template of bernardladenthin/actions v1.2.1: every minor and patch update of Maven, and of the GitHub Actions, comes in one pull request (a major update stays a pull request of its own), on Saturday, so a week of updates is one merge and one run of the pipeline onmain.common/check-dependabotinrepository-checksholds the file to that rule, and every action and reusable workflow of bernardladenthin/actions is pinned at v1.2.1.
- Dependabot: one grouped pull request per ecosystem a week, on Saturday.
-
bernardladenthin/actions v1.2.2. The snapshot publication no longer deletes the
snapshottag when GitHub refuses the workflow token the move (a 403 after a change to a workflow file): the tag stays where it is, the run warns with the command that moves it, and the pre-release is updated all the same. Before, the fallback deleted the tag, could not create it again, and left the pre-release a draft (java-llama.cpp, Publish run 42708115919). -
CI: every artifact upload and download step times out, after the minutes of the repository variable
ARTIFACT_TIMEOUT_MINUTES, else 20 (7 steps inpublish.yml). In a Publish run of java-llama.cpp anupload-artifactstep hung for hours after a build of a minute and a half, with GitHub's 360-minute default the only limit; the longest transfer measured over two complete runs of that pipeline took 18 s. The workflow's top-levelenvreads the variable with the default, each step saystimeout-minutes: ${{ fromJSON(env.ARTIFACT_TIMEOUT_MINUTES) }}, andcommon/check-artifact-timeoutsof bernardladenthin/actions v1.2.0 holds that layout inrepository-checks. -
CI: a pull request Dependabot opens or rebases does not run the Publish pipeline. Its
startgatejob, which every other job needs, is skipped for it (github.actor), so every job of the run is skipped. Such a run got none of the repository's secrets and Dependabot rebases its pull requests on every change ofmain. The merge tomainruns the pipeline; a push of a person to the branch, or a dispatch on it, runs it too. -
bernardladenthin/actions v1.2.0 (from v1.1.1) for every action and reusable workflow it pins.
-
exec-maven-plugin 3.7.0 (
srcmorph-maven-plugin) andanthropics/claude-code-actionv1.0.271, in step with the sibling repositories. exec-maven-plugin 3.7.0 needs Maven 3.9.0, so maven-enforcer requires Maven 3.9.0 instead of 3.6.3 to build the reactor. Using the plugin still needs Maven 3.6.3 (its descriptor is unchanged). -
Dependencies, build plugins and Actions, in step with the sibling repositories: Jackson 2.23.0, maven-enforcer-plugin 3.7.0 in
examples/build-your-own-jar(the reactor had it already),actions/setup-java6.2.0 in thelinearity-jdk8job andanthropics/claude-code-actionv1.0.269.bernardladenthin/actionsis pinned at v1.1.1: its actions and reusable workflows runactions/setup-java6.2.0,github/codeql-action4.39.3,gradle/actions/setup-gradle6.4.1 andsoftprops/action-gh-release3.0.4, andverify-signing-key-gradlesigns with Gradle 9.9.0 by default. Error Prone 2.52.0 (#273, already onmain) reports theStringBufferof the${name}substitution inSrcMorphConfigurationFileLoaderasJdkObsolete, which-Werrorturned into a failed compile; the substitution copies into aStringBuilderby hand now, because theStringBuilderoverload ofMatcher.appendReplacementis Java 9+ and the core is release 8. The output is unchanged. -
CI: the live test's llama-server is b12211 (was b11946), the newest
ghcr.io/ggml-org/llama.cpp:server-*image on 2026-11-09, pinned by digest.OpenAiCompatibleLiveTestpasses against it locally (the running-server and the jllama-server cases). -
The configuration loader refuses a fraction in a whole-number key (breaking), naming the key: Jackson cut
charsPerToken: 2.5to2and0.5to0, which switches srcmorph to the staticmaxInputChars; run.py's--chars-per-token 0.5reached srcmorph that way. A floating-point key (temperature: 1) still takes a whole number. -
The XML size tiers leave a kind's file to its rule:
huge-xmlandbig-xmltake only what thexmlrule takes and leave out every file an XML kind rule takes by name or content (jee, mybatis, bpmn, persistence, integration; their whole conditions copied in rule order, a test holds them equal, so a large.xsdor.xslthat only names a kind's namespace stays with the tiers), and.wsdland.xhtmlby extension. From 260 KB apersistence.xml, MyBatis mapper or WSDL got the genericxml-factson Granite; it now keeps its kind's counters on gpt-oss, which map-reduces it. POMs keep their own tiers. -
Secrets skip a properties file with a secret key, another spelling or a blank separator: the properties branch of the secrets rule (self-test, CLI twin, Luna example, run.py) also takes
aws.secretKey=,app.secret-key=,aws.secret-access-key=,jwt.secret.key=,db.pwd=,keystore.pass=,signing.passphrase=,gpgPassphrase=,jwt.private-key=,credentials=andmail.smtp.password hunter2(a blank is a separator in a properties file), and a value that only starts with a placeholder (${A}hunter2) or gives a default (${DB_PASSWORD:hunter2}). A whole placeholder (password = ${DB_PASSWORD},"${X}", Maven's@db.password@, Spring's#{...}), a flag, a number or a log level (token=false,secret=3600,logging.level.org.secret=DEBUG), an empty value and a key that only starts so (password.label,token.expiry,passwordless) stay sources; a message bundle that labels a password field (login.password=Password) is skipped, as before. Not seen: a plural or a qualifier after the key (passwords=,password.prod=), credentials in a URL, anaccount-keyand a value on a continuation line. run.py readscontentMatchesin ASCII, as Java does. -
score.py wants
--metaand--scoreswith an--outother than the card's (breaking for a script that left them out): it applied the cut-off lists of the card's runs to any output tree and wrote over the card'sscores/, which it now refuses for another--outtoo. The card's quality metric moved from plot.py toscore.card_metrics, with a test, and so did the cost per MB (score.cost_per_mb): a run whose cost lookup failed (costUsdnull) is left out of the cost charts, where plot.py divided it and stopped. run.py likewise wants--metawith another--out: it wrote that run's records over the card'sruns/. -
OpenAPI, Camel YAML and Dockerfile rules take no lookalike: the
openapirule wants a version afteropenapi:orswagger:(a Springapplication.ymlwithswagger: enabled: truegoes toops), a Camel route's- from:names an endpoint (timer:,kafka:, nothttp:/https:, which no Camel consumer is) or holdsuri:,steps:orparameters:, after anid:, adescription:, a comment or a blank line too (a list of redirects- from: /docs/oldis no route), and the configs and run.py exclude**/*.dockerignore, BuildKit'sDockerfile.dockerignorethatDockerfile.*took. -
Secrets skip a Kubernetes Secret by its kind: the secrets rule of the self-test, its CLI twin, the Luna example and run.py skips a YAML file whose head declares
kind: Secret, whatever its name;deploy/postgres.yamlwent to thek8srule, in the Luna example to a remote model. A Helm chart's template stays out, as in the name branch. -
The
openai-compatiblerequest body is shaped for OpenRouter and llama-server: a configured penalty goes out asrepetition_penaltyandrepeat_penalty, a configured effort asreasoning.effortandreasoning_effort(OpenAI's and llama-server's name), so llama-server sees both; OpenRouter answers a request with both forms (checked 2026-11-07). -
The self-test's large tiers start at ~260 KB, where gpt-oss-20b's window really ends. They started at 275,000 bytes, but the 96K window leaves about 270,000 characters next to the XML prompt, so a 272 KB Java, SQL, XML or POM file failed the run's plan (
onOversize: fail). And every other kind the self-test routes (shell, Perl, TypeScript, JSP, BPMN, DMN, Protocol Buffers, GitHub Actions, Kubernetes, Dockerfiles, Terraform, npm,opsand the fallback) map-reduces a source over the window now instead of failing the run: one 400 KB OpenAPI file did. The CLI'sbuild-your-own-jar/srcmorph.yamlthe same; its header says what its one subtree indexes and how to widen it. -
Counters closer to the source (the facts hash of these kinds changes, so their summaries are written again): the integration counter's property placeholders no longer count Camel Simple expressions (
${body},${header.id}), Facelets input fields no longer count<f:selectItem(s)>, Dockerfiles blank BuildKit heredoc bodies, PowerShell counts a quoted dot-source and noMandatory = $false, Gradle reads the Kotlin DSL's bare plugins andtasks.<name> { }, Kotlin no longer counts awhen (val x = ...)subject as a constructor property, Java counts an annotation element with adefaultas a method, GitHub Actions no longer counts a reusable-workflow call as an action, and Compose no longer counts the$${X}escape as a variable. -
Every fact counter reads linearly within one line too, and more of them count what their label says (the facts hash of these kinds changes, so their summaries are written again). The linearity test also repeats each adversarial fragment on one long line (
\',#{,${,rpc a(b, ...), which found string literals full of escaped quotes, nested placeholders and Java declarations under thousands of annotations reading quadratically or overflowing the stack; a literal not closed on its line is blanked to the line end now (java, kotlin, groovy, gradle, scala, proto, terraform, graphql), an unclosed SQL'to the end of the file, a PowerShell string across lines, and the rest of a shell or Perl heredoc-opener line that cannot be read to its end. Counters: integration property placeholders count Mule's and Spring's${env},${file.dir}and${type.handler}again (only Camel's Simple forms are left out); a Dockerfile heredoc opens only where<<starts a word on a RUN, COPY or ADD line (also a continued one); Java methods and constructors count a wrappedthrowsand a brace on the next line, and@java.lang.Override; Kotlin extension functions take a nested generic receiver; properties count a secret key that starts with the keyword (password=); Gradle's included builds areincludeat a line start, its dependencies count after{or;; GitHub Actions count every secret in an expression; Compose counts$NAME; Ansible counts a quotedbecome; batch blanks& rem; proto's streaming rpcs read CRLF files; MyBatis and properties placeholders count the innermost of nested ones. SQL's'\'(T-SQL, PostgreSQL, Oracle) andESCAPE '\'close their literal instead of blanking the rest of the file (a backslash escapes a quote only where no delimiter follows, as MySQL's'it\'s'); a Dockerfile<<inside quotes opens no heredoc, and a second heredoc on one line is blanked too; Compose counts the substitutions in a block scalar (command: >), which it interpolates, and none in a comment. -
The model card's scorer reads what the model answered.
score.pystrips a leading reasoning block before recall, invented names, numbers and the format checks (the leak is still counted), matches a path- or colon-like name whole (/ordersis not found in/orders/{orderId}; a Perl arrow,JSON::PP->new, still ends a name), opens a code fence only where CommonMark does (at a line start, or indented to its list item's content), reads the methods of a static interface and annotation elements, takes no digits stuck to a letter or a date part as a stated number, splits a name after an acronym (XMLParser), and takesorder.coritem.tagfor a member, not a file. A complete answer after a closed reasoning block is a leak, not a leak-only failure; one whose block never closes is.run.pymaps a package call's log name to itspackage.ai.mdagain, so a cut package counts as cut, and runs the project phase after some packages failed;search.pyreads the answer the same way and keeps a leading dot of a path (.github/...). The card's numbers do not move (its outputs have none of these cases). Two published numbers elsewhere differ from a re-score: the generic-prompt languages run 3 gives gpt-oss-20b 94.9 % instead of 95.3 % (the mean 95.5 % instead of 95.7 %), becauseinvoice-service.jarno longer credits the nameinvoice-service; and the large-corpus run 2 of 2026-11-01 gives gpt-oss-20b 49.1 % instead of 55.2 % (a stricter format check of an earlier round), so the gpt-oss mean over the three large runs re-scores to 76.0 %, published 78.1 % (run 3 keeps its 88.3 %). The writeups keep the numbers they were published with.scores/search.jsonis re-run (2026-11-07, 1.46 USD) on the text without the facts and coverage lines thatsearch.pyreads now: BM25's hit@1 stays in every run (MRR up to 0.016), the reader's hit@1 moves in 27 of 37 runs by up to 12 points (mean 93.6 % over the 35 runs; the two baselines, whose input is the same, keep their hit@1 and move by up to 7 points at hit@3, so part of the change is the reader's variance), and the correlation of an index's size with it is +0.13 with hit@1, +0.15 with the MRR. The search runs of the repeat, prompt-v2, prompt-v3, language and the two re-measurement directories are re-run the same way (1.62 USD; BM25 again unchanged). The model card takes the new numbers in its Search section, its recommendations and findings 5 to 7: the reader finds 84 to 100 % first with the 32 scored runs (it said 86 to 98 %), Qwen3.8-27B with reasoning off is no longer the one exception that searches worst in every run (86 to 98 % over its three), and an index's size does not go with the reader's result. The writeups of 2026-11-01 and 2026-11-02 and the deck (slides 5 and 13) take them as well, the writeups naming the published values. -
The model card's search prose and caveats match its data: BM25 over the summaries finds 67 to 91 % first (it said 67 to 86 %), 81 % on average against 84 % over the source, and nine of the 35 runs beat the source; the index length goes with the reader's hit@1 at +0.13 over the 32 scored runs (it said −0.46, from an older set of runs). The caveats add that recall does not measure faithfulness (a bare name list scores 100 %), that three runs against three reach p = 0.1 at best in Fisher's test and that search was compared unpaired (McNemar per question was not run), and that the prompts were tuned on the corpora they are scored on. No table or number of
report.pymoved. -
The self-test routes the new kinds measured on real projects: TypeScript, the Jakarta EE descriptors, MyBatis mappers, BPMN, DMN, JSP, Protocol Buffers, GitHub Actions, Kubernetes manifests, Dockerfiles, Terraform, npm manifests and other YAML (
ops), with the Luna example's rules on the self-test's model, and the extensions and file names they need. Measured on jpetstore-6, microservices-demo and camunda-bpm-examples (218 files of 13 kinds, 1072 names derived from the sources, two runs each): with the kinds' rules gpt-oss-20b names 97.3 % of the names, with every file through the fallback prompt 81.2 %; GPT-6 Luna 97.7 % against 92.2 %. No kind is worse with its rule for gpt-oss-20b; for Luna the Spring XML files through thexmlrule are (90 % against 100 %). The remaining kinds followed on 2026-11-07, measured the same way on 16 open-source projects and 12 Python projects'pyproject.toml(272 files of 20 kinds, 2046 names): persistence and integration XML, WSDL and Facelets (which leave thexmlrule), OpenAPI, AsyncAPI, Avro, GraphQL, Camel YAML, Compose, Gradle, properties, PowerShell, batch, Kotlin, Groovy, Scala andpyproject.toml(the ops prompt: gpt-oss-20b 98 % against 88 %, Luna 97 % against 86 %) take the Luna example's rules; Ansible takesfile-body-fallbackwithout counters, where gpt-oss-20b names 95 % (86 % with the ops prompt and its counters); the Java EE classes stay on thejavarule (file-body-oopeither way). Over all kinds gpt-oss-20b names 88.9 % with the rules against 79.7 % through the fallback, Luna 96.6 % against 88.6 %. The Luna example routespackage.jsonandpyproject.tomltoo. -
The self-test's
big-pomrule takesfile-body-xmlinstead of the generic prompt: on a real 590 KB POM Granite H-Tiny invents no names with it and about half of them with the generic one, at the same low recall (6 and 4 of 99;profiles/README.md).huge-pomis unchanged. -
The core library's packages say what their classes do (the review of 2026-11-05,
docs/code-review-2026-11-05.md, moves M1 to M10 and renames N1 to N3).configholds only the beans and the file loader: the routing (AiConditionEvaluator,AiFieldGenerationSelector,AiFileContext) is the new packagerouting, the fact counters (AiFactExtractor,AiFactDefinitionLoader,AiFactDefinitionSupport) the new packagefacts, andAiPromptDefinition, a bean like the others, moved frompromptintoconfig. The generation of one summary body is the new sub-packageindexer.generation:SummaryGenerator(wasAiFieldGenerationSupport) withAiMapReduceSupport,AiTokenWindow,AiOverflowBackoff,AiInputWindowCalculator,AiRouteEstimator,AiGenerationResult(fromdocument) andAiSourceChunker,AiDeterministicSummary,AiGenerationTimeEstimator(fromsupport).AiGenerationRequestmoved fromdocumenttopromptand lost the header it carried and no provider read;AiProgressBarmoved fromsupporttoengine.supportgainedReadBudget(the read limit of a pattern, fromAiFactExtractor) andUrlUserInfo(the masking of a URL's user info, which the CLI reached through a provider class).LlamaCppJniConfigisProviderConfig, since it configures every provider, and the plugin's base classAbstractSrcMorphMojo. The ArchUnit layering names every package with the exact layers that use it:configno longer depends onprompt,promptandproviderno longer ondocument, and a new rule keepsindexer.generationfrom depending on the walkers. -
The map-reduce decisions are on the PIT gate.
SummaryGeneratorkept the chunk plan, the packing and shortening of a reduce round and the decision on an incomplete answer next to its provider calls, off the gate; they are the package-privateChunkPlan,ReduceRoundandIncompleteAnswerActionnow, 46 mutations, all killed. The behavior is unchanged. -
The benchmark's runs and tools are pinned better.
run.py's meta records the srcmorph commit (+dirtywith local changes), the Python version and a SHA-256 over the corpus; a test pins the byteslarge.pywrites at--scale 1.0, so a Python whoserandomdraws differently fails it.score.pyscores only the files srcmorph indexes and reads a source as srcmorph decodes it (UTF-8, UTF-16 after a byte order mark, U+FFFD for an invalid byte);report.pytakes--cardand a--helpthat does not render. A mutation pass over the tools found most of their logic untested:score.pynow has 20 of 114 mutants surviving (41 before),report.py5 of 17 (15). -
The run record and the plan say more of where the time and the tokens go. The record's totals name the files the plan skipped (
skipped), the run's wall time (wallSeconds) and each model's load time (modelLoadSeconds): a local model loads on its first file, whose own seconds leave the load out now (GenerationProvider.takeLoadNanos,0by default). The plan's model header adds the tokens the model is sent (~N source tokens (~M sent)): none of a filedeterministicsummarizes, one window of asampleorfail,maxChunkswindows of asampleChunks. Its map-reduce estimate packs as many part summaries into a reduce call as fit the window, as the run does, instead of always 16 (MapReduceMath.expectedFanIn), so a narrow window estimates more reduce calls. -
requestsPerMinutekeeps one window per server: two models of one server (samebaseUrlandapiKeyEnv) with 20 and 10 requests a minute had two windows and sent up to 30 together. Now every request waits until fewer than its own model's limit started in the server's window, whichever model started them, so the model at 10 waits for the other's requests too. The window lives as long as the process, so a later run of a long-lived JVM (mvnd, an IDE, an embedder) that names a higher limit sends at it. -
Gis srcmorph's own version, in the CLI and the plugin alike. The plugin recorded the version of the user's project as the generator and the CLI recorded none; a summary now names the srcmorph release that wrote it (support.GeneratorVersion, which the core's build fills in), andgeneratorVersionin the file still overrides it.Gnever causes a rewrite, so no summary is rewritten for this. -
The examples, the self-test and
run.pyask once more on an incomplete answer (onIncomplete: retryOnceon every rule, the package and project rules too). gpt-oss-20b ran away to the output budget on one big TypeScript file in six runs and summarized the same file in the next one; a larger budget only made the runaway dearer (measured on 2026-11-05: finished answers took 1229 to 1618 tokens, the runaway 16384).run.pyrecords it in the meta (onIncomplete); the runs before failed such a file at once. -
The time estimate is based on GPT-6 Luna, the benchmark's remote model, where it was a local CPU (gpt-oss-20b on a Ryzen 7 5800H):
2.1 s + 0.105 msper prompt token+ 6.45 msper output token, fitted over 1423 calls of the 2026-11 measurements and linear up to their largest prompts (98K tokens), 4.0 characters per token, 1000 tokens of prompt around the source, 950 output tokens assumed. For a 24K-token prompt the CPU model said 25 min, Luna takes about 12 s, so a remote run's plan was off by more than an order of magnitude. The per-file log line ends in(rough; based on GPT-6 Luna), and the plan names the basis in every model's header; a model with acalibrationreadsfrom the model's calibrationin both instead. A local model is far slower than the estimate:Calibratemeasures it. -
SourceFileIndexeris final, and its public constructor takesfileNames,fileInterpreters,regenerateOnandcontentHeadChars(Upgrading from 1.4.0). The file phase's log lines name a file by its path below the base directory (as the model reads it), with the system's separators, where they named the absolute path. -
Java takes
file-body-oop;file-body-javais gone from profile v3. A rule namingpromptKey: file-body-javafails the run (no such file in the profile): namefile-body-oop. Three runs each with both prompts on the model-card corpus gave the same Java recall, GPT-6 Luna 100 % in all three with the shared prompt (remeasure-2026-11-04/). The self-test, the examples, the plugin-it fixture andrun.py(--java-prompt, defaultfile-body-oop) name it. -
What rewrites a summary is set per cause (
regenerateOnundersrcMorph:source,prompt,facts,model,children, eachtrueby default), next toforce, which rewrites every summary whatever the switches say. Each switch compares one header field:C(the source; a package's files, the project's packages),P(the rule's prompts, map and reduce included),N(its fact counters), the newM(the model: definition key,remoteModel, model file name) and the newK(the child summaries a package or the project is written from). A missing or empty summary, another header format, node, title oraiVersionrewrite it always;D,TandGare recorded only. Up to now only a changed source,aiVersionorforcerewrote a summary (1.4.0 recordedPandNwithout acting on them), so a summary of an older prompt or model stayed; and a package was never rewritten when its files were re-summarized over unchanged sources, since its checksum is built from theirs, nor when its own prompt changed: packages and the project now recordP,MandKtoo (the project's checksum is built from its packages and their checksums, as a package's from its children, no longer folded with the overview's prompt key and model).Preads a prompt's line endings as\n, so a profile checked out with CRLF on one machine and LF on another hashes the same. A field a summary does not record counts as changed, so the first run after the upgrade rewrites every summary, one model call each (Upgrading from 1.4.0 says how to keep them).AiMdRewritePolicyTestpins every switch on and off,PackageIndexerTestandProjectIndexerTestthe children and overview hashes; the plugin README's header section says what rewrites a summary, and every example configuration shows the switches. -
java-factscounts code only: an[ignore]line blanks comments, Javadoc, text blocks, strings and character literals, so apublicorclassin them no longer counts, andTODO/FIXMEis a[raw]counter, counted in the comments as before. The facts line of a Java file with such text changes (java-ee-factsincludes it), and so does the facts hash, which withregenerateOn.factsrewrites the Java summaries that carry facts. -
DMN has its own counters:
bpmn-factscounts BPMN only (its "decisions" and "decision rules" went todmn-facts), and the Luna example andrun.pyroute.dmntodmn-facts. -
The model reads a file's path below the base directory, not its name alone, at the head of the user message (
billing/src/main/webapp/WEB-INF/web.xml), so two files of one name in different modules read apart; every file prompt of profile v3 says so in its last paragraph. The plan's window check counts the same line. A package's message starts with the package's path (see Fixed), the project's withproject.ai.md. -
Profile v3's file prompts ask for more (
remeasure-2026-11-04b/: the enterprise corpus 97.9 to 99.7 % with gpt-oss-20b, 97.3 to 99.9 % with GPT-6 Luna; the card corpus 99.5 % and 100 %, the language corpus 99.0 % over nine gpt-oss runs and 99.5 %):file-body-oopnames the private functions, constants and fields the logic uses and each call on a dependency, which the enterprise runs found by chance only (nextNumber,PREFIX, arefundcall);file-body-opsnames a CI workflow or pipeline by its name (release-backofficewas missing in every gpt-oss run) and defines Gradle build and settings scripts (plugins, dependencies by configuration, repositories, tasks, included projects, the toolchain), which take it now. -
pathGlobmatches in the case written, also on Windows, so a rule routes the same file on every system;fileNamesandextensionsmatch in any case, as names of a kind are written in either. Decided and documented, unchanged. -
A
{in a glob opens an alternative; it matched itself before. An exclude or apathGlobwith a literal brace pair now means its alternatives. -
A source file with a byte order mark is read without it, and one with the UTF-16 mark as UTF-16. Before, a UTF-8 file saved with a mark (as some Windows editors do) began with U+FEFF, so its first line did not start with its first word for the fact counters and the prompt, and a UTF-16 file read as replacement characters. The checksum of a UTF-8 file is still over its bytes, so such a file is not rewritten for this alone; a UTF-16 file's is over its text (see Fixed).
-
The fact counters of ten kinds count what their label says: Java counts a constructor whose first parameter is generic (
Reg(List<String> names)) and names with a$(FormatMessageW$MH), and its header no longer says the methods hold every constructor; OpenAPI counts an operation only one level below a path key, not a schema property namedoptionsorhead; batch'serrorlevelis a word (:exitWithErrorLevelis none); SQL counts PostgreSQL'sTEMP/RECURSIVEviews,EVENT TRIGGERandDOMAIN, Oracle'sNOFORCE/EDITIONINGviews andPRIVATE TEMPORARY TABLE; GitHub Actions blanksinputs:/outputs:entries (an inputruns-onis no runner) and countsschedule: [{cron: ...}]; Ansible counts an FQCN module only as a task key; Kotlin readsactual val, Scalaimplicit final class, Groovy<T extends Comparable<T>>, JSP Spring'sform:checkboxes,radiobutton(s)andhidden. -
Pretty-printed JSON OpenAPI and AsyncAPI documents count as their YAML form: the
[ignore]alternative that blanks a double-quoted string holding a#could start at a key's closing quote and blank the rest of the line, so"get": {,"/orders": {or"publish": {lost their colon and counted no path, operation, payload ordeprecated: true, nor did a double-quoted YAML key. A string now opens only at a line's start or after a:,,,[,{or-. -
No summary in the output directory is a source: a subtree that holds it (
subtrees: ["."]with the defaultsrc/site/ai) fed srcmorph's own summaries back in:Dockerfile.ai.mdmatchedDockerfile.*and was summarized, one level deeper with every run. The walk leaves out every.ai.mdbelow the output directory and still takes the sources an output directory holds (summaries next to their sources,outputDirectory: src); nor are srcmorph's other files a source, the calibration reportssrcmorph-calibration.json/.yamlin the output directory, therunRecordFileand the.parts.mdfiles belowpartsDirectory(a source there stays one; a*.parts.mdwritten by hand below it is no source either, andpruneOrphansremoves its summary). -
No walk enters a link to a directory below its root: on Windows the file walk and the project phase entered a junction (and so did the new orphan scan of
pruneOrphans), and the package phase followed a symbolic link on every system, so a link beside its target summarized every file twice and a link back up looped (64 copies of one file measured; on Linux the package phase failed withELOOP), and a link likedocs/latest -> v2rewrote the target'spackage.ai.mdon every run. Such a link is now left out with a warning (a debug line when an exclude takes it, or when it is the output directory or a subtree of its own). The start of a walk (a subtree, the output directory) is entered when it is a link: the file walk, the orphan scan and the project phase took a symbolic link there for a file and found nothing, so a linked output directory got noproject.ai.md. A link to a file is still taken, and a directory with another reparse point (a cloud file placeholder) still counts as a directory; a volume mounted into a folder carries a junction's reparse point and is taken for a link. Subtrees are compared by their real paths: one behind a link inside another subtree is now kept as a subtree of its own (it was dropped as part of the other), unless the link leads into a subtree, whose walk collects its files once. A junction whose target is gone is such a link too (the project phase failed on it, the file walk could not read it, and the orphan scan kept every package above it). A link to a summary file counts as that summary in every phase, andpruneOrphansremoves the link, not its target. A walk of a linked root names the target in an error message. The summaries an earlier run wrote below a link are orphans now. -
The file walk does not enter an excluded directory and goes on past an unreadable one: it descended everywhere and applied the excludes per file, so a directory it could not list ended the run with an
UncheckedIOException, also when an exclude covered it. A directory an exclude ending in/**takes as a whole (target/**,**/generated/**) is no longer entered; a file or directory that cannot be read is left out with a warning (a debug line when excluded), and so is a source whose content, size or modification time cannot be read (no read permission, a lock, deleted after the walk), which ended plan, check and generate; its summary is kept. The candidates and their order are unchanged. The package and project phases fail with theIOExceptionof a directory they cannot list, also when the listing breaks off midway, not with anUncheckedIOException. -
Allchecks the package and the project phase before the file phase (their rule, prompts and model, asCheckdoes), so a missingpackageRulesor a wrong phase rule fails before any file is sent and billed, not after the whole file phase; withplanOnlyonly the file phase is planned. The package and project checks now also fail on a rule prompt that has no template. -
calibrateleaves skip rules out and checks the rules as a run does: a skip rule's model is no longer loaded and measured, and one noaiDefinitionsentry defines no longer ends the run with anIllegalArgumentException. The rule set, the routed models and the prompts are validated first, and a provider's unchecked failure (the JNI binding'sIllegalStateException) fails naming the model. -
Calibration's prompt-cache line fits the provider: a remote model that reports no cached prompt tokens gets an INFO line instead of the warning about
cachePrompt/cacheReuse/swaFull, which never reach a server. Thecalibrategoal's log and the Javadoc say that anopenai-compatiblemodel's numbers are a wall-clock estimate with one extra billed probe request (the mock reports its rates), that a chat request is not resent after a read timeout, and thatProviderConfighas 44 fields. -
run.py keeps a paid run's meta when a cost lookup fails: the meta is written before the OpenRouter price and usage lookups and again with them; a timeout or
URLErrorrecords the cost asnullwith the reason undercostErrors, and the remaining models still run. A relative--model-diris made absolute; it resolved against the generated configuration's directory. -
score.py and search.py read the skip rule: a file the secrets rule skips no longer scores as a missing summary or counts as a search file (
run.skipped); search.py reads a source as srcmorph does (score.source_text) instead of strict UTF-8. An unclosed fence in a list item ends with the item, so a following heading is counted (no stored summary reads differently); correlate.py printsn/afor a correlation of no runs instead of failing. -
Example comments say what the configurations do: the Luna example does not index JavaScript; gpt-oss takes every file but the large Java, SQL, POM and plain XML files (a 500 KB shell or Kotlin file stays on it, map-reduced); the self-test names two packages instead of reading compile source roots.
-
A
sizecondition measures the same file on every checkout. It read the bytes on disk, so a file checked out with CRLF line endings was larger than the same file with LF and could reach another size tier, model and summary (a 272 KB Java file, 5000 lines). It measures the source as srcmorph reads it now: the bytes the checksum is taken over,\nline endings and no byte order mark (a UTF-16 file: the UTF-8 length of its text); the plan's time estimate of a file that fits does too. The run'sminFileSizeBytes/maxFileSizeBytesband stays a filter on the bytes on disk. -
A package's lead in
project.ai.mdis its summary's lead again for a package read in chunks (packages default tomapReduce) or with counters: the project listed the coverage or the facts line, which stands in front of the summary, as the package's lead. -
A UTF-16 source has one checksum on a CRLF and an LF checkout. Its line endings were normalized byte by byte, which turns the two bytes of a UTF-16
\r\ninto two line feeds, so a checkout switch rewrote its summary; its checksum is taken over its normalized text now. -
A leading
Src/stays in the summary's path on every platform:Path.startsWithignores the case on Windows, so the summary ofSrc/main/Foo.javalost itsSrcthere and kept it on Linux. -
A model id with a line break cannot break out of
ListModels' output: control characters and YAML line breaks are written as\uXXXXin the table and the paste-ready block. -
A
baseUrlwith a query or a fragment fails the check: srcmorph appends/chat/completionsto it, which landed in the query, and the query (possibly a key) was printed with the endpoint. Every URL srcmorph prints (a message, a log line,ListModels, the CLI's configuration echo) shows its query and fragment as?***now, as it shows user info as***@. -
A subtree that is
srcitself (subtrees: ["src"], or a POM's<sourceDirectory>src) no longer fails the package phase: its summaries' root is the output directory. -
An interrupt ends the package pass, as it ends the file phase, instead of failing every remaining package; a timeout is still one package's failure.
-
A request that failed its retries with a timeout last says "failed N times, the last with a timeout", not "timed out N times" after three 503 answers and one timeout.
-
A password with
',?or#in abaseUrlis masked and refused as user info. -
A long password in a URL is masked too. The mask of a URL's user info stopped at 256 characters, so a longer token (a JWT) stayed in the message that refuses such a
baseUrl; an@in the password left its tail visible. It masks up to the last@before the host now, of any length, andtoStringof a model definition and of a provider's configuration masks thebaseUrlas well. -
A
lazyMode,cacheTypeKorcacheTypeVthe binding does not know fails the check, before a model is loaded, instead of the first file's load; and a local model that does not load is tried once per run: every later file fails with the same error at once instead of loading the model again. -
One unchecked provider error no longer ends the package pass: the package is counted as failed and the others are written, as in the file phase (the llama.cpp binding throws only unchecked exceptions, so one context overflow left every later package unwritten).
-
A package call names the package, as
package-bodyandprofiles/README.mdsay: its user message started withpackage.ai.md(the file name of the absolute path), now with the package's path below the output root (main/java/com/example; the root package by the output directory's name). -
A prompt or facts file checked out with CRLF line endings reads as with LF. Its hash ignored the line endings already, but the template was sent and measured as written, so the same profile gave another window and other chunk boundaries on Windows than on Linux.
-
-Dsrcmorph.planOnly=truereaches the aggregation goals too. Onlygenerateread it, so a build with the self-test profile planned the file phase and then letaggregate-packagesandaggregate-projectload the model and rewrite every package; they do not run now, as the CLI'sAllcommand never did. -
The window counted the prompt template twice.
InputWindowCalculator.maxInputCharstook the template out of the budget the whole prompt (template and source) is then compared with, so every file lost one template's length of window (3 to 6 KB with profile v3's prompts) and some that fit were read in chunks, or failed underonOversize: fail.maxInputChars(config)is the budget of the rendered prompt now (the window less the output and the safety margin), and the source gets it less the template, once; the plan and the run still agree. Java callers drop the second argument. -
A phase rule's counters from an inline
factDefinitionsentry are checked before a model loads: they were resolved after the rule's check, so a pattern that does not compile failed the run at the first package instead. -
partsDirectoryno longer writes a chunked file's answers under the next file's name when that file was unchanged (or failed before it was read):SummaryGenerator.takeParts()returns the answers once and starts anew (parts()is gone). -
Files are planned and summarized in path order, on every machine. The walk of a subtree came in the file system's order, so the plan, the progress and the run record listed files differently on Linux and Windows, and a test that relied on two files' order failed on one CI runner and not on another. The candidates are sorted by their path with
/separators now. -
A template that opens the thinking block itself leaves no reasoning in the summary. Qwen3 and DeepSeek-R1 templates put
<think>into the prompt, so the answer starts with the reasoning and holds only</think>;CompletionParserkept all of it. The first</think>alone on its line, with no<think>before it, now ends the reasoning, and without one the first that ends its line glued to the reasoning's last word (I will write it.</think>) when no line before it starts like an answer (a>lead, a code fence); a tag inside a line, or after a space, a quote or a backtick (a sentence, quotes, a list item naming it) is text, in the answer or in reasoning that names it before its real end. -
The benchmark's scorer (
score.py,report.py; the card's published numbers stay until it is re-rendered): an everyday word between two backtick spans no longer counts as marked as code, one in a column listing with its type (email VARCHAR(254)) does; a reasoning marker in backticks is no leak, and a complete summary after a reasoning block is a leak, not a leak-only failure; the card's invented-name rate sums the names of every kind instead of averaging the kinds' rates (summary.jsoncarriestickedandinvented;report.pyrefuses one without them);CHECKOUT,RETURN,Sinceandhistogramare everyday words of their files. -
The scorer's copy filter and invented names (TODO row 43; the card stays until it is re-rendered): a
~~~fence, an indented code block and a copy with its whitespace changed copy the source as a ``` fence does, and do not count. A name is invented unless it is an identifier of the source, a run of an identifier's word parts (split at_and at a capital) or their plural, so a piece inside a word (`ame` of `getName`) no longer passes as a substring; the names are those in backticks and the code-shaped words of the prose (`orderTotal`, `OrderService`, `order_total`), and the names of languages, tools and products (`JavaBean`, `PostgreSQL`) and Python's builtin types are none. Over the outputs of 2026-10 no recall changes, and the invented-name rate of a model's kind rises by at most 1.3 per 100 names; qwen3-coder-30b's SQL falls from 18.6 to 3.7, since its few backticked names are joined by the names of its prose. -
The fact counters read more of what they count, and less of what they do not (every facts hash of the changed files changes, so
regenerateOn.factsrewrites their summaries once):- shell: a
<<after up to three groups in parentheses ($(( (a) + ((b)) << n ))) is a shift, andMSG=${MSG:-a b}is an assignment, not the environment prefix of a command; - Perl: a
<<after a hash element ($h{x} <<BITS,$r->{k}) is a shift, one insideq{},qq//,qw,qr,m//,s///,yortropens no heredoc (print {$fh} <<EOTstill does); - SQL: an apostrophe in PostgreSQL's one-line dollar quoting, Oracle's
q'[...]', a MySQL"..."string or a MySQL#comment opens no literal that hid the next statements (quoted identifiers, T-SQL#temptables and multi-line$$function bodies stay); MariaDB'sCREATE OR REPLACE TABLE, PostgreSQL'sCREATE FOREIGN TABLEand T-SQL'sCOLUMNSTOREindexes count; MySQL'sVALUES(col)ofON DUPLICATE KEY UPDATEis no row; - Java: annotations on the declaration's own line, type parameters three deep and parameters
that start on the next line count;
yield f(x);, a record header and an enum constant with a body (PLUS("+") {) are no method or constructor; - XML:
idandnamecount inside a tag only, not asname = 'a'in element text; - a blanked region keeps every line terminator Java's regex reads (NEL, U+2028, U+2029 too), so
^sees the same lines as before.
- shell: a
-
The fallback prompt's example
MAX_RETRIESisPOOL_SIZE:MAX_RETRIESis a scored name of the language corpus, which a model could copy from the prompt (the prompt hash changes). -
An answer the provider stopped was written as a finished summary. OpenRouter answers a generation that failed half-way with a 2xx, the text so far and finish reason
error; onlylengthcounted as incomplete, so the summary ended mid-sentence, without a warning (measured on 2026-11-05: gpt-oss-20b, 694 tokens, ending intest paging (). Finish reasonerrororcontent_filteris now an incomplete answer of the new reasonABORTED: not written and left for the next run (onIncomplete: retryOnceasks once more first), never kept as a part or underonIncomplete: keep; the calibration warns about it. Any other reason (stop, a server's own word for it, none) is the model's, as before. -
An ignored region at the start of a line left a space there, so a counter anchored at the line start (
^class,^import, Perl's^my) missed the code after it:/* x */class A {}counted no type, the same code on two lines one. Such a region now leaves nothing; within a line it still leaves a space, so the words around it stay apart. The facts hash is unchanged (it hashes the counters, not the counts). -
A JSON configuration file with a UTF-8 byte order mark did not load: Windows PowerShell 5.1 writes one (
Out-File -Encoding utf8), and Jackson read it as a character before the{. The loader reads a configuration file as it reads a source, without the mark (YAML too). -
A key defined twice in
aiDefinitions,promptDefinitionsorfactDefinitionswas taken silently, the last one winning, so a rule could run with another model or prompt than the one read first. It now fails the run, naming the list, the index and the key. -
docs/copilot-api-gateway.mdshowedmaxOutputTokens: 2048for the example, which has 4096. -
The XML and POM fact counters of profile v3 were quadratic over unclosed markup. In
xml-facts.tsvthe[ignore]line scanned to the end of the file again from every<!--,<![CDATA[,<?or<!DOCTYPEthat is never closed (100 KB of unclosed<?took 8.4 s; a<!DOCTYPE r [ ]without>repeated ran for minutes), and so did the XSD counter from every unclosed<complexType/<simpleType; inpom-facts.tsvthe comment and CDATA patterns and the plugin-body pattern from every unclosed<plugin>. An unclosed region now runs to the end of the file, as insql-facts.tsv, a type tag ends at the next<, and a nested plugin body at the next plugin tag. Well-formed files count as before (the corpus counts are unchanged); 11 newXmlFactPatternsTestcases pin the unclosed regions and the run time of each pattern. -
Three more fact counters were quadratic, found by the new linearity test: in
java-facts.tsvthe type arguments of "methods (approx)" and "field declarations" and the field initializer read across lines to the end of the file from every declaration without its>or;, and inshell-facts.tsv"dash-option case arms" backtracked over a long run of blanks. Type arguments are now read within one line, an initializer up to 10,000 characters (a generic type split over lines, or a longer initializer, is no longer counted;JavaStructureFactPatternsTestpins both), and the blanks possessively. -
search.py no longer overwrites the card's search scores with a partial run. With the card's scores/ directory,
--no-reader(which nulled every reader result) or a non-default--out(which replaced the card's runs) is refused with a hint to pass a--scoresof its own, as repeat.py already refuses--runsthere. -
run.py hands each model only its own API key. The CLI's environment drops every
apiKeyEnvnamed in models.json except the one the model's definition names;COPILOT_GATEWAY_API_KEYreached every model's subprocess, against the script's docstring. -
search.py's reader no longer loses or freezes answers. An empty answer is asked again up to four times and never cached (one already in an older cache is asked again too);
"usage": nullno longer raises before a paid answer is cached; a 200 whose body is no JSON or is cut off is retried like a 5xx. -
score.py counts a heading indented by one to three spaces, as CommonMark reads it; no published summary has one, so the card's figures do not change. Three of its tests are renamed or split to say what they check.
-
SQL: an identifier's closing quote opens no "..." string. In
INSERT INTO "cfg" ("data") VALUES ('{"k": "v"}')the[ignore]alternative for a MySQL string with an apostrophe ran from the closing quote of"data"to the JSON literal's first", and the literal's apostrophe, left unpaired, opened a literal over the rest of the file (INSERT 1, VALUES rows 0, tables 0, indexes 0 of 2, 2, 1, 1). -
A SQL*Plus remark or pg_dump's COPY rows open no SQL string literal. The sql-facts
[ignore]line opened a'...'literal at any apostrophe, also on a line that is no SQL:REM Creates the shop's schema/PROMPT Don't ...before twoCREATE TABLEand oneCREATE INDEXcounted tables 0 and indexes 0, now 2 and 1; a pg_dump whoseCOPY ... FROM stdinrows hold anO'Brienhid theALTER TABLEand index after them (0 and 0, now 1 and 1). AREM[ARK]/PRO[MPT]line and the rows of aCOPY ... FROM stdinup to\.are blanked; without the\.line the data runs to the end of the file, as psql reads it. -
A Groovy method with a generic return type counts. "methods (approx)" read the type arguments possessively up to the line end, which ate their
>:List<Order> all() {,Map<String,Integer> counts(int x) {,public static List<String> names() {andString plain() {counted 1, now 4. An annotation on the method's line and a brace on the next line count too. -
A type-use annotation after a field's modifier no longer hides the field (java-facts, and so java-ee-facts):
@Nullable private String a;,private @Nullable String b;andprivate final @NonNull List<String> c = ...;counted 1, now 3. -
A CDI, Guice or Dagger
@Singletonis no EJB session bean. java-ee-facts counted every bare@Singleton(import javax.inject.Singleton; @Singleton class Cache {}counted 1, now 0). An EJB singleton counts through itsjakarta.ejb/javax.ejbimport or when written with its package; one imported throughjavax.ejb.*is no longer seen. -
Kotlin object expressions, annotated and vararg constructors count as written: two
object : Runnable {expressions next toobject Registryanddata object Emptycounted 4 type declarations, now 2;@Inject constructor(...)and@JvmOverloads internal constructor(...)0 secondary constructors, now 2;vararg valandfinal val0 of 3 constructor properties, now 3. -
Scala package objects, infix defs and abstract type members count as written:
package object shop {counted 0 type declarations, now 1;infix def max(...)andprivate infix def min(...)0 methods, now 2;type T <: AnyRef,type Uandtype V >: Null <: F[A => B]next to five aliases 8 type aliases, now 5. -
A context overflow's retry sends a shorter prompt, not the same one again. The next window was 75 % of the window budget, so a prompt far below it (a dense 12 000-char prompt in a 30 000-char window, on a provider that cannot count tokens) was sent four times unchanged and the file failed: a
mapReducerule never chunked it. The next window is 75 % of the refused prompt; a prompt of 1000 chars or fewer is not sent again. -
A trimmed prompt reads the long line behind a short one. The trim backed up to the last line break before the window, so a cut map call of a one-line (minified) chunk kept only its label ("Chunk 3 of 9, half 2, half 1:"), and the sampled head of "header + one long line" read the header alone. A line break that keeps less than half of what fits is no longer the cut.
-
No cut splits a character outside the Basic Multilingual Plane: trimmed prompts, chunks, shortened part summaries, halved chunks and the counted head stop before a surrogate pair.
SourceChunker's Javadoc no longer claims amaxChunksof 1 keeps the last chunk. -
The whole plan is in path order: files were sorted within each subtree and the subtrees taken in configured order; all candidates are sorted by path together now, as this file said they were.
-
An interrupt ends the file run and the package pass however it surfaces. A
ClosedByInterruptException, or anyIOExceptionafter the interrupt flag was set, was taken as one file's or one package's failure, and every remaining one then failed too. -
The plan's time counts what the run does: no reduce call for a one-partial batch that is passed on (17 partials: 2, not 3); characters, not bytes (a CJK source was about 3 times too long); 0 s for a mapReduce or sampleChunks file whose window leaves less than 1000 chars for a chunk, which the run refuses; the plan and
checkfail on such a file now, before a model loads, naming the room its chunks would get. -
A
nullrule in a list built in code fails the run naming it (Invalid rules: srcMorph.rules[1] is null) instead of aNullPointerException; the configuration file's loader refuses one already. The classes that skip anullrule or counter say so in their parameter types (? extends @Nullable), null-safety finding S9. -
A deterministic body fences a Markdown source safely: the fence is one backtick longer than the longest backtick run in the lines it shows (at least three), so a ``` line in the source no longer closes the block.
-
A
#in a JSON string no longer blanks the rest of an OpenAPI or AsyncAPI spec. A double-quoted string holding a#is blanked whole instead of opening a YAML comment: a one-line JSON spec with"Fixes issue #12"counted 0/0/0 paths, operations and operationIds, now 2/3/3. -
An empty YAML block scalar no longer blanks its sibling keys (asyncapi, compose, ansible, camel-yaml and openapi facts): the body must be deeper than its key, past a list item's dash, and the key line stays. After an empty
command: >healthchecks and port lists went from 0/0 to 1/1. -
properties-facts reads keys as
java.util.Propertiesdoes: a continuation line is no key; a key separated by a blank alone (welcome message) or with no value counts. -
batch-facts counts no echoed text and no exit labels: the text an echo prints is blanked, and "exits" counts the exit command, not
:exitorgoto :exit(6 to 2 on the pinned sample). -
gradle-facts counts no plugin declarations, publications or task references: plugins went from 5 to 2 (
gradlePlugin { plugins { id } }, developer ids), repositories from 2 to 1 (maven(MavenPublication)), tasks from 3 to 1 (dependsOn tasks.named(...)). -
JSP documents count their directive elements and blank their XML comments: jsp-facts counts
<jsp:directive.page|include|attribute>, and the new jspx-facts blanks XML comments (a .jspx went from 0/0/1/2 page directives, includes, scriptlets and EL to 1/1/0/1). -
Counters that counted the wrong items: GitHub Actions matrix entries are no steps (5 to 2); Ansible loop items are no tasks (5 to 2) and template expressions in block scalars count; PowerShell Verb-Noun calls count in any case and a function definition is no call; a GraphQL schema definition holds no fields and a field with nested argument parentheses counts; WSDL counters accept any element prefix and count an operation once; a manifest embedded in a ConfigMap counts no container; AsyncAPI message examples are no payloads (3 to 1); npm contributed commands count with their keys in any order (0 to 2).
-
file-body-oop names TypeScript, and file-body-xml indexes SQL, scripts and conditions in CDATA.
-
The secrets skip rule skips credential files, not CI workflows, Helm templates or Android settings: secret-scan.yml, a Helm templates/secret.yaml, vault-deployment.yaml and Android's res/xml/settings.xml are routed again; secrets.hcl and *.pkrvars.hcl are skipped. run.py gets the same rule and routes build.gradle.kts to gradle instead of kotlin.
-
A password with an unencoded
/in abaseUrlis shown nowhere. The mask stopped at the first/, sohttps://user:pa/ss@host/v1reached the CLI's configuration echo, the ListModels log line and error and bothtoStrings unchanged; an@after a URL's first/masks everything after the scheme now (https://***). The check refuses everybaseUrlthat holds an@, includinghttps://u:123/abc@host/v1, which parsed as hostu, port123, and passed. -
A redirect is not followed with the key. The models GET followed a 30x with
Authorizationto whatever server theLocationnamed, and a streamed POST reported a redirect as a refused key; a 3xx fails the request now, naming the status and the maskedLocation. -
A
</think>alone on its line inside an answer is kept: a summary that quoted the tag in a code fence lost its lead and was stored as complete from the fence on. The tag ends the reasoning only outside an open fence and before any>lead line. -
A pattern that overflows the stack is stopped like one past its read budget. A fact counter or
contentMatchessuch as(?:a|b)*over a long text threw aStackOverflowErrorthat aborted the plan or the run; the counter is left out now, an[ignore]line blanks nothing, the condition does not match, and a warning names the pattern. -
Two ignored regions back to back at a line start leave the line starting with what follows (
/* a *//* b */class A {}counts as a^classline again), and a facts file of[ignore]lines only can be included (the file that includes it must still count something). -
An answer the provider stopped inside open reasoning is ABORTED, not TRUNCATED: one that ended with finish reason
errororcontent_filterinside<think>asked to raisemaxOutputTokens. -
requestsPerMinutewaits once, to the start that frees the window, instead of once per start with a "Holding ..." line each; the server in that line and intoStringis masked.
1.4.0 - 2026-11-02
The entries below hold the details; what a 1.3.0 user changes:
- Maven plugin: the configuration moves into a file. The goals' XML parameters are gone. What the
POM's
<configuration>held (aiDefinitions,fieldGenerations,promptDefinitions,factDefinitions, the paths,generationProvider, ...) goes intosrcmorph.yamlnext to the POM, undersrcMorph:, in the CLI's format; the package and the project rule go into the top-levelpackageFieldGenerations/projectFieldGenerations. The POM keeps the goals,configurationFile, the skip switches andsrcmorph.force/srcmorph.planOnly. Without the file every goal fails, naming where it looked. The worked example is insrcmorph-maven-plugin/README.md. - One rule per aggregation phase, no
llama*run parameters.packageFieldGenerationstakes exactly one rule andprojectFieldGenerationsnone or one; neither falls back to the file rules any more.llamaModelPath,llamaContextSize,llamaMaxOutputTokens,llamaTemperatureandllamaThreadsare removed: every model parameter is a field of anaiDefinition. - Paths resolve against the configuration file's directory, not the working directory, for the
CLI too; a
${name}that no property sets fails the load when it stands in a path. - Defaults that changed: the package and project rules map-reduce a long input
(
onOversize: mapReduce); an incomplete answer is never written (onIncomplete), so a file whose answer stays empty or cut off is reported, fails the run at its end and is tried again by the next run. - No runnable jar is attached to the release any more: build one for your machine with
examples/build-your-own-jar(root README). - Existing summaries stay: a new srcmorph version alone regenerates nothing (the plugin version and the source's modification time are no longer compared); a summary 1.3.0 wrote from an incomplete answer is written again.
- The plan shows the source tokens. Every routed file has an estimated token count (its
characters over its model's
charsPerToken), every model section the sum, and a section sums them per directory, so the plan says where a run's tokens and cost go before it starts. - A record of every run. With
runRecordFileset, the file phase writes a JSON record: per routed file its rule, prompt, model, outcome, calls, prompt / cached / completion tokens, the finish reason of its last call, its wall time and the CRC32 of its prompt and its facts, and the totals (engine.RunRecord). The providers report each call's tokens and finish reason (AiGenerationProvider.lastUsage(),provider.AiGenerationUsage), also for a call that ended incomplete;AiFieldGenerationSupport.takeUsage()sums the calls of a file, a map-reduce's included. mapReducereads every chunk; sampling issampleChunks, and every summary says what it read.onOversize: mapReducesummarizes every chunk and reduces in batches until one summary is left; amaxChunkscap used to sample a subset silently, so a 1.6 MB dump was summarized from a third of it while the configuration read as mapReduce. Sampling is the new strategysampleChunkswithmaxChunks(required there, and refused with any other strategy). An over-window summary begins with a coverage line:**Coverage:** 9 of 9 chunks summarized,... 6 of 9 chunks summarized (sampled), or undersample**Coverage:** the first N of M characters (sampled head). The plan's time estimate ofmapReduce/sampleChunkscounts the chunks with the run's budget and overlap and the reduce calls by the fan-in. The self-test and example configurations keep their six-chunk huge-file rules assampleChunks. A rule-set error ofgenerate(two fallbacks, a maxChunks without sampleChunks, ...) is aSrcMorphException, so the plugin reports it as a configuration error.- The jar smoke test counts what it planned.
smoke-jarrunsconfig_Plan.jsona second time and asserts the plan's one file of the demo tree, so a launch that finds no sources fails. The examples say they need 1.4.0 (examples/build-your-own-jar/pom.xml, the root README); stale comments and references inpublish.yml, the poms and CLAUDE.md (versions, layout, test lists) match the code, and the root pom declaressonar.organizationonce. - The Maven plugin is configured by the CLI's file;
checkgoal and command. The plugin reads one JSON or YAML file, the same one the CLI reads (<configurationFile>,-Dsrcmorph.configurationFile, default${project.basedir}/srcmorph.yaml), through one loader in the core (SrcMorphConfigurationFileLoader,SrcMorphConfigurationFile). Breaking: the XML parameters of the goals are gone (aiDefinitions,fieldGenerations,promptDefinitions,factDefinitions,subtrees,excludes,fileExtensions, the size band,outputDirectory,generationProvider,modelDirectory,profileDirectory,aiVersionand theirsrcmorph.*properties); the POM names the goals,configurationFile, the skip switches and the run switchessrcmorph.force/srcmorph.planOnly, which override the file whentrue. What the file leaves unset, Maven fills:baseDirectory,subtrees(the source directory and the resource directories that exist below the base directory),projectName,pluginVersion. A${name}in a value of the file is a Maven property (the CLI: a system property), filled after parsing, so a Windows path needs no escaping. Each goal takes its section:generateandcalibratethe file rules,aggregate-packagesthepackageFieldGenerations,aggregate-projecttheprojectFieldGenerations. The self-test profile issrcmorph-maven-plugin/srcmorph-selftest.yaml(its-Dai.*still work), with the repository root as base, so its two subtrees lie inside it. Relative paths of the file, for the CLI too, resolve against the file's directory instead of the working directory (the root README's quickstart fromexamples/build-your-own-jarworks now); an unset CLIbaseDirectoryis that directory.srcmorph:checkand the CLI'sCheckcheck the plan, the profile files and every model of the three phases (file and GGUF header,baseUrl,apiKeyEnv) without loading one (GenerateEngine.check(),AggregatePackagesEngine.check(),AggregateProjectEngine.check()). A rule naming a model noaiDefinitiondefines fails as a configuration error naming the rule, not as anIllegalArgumentException. The CLI prints its usage with--help/-hand fails with it on a wrong argument count (it used to log and exit 0); an unknowncommandnames the commands. The plugin descriptor states Maven 3.6.3 and Java 1.8 (it said 3.10.0 and Java 11). - A source is read once, with
\nline endings. The checksum, the plan's window check, the prompt, the facts and the map-reduce chunks see the same text, whether the working copy has\r\n(Windows withcore.autocrlf),\ror\n: a CRLF checkout no longer has another checksum than an LF one, and no\rreaches the model. The checksum is now taken over that text instead of the file's bytes, so a summary written from a CRLF working copy is regenerated once. The line count of alinescondition and the deterministic body count\r\nand\ras line endings too (AiSourceLines); the deterministic body no longer carries\rinto its fenced lines. The pure parts of the field generation moved to classes on the PIT gate: the time estimate of the plan and of a run (AiRouteEstimator, one code path instead of two copies), the map-reduce sizing and reduce batches (AiMapReduceSupport), and the smaller window after a context overflow (AiOverflowBackoff, which also recognizes a remoteAiContextOverflowExceptionin a cause chain). A reduce round whose partials are too long to combine now goes straight to the last-resort single reduce instead of first summarizing every batch for nothing.LlamaCppJniConfig$BuilderandAiIndexPlan$Entryare on the gate;AiChecksumSupport.calculateCrc32Hex(Path)is gone. - One rule per aggregation phase; the
llama*fallback parameters are gone.aggregate-packagestakes exactly one rule (its prompt and model) andaggregate-projectnone or one; more fail the goal before anything is written, as does a missing package rule. Before, the package phase generated every package once per rule and kept the last answer, loaded the provider of the first rule's model whatever the others named, and the project overview took the first rule, so the file rules of a CLI run reached both phases unnoticed. The CLI'spackageFieldGenerations/projectFieldGenerationsno longer fall back tosrcMorph.fieldGenerations: without them the package phase fails and the project index has no overview (examples/config_All.jsonsets both). The aggregation phases check their model before the run likegeneratedoes (model file, key,baseUrl). With a rule required, the run-level model parameters were dead:llamaModelPath,llamaContextSize,llamaMaxOutputTokens,llamaTemperature,llamaThreads(srcmorph.llama.*) are removed from the plugin and fromSrcMorphConfiguration, a CLI configuration naming one fails to bind, andLlamaCppJniConfigFactory.fromFallbackParametersis gone; every model parameter is a field of anaiDefinition.AiFieldGenerationSupport.processFieldGenerations(List, …)isgenerate(rule, …). The indexers keep their rule as a full copy (AiFieldGenerationConfig.copy()); the project overview used to copy only the prompt, the model andonIncomplete, so it ignored its rule'sonOversizeand facts. - The engine plans and runs alike, and checks its configuration before the run. The source budget
is computed for every file from the prompt with that file's name, as the plan computes it; before, it
was cached from the first file of a model group, so a later file with a shorter name could be
trimmed although the plan said it fits, and
onOversize: failwas not enforced at run time (it is now, as a failed file). A configured subtree that does not exist or lies outsidebaseDirectory(another drive included) fails the run instead of being skipped with a warning, which silently indexed the default tree; two sources summarized into one.ai.mdand a prompt that leaves its model's window no room for the source fail before the run. Extensions match with their dot and in any case (mdno longer matchesREADME.cmd). The incremental check no longer compares the source's modification time or the plugin version, and a package's checksum no longer includes its children's times: a fresh clone, a touched file or a new srcmorph release regenerate nothing whose content is the same. - Retries and a context overflow for the
openai-compatibleprovider (maxRetries). A request that failed with408,429, a5xxstatus or a read timeout is sent again up tomaxRetriestimes (default 3), after the server'sRetry-Afterin seconds or an exponential backoff from 1 s with jitter; other failures are final at once. A prompt the server refuses as longer than the context (413, or400saying so) is anAiContextOverflowException, and the window is shrunk and the prompt asked again, as for a local model. One file or package whose request still fails no longer ends the run: it is counted, the others are indexed, and the run fails at its end. A namedapiKeyEnvthat is not set, or a key that would go over plainhttpto another machine, fails the run before any request;apiKeyEnvis trimmed, andlocalhostand the IPv6 loopback count as loopback in any written form. - An incomplete answer is never a summary (
onIncomplete). A model answer that is empty (also a reasoning-only one), cut off at the output budget, or a response without an answer (no choice, an HTML page or an{"error": ...}object with status 200) is refused by the provider (AiIncompleteAnswerException) instead of being written as the file's summary. Nothing is written for that file, package or project overview, so an older summary stays and the next run tries again; every other file is still indexed, and the run fails at its end naming the count. Per rule,onIncompleteisfail(default),retryOnce(ask once more) orkeep(write a cut-off answer that has text, the behaviour before). Before, the facts line alone or the package'sContentslist was written in place of the missing summary and counted as up to date, so an incremental run never tried it again. A leading<think> ... </think>block is stripped like Gemma's, and an unclosed one is a cut-off answer.AggregatePackagesEnginereturns the packages written (no longer the unchanged ones too);SourceFileIndexer.indexFile,PackageIndexer.aggregateandProjectIndexer.aggregatereturn anAiIndexOutcome/AiIndexTally. - Profiles: prompts and fact counters as files, named by key (
profileDirectory). A profile is one directory with a file per prompt (<key>.md, the system prompt verbatim) and per group of exact fact counters (<key>.tsv:label<TAB>regexper line,#comments). WithprofileDirectoryset (in the configuration file: relative to the file's directory; in aSrcMorphConfigurationbuilt in code: relative to the working directory; absolute always as it is), a rule'spromptKeyandfactsKeythat no definition has are read from there, so a configuration names keys only and repeats no prompt text and no regex. A key found nowhere fails the run before a model loads, naming the file it looked for; a pattern that does not compile fails it with file and line. A definition still works and wins over a file of the same key: a prompt is atemplateor atemplateFile, a fact groupfactsor the newfactsFile(both files relative toprofileDirectory). Every file is read as UTF-8, a leading byte order mark dropped; an empty file is an error. In the library:AiPromptTemplateLoader,AiFactDefinitionLoader,AiProfileFileSupport. - Fact counters can leave regions out (
[ignore]). A counter whose label starts with[ignore]in a facts file (in a configuration:ignore: true) is not reported: its matches are blanked out of the source, every character except the line breaks, before the other counters count. A sub named in Perl's POD, a function or aset -einside a shell heredoc, a commented-out POM dependency or anINSERTin a SQL block comment then no longer counts, while^anchors see the same lines.AiFactExtractor.countsreturns the counts by label. A configuration'sfactsand a facts file pass one validation,AiFactExtractor.validate(where, counters): every counter present, with a label and a compiling pattern, no label twice or starting with[, and not only ignore counters; the[ignore]prefix is read in any case, another[...]marker is refused. Several ignore lines blank in file order, so the profile's files have one per language, an alternation of all its regions where the one that starts first wins, as in a lexer (a<!--inside CDATA, POD inside a Perl heredoc or a/*inside a SQL string opens nothing), and their counters anchor with^[ \t]*, not^\s*, which is quadratic over the line breaks a long blanked region leaves (a 40 000-line SQL comment: 44 s before, 0.05 s now). - Fact files for shell, Perl and XML:
shell-facts,perl-facts,xml-facts,pom-facts. Bash, POSIX sh and ksh93 (functions, getopts loops and their option letters, dash-option case arms, named positional parameters, exported and readonly variables by name, defaulted variables, script-level assignments, sourced files, traps,set -e, exit statuses, kshtypesetand discipline functions), Perl (packages, subs, modules used and required,use constant,ourand file-scopemy,%ENVkeys,GetOptions,die/croak/confess), XML (elements, namespaces,id/nameattributes, XSD types, XSLT templates, Spring beans) and Maven POMs (dependencies, managed sections, plugins, modules, profiles, test scope), each with its comments, heredocs, POD or CDATA ignored. SQLINSERTcounts withoutINTO, after an Oracle hint and asINSERT ALL/FIRST; MySQL'sDEFINER=views and Oracle'sCREATE OR REPLACE TYPEcount, aTYPE BODY,MATERIALIZED VIEW LOGandDATABASE LINKdo not. Shell comments after code are blanked, a function body may be any compound command,sourcecounts after&&/then/;, an environment prefix (LC_ALL=C sort) is no assignment, anddeclarecounts withtypeset(labeltypeset and declare declarations). Perl counts 5.38class/methodanduse parent/base, and%ENVreads without its assignments. XML blanks the DOCTYPE; POMs blank every plugin body, so a plugin's dependencies and configuration do not count, andmanaged plugin sectionsis new. Routing takes.zsh,.bats,.t,.cgi,.psgi,.wsdland.xhtmltoo, and a POM above 275 KB keeps the Maven counters in the Granite tiers (huge-pom,big-pom).ShellFactPatternsTest,PerlFactPatternsTestandXmlFactPatternsTestpin every counter. The self-test profile andexamples/build-your-own-jar/srcmorph.yamlroute.sh/.bash/.ksh,.pl/.pm,**/*pom.xmland.xml/.xsd/.xsl/.xsltto them, and big XML files take the Granite size tiers Java and SQL have. - Prompts for shell, Perl and XML:
file-body-shell,file-body-perl,file-body-xml. Up to now these files tookfile-body-fallback. The shell prompt asks for every argument by its variable, every option as typed with its parser, every top-level variable asname=value(the lower-case and derived ones too), every function and the tables or hosts of its heredocs; the Perl prompt for subs,Getopt::Longoptions,%ENVkeys, constants and imports; the XML prompt for what the file defines by kind (POM, Spring context, XSD, XSLT, other data). On the benchmark's language corpus gpt-oss-20b went from 95.7 to 97.7 % over six runs (99.1 % without one empty answer of the provider), GPT-6 Luna from 99.6 to 99.9 %; on the other two corpora gpt-oss lost a few names of unchanged Java and SQL prompts, which three runs cannot tell from variation (profiles/README.md, model card finding 7). The self-test'sshell,perl,pomandxmlrules and the example configuration use them. The Granite tiers for big XML,big-xml(H-Tiny) andhuge-xml(H-1B), takefile-body-xmltoo: on an 870 KB catalog H-Tiny named 7 of 8 names with the XML prompt and 4 with the generic one; on a 1.15 MB catalog, sampled ashuge-xmlsamples it, H-1B wrote a short summary with the XML prompt and failed the file with the generic one, in each of three runs (docs/ai-index-benchmark/remeasure-2026-11-02/). - Benchmark: a language corpus, a large-file corpus and the fact counters in
run.py.docs/ai-index-benchmark/corpus-languages/holds nine files written for it (Perl, Bash, ksh93, POSIX sh, a POM, a Spring context, an XSD, an XSLT), with their expected names inmodelcard/facts-languages.json;modelcard/large.pywrites a reproducible SQL dump, XML catalog and Java enum far beyond a small window, with their facts.run.pytakes--facts(every rule names its counters),--oversize/--max-chunks, a corpus without one of its kinds, and gives a language with a prompt of its own in the profile (file-body-shell,-perl,-xml) a rule;models.jsonhas the two Granite presets of the self-test for the large corpus;--chars-per-tokenreplaces every model's estimate, for data that tokenizes denser than code. - CLI: rules per aggregation phase,
packageFieldGenerationsandprojectFieldGenerations. Top-level keys besidesrcMorph, read by the CLI's package phase (AggregatePackages, and the package step ofAll) and project phase and by the plugin'saggregate-packages/aggregate-project: the package phase takes exactly one rule, the project phase none or one. Before, both phases took the first file rule, soAllsummarized packages with a file prompt and model. examples/build-your-own-jar/srcmorph.yamlis the plugin's self-test profile as a CLI configuration, nearly 1:1: gpt-oss-20b (native MXFP4, 96K window) for every file, package and the project overview, Granite 4.0 H-Tiny (64K window,mapReduce) for ~275 KB .. 1 MB and Granite 4.0 H-1B (16K window,sampleChunkswithmaxChunks: 6) beyond, with the exact whole-file counts of the profile's fact files (Java: types, methods, fields, ...; SQL: INSERT statements and rows, tables, views, indexes, routines, triggers, types; and the shell, Perl, XML and POM counters) on every summary of those kinds, so a sampled summary of a multi-MB SQL dump still states what the data holds. It replaces the one-model, one-placeholder-prompt example; a test compares it withsrcmorph-maven-plugin/srcmorph-selftest.yaml.- The plugin's
srcmorph-selftestprofile carries three model definitions only: gpt-oss-20b (gpt-oss-20B-mxfp4) and the two Granite presets. The other model definitions (Mistral, Llama, Codestral, Qwen, Gemma, Ministral, theEXP-*benchmark keys, the other gpt-oss quantizations and windows) are gone; their measurements stay indocs/ai-index-benchmark. profiles/: the production prompts and fact counters once in the repository, now v3.profiles/v3/holds the prompts (file-body-java,-sql,-shell,-perl,-xml,-fallback,package-body,project-body) and the fact counters (java-facts,sql-facts,shell-facts,perl-facts,xml-facts,pom-facts), which the self-test and the example configurations no longer write out;SqlFactPatternsTestandJavaStructureFactPatternsTestread these files instead of copies. The prompts are v1, which the October 2026 model card measured, with targeted additions (the type's own name; getters and setters named one by one; SQL constraint and index names and row counts per table; constants, configuration variables, helpers and library calls of other files). Three runs each on OpenRouter: GPT-6 Luna 100 % in all three (v1 93.6 to 99.0 %), gpt-oss-20b 99.0 % (v1 96.8 %), search unchanged within its noise (model card, finding 5). v3 adds what gpt-oss-20b still lost on a second corpus: a table's row total over severalINSERTs as an addition, extensions and types by name, every index as its own bullet, the entry pointmain, command-line commands and options by the name a user types, and a script's arguments by their variable; gpt-oss-20b 100 % in three runs on the benchmark corpus and 99.8 % on the second (v2: 99.0 % on both), GPT-6 Luna 100 % on both (finding 6). v1, v2, the June v2 experiment and v1 as the June benchmark ran it are in the history (profiles/README.md). The plugin'ssrcmorph-selftestprofile, the model-card benchmark (run.py --profile <name>, default v3, replacingmodelcard/prompts.json), the examples (which had one-line placeholder prompts) and the plugin's integration-test fixture all read these files by key from the profile directory, so the CI smoke of the user-built jar and the plugin IT read the profile as well (the IT also atemplateFile).openai-compatibleprovider: a model can run on any OpenAI-compatible API. OpenRouter (the defaultbaseUrl), a llama.cppllama-server, jllama'sNativeServeror any server answeringPOST <baseUrl>/chat/completions. Four new per-model settings:provider(overrides the run'sgenerationProviderfor this model, so one run can mix local and remote models),remoteModel,baseUrlandapiKeyEnv(the name of the environment variable holding the key, e.g.OPENROUTER_API_KEY; empty by default, so no key is sent unless a model names its variable and a custom endpoint never receives the OpenRouter key by accident; the key never goes into a configuration, and it is sent only overhttpsor to a loopbackhttpserver, never in cleartext to another machine). The request carries the same prompts as the local provider and only the portable sampling knobs; the answer goes through the same parsing, so the.ai.mdcomes out the same way. Remote models skip the GGUF checks before a run and instead must name anhttp(s)baseUrl. Local stays the default: nothing leaves the machine unless a model saysopenai-compatible.openai-compatibleprovider:stream. A model withstream: trueasks for the answer as Server-Sent Events (withstream_options.include_usage, so the token counts still arrive) and assembles it into the same chat response, so the summary does not change; a stream withoutdata: [DONE]or with anerrorchunk fails the file instead of passing for a short answer. Defaultfalse. Needed for the GitHub Copilot API Gateway VS Code extension, whose non-streaming path adds every VS Code language-model tool to the request and runs the tool calls the model makes; its streaming path does not.- Guide: the models of a GitHub Copilot subscription through the Copilot API Gateway
(docs/copilot-api-gateway.md): setup, the settings that keep
the gateway private, an example configuration (
examples/copilot-api-gateway/srcmorph.yaml), model choice and the limits of the route. The model card (docs/ai-index-benchmark/MODEL-CARD.md) adds four runs streamed through the gateway: Claude Haiku 5.5 (98.2 %, 6.8 s per file), GPT-6 Luna (97.6 %), MAI-Code-1.1-Flash (93.0 %) and Grok 4.7 (76.4 %, its reasoning summary leaks into the answer). Two more runs: Mistral Large 4 on OpenRouter (95.0 %, 32.1 s per file, 0.38 USD; its reasoning cannot be switched off) and Aleph Alpha's Kolibri-1 locally (92.2 %, 61.8 s per file, Q4_K_M with all 50 expert layers in RAM). Nine more on 2026-10-31: Muse Spark 1.3 in Meta's Contributor tier (100 %, the best search; Meta trains on its prompts), Nemotron 3.5 Super VL (97.9 % with reasoning off, 90.4 % with it), Mercury 3 (96.6 % at 2.6 s per file, the fastest), Claude Haiku 5.5 on OpenRouter (93.9 %: OpenRouter enforces the 4096-token limit the Copilot gateway ignores), and locally gpt-oss-20b (95.8 %, its hosted quality, the best local run), Granite 4.2 8B (94.2 %), Gemma 4 26B-A4B (93.2 %) and LFM 2.5 24B-A2B (87.0 %, always reasons). openai-compatibleprovider:requestsPerMinute. At most this many requests to a model's server start within any minute; a request beyond it waits until the oldest of the last minute is a minute old, instead of being answered with429. Models of one server with the same key and limit share the window, across the phases of a run. Default0, no limit. The Copilot API Gateway example sets 25 against the gateway'srateLimitPerMinute30;ListModelscopies it into every template entry.- CLI command
ListModels: asks every OpenAI-compatible server theaiDefinitionsname which models it serves (GET <baseUrl>/models, with the definition's key and the same refusal to send it in cleartext to another machine) and prints a table (id, token limit, OpenRouter's price) and a paste-readyaiDefinitionsYAML block, one entry per model, copied from the server's first definition withcontextSizelowered to a smaller model limit.modelFilterkeeps the ids that contain it. Reads OpenRouter, the Copilot API Gateway, llama.cpp'sllama-serverand jllama's server. In the library:ListModelsEngine,ModelListReport,OpenAiCompatibleAiGenerationProvider.listModels(). - CI tests the provider against three real APIs on every run, each in addition to the others:
jllama's
NativeServer(started in-process byOpenAiCompatibleLiveTest), a llama.cppllama-servercontainer (both on the committed SmolLM2 GGUF), and OpenRouter withopenai/gpt-6-lunaand theOPENROUTER_API_KEYsecret. A check on the Surefire report fails thetestjob when any of them was skipped; only the OpenRouter case may skip, and only in a run without the secret (a pull request from a fork or from Dependabot).
- No more fat jars on the GitHub release. Up to 1.3.0 every release attached 15 signed
srcmorph-cli-<version>-jar-with-dependencies[-<classifier>].jarfiles (a CPU jar for every desktop platform plus one per llama GPU module); they were never on Maven Central. There are too many OS, CPU and GPU combinations to ship one per machine, so from 1.4.0 you build the runnable jar yourself (see Changed). Gone with them: thepackage-bound assembly insrcmorph-cli, itsfat-jar.xmldescriptor, the runtime license collector that only those jars needed (and with it the build's dependency on a bernardladenthin/actions checkout andthird-party-notices/), and in CI the classifier build loop,sign-fatjars.sh,verify-classifier-fatjars.shand its tests..github/release-assets.jsonnames no files beyond the Maven artifacts.
-
The prompts lose what models copied or tripped over (the prompt review of the same round). No prompt quotes an example lead any more ("Calculates VAT for invoices" came back as Granite's lead of an enum and as an invented class in a reduce), and each says not to copy wording from the instructions or to write about them; the closing paragraph says that the user message starts with the file name, not a path; headings are named plainly (
'#', '##', '###'); Java and SQL no longer ask to recount before finalizing (a reasoning model spends its budget on it).map-chunkandreduce-partsdrop their own##headings, say that they replace what the rule's prompt says about the user message, name the sample and half labels and the overlap, keep part summaries short, and merge format-neutrally (the project overview has no lead line or sections): the merged lead is about the whole input, a family stays a family, getters and setters stay apart, the facts line's counts are used without copying the line, a total added up from overlapping parts is "about N", an addition the rule's prompt asks for is written as its total. A reduce over sampled chunks reads a note saying so (AiMapReduceSupport.sampleNote), and states the sample's counts as the sample's. -
A local model's windows are measured in tokens, not estimated.
charsPerTokenis an integer estimate, and source code, SQL and data tokenize denser than it assumes: Granite's 64K and 16K windows overflowed on every chunk of the large corpus's dump atcharsPerToken: 3, and each retry cut a quarter of its chunk away.AiGenerationProvider.promptTokenscounts a prompt as the chat template renders it, with the model's own tokenizer (llamacpp-jni:applyTemplateandencode; a remote API cannot, so it keeps the estimate and the overflow retry). With a count, a source over the estimated window is counted whole: it goes in one call when it fits after all, and is otherwise chunked by its measured density; and every call is counted before it is sent, so a prompt overcontextSize - maxOutputTokensis not sent but its window shrinks by the density (indexer.AiTokenWindow). The plan still estimates, since it loads no model. -
Map and reduce calls say what their input is:
mapPromptKeyandreducePromptKey. Every map and reduce call ofmapReduceandsampleChunkstook the rule's prompt, which describes a whole file: gpt-oss-20b indexed a chunk of INSERT rows of the large corpus's dump as a file and invented three tables, a procedure and a trigger, and Granite 4.0 H-Tiny answered the reduce of the catalog with the prompt's own example, aVatCalculatorclass, repeated underPart 8:..Part 18:until its output budget ended. A rule'smapPromptKeyandreducePromptKeyname prompts appended to its prompt in those calls (AiPromptSupport.composedKey, rendered as the two templates and a blank line between); the profile'smap-chunkandreduce-partssay to index only the chunk, and to merge the parts keeping every name a part states. Every map call's input starts withChunk i of n:(AiMapReduceSupport.chunkLabel). The keys apply only to the two chunked strategies; a prompt key may not contain+. The self-test, the CLI example andrun.pyname them in every chunked rule and the package and project rules; the header's- P:covers them. Every reduce call also reads the exact facts line of the whole input in front of the parts, andreduce-partssays to state its counts as given: gpt-oss-20b spent the whole 4096-token budget of the dump's final reduce reasoning, adding up the row counts of nine parts, and answered nothing. -
The self-test's ~275 KB .. 1 MB tier map-reduces with Granite 4.0 H-Tiny over a 64K window instead of one pass over 384K: in one pass over the large corpus's 870 KB catalog H-Tiny copied the XML into a code fence with invented values (every price 972.00, every stock 0) until its output budget ended. The definition is
granite-4.0-h-tiny-mapreduce(wasgranite-4.0-h-tiny-bigwindow), its rulesbig-java,big-sql,big-xmlandbig-pom(werebig-window-*) withonOversize: mapReduce;models.jsonkeeps the old preset for the earlier measurements. -
The family rule of the prompts names no analysis channel.
file-body-javatold the model to count a family of near-identical members "in the analysis channel", which only gpt-oss has; it now asks for the exact total and the index range only (a clause that also had the model name every member breaking the pattern was dropped again: GPT-6 Luna then reasoned at length over one Java file in three of six runs, against none of three before; Granite 4.0 H-1B had listed 6 000 enum constants one by one until its output budget ended, in five of eleven chunks).file-body-sql,-xmland-fallbackget a family rule for the long runs of the large corpus's dump and catalog: rows, records and numbered constants. Infile-body-sqlmore than five INSERTs into one table are a family too: the number of INSERTs and the total rows, not each count as an addition. -
The package and project rules default to
onOversize: mapReduce. Their input is the summaries below them; with the general defaultfaila package over its window failed on every run and its parent lost the subtree,samplewould drop the children behind its head. A rule that setsonOversizekeeps it. -
A file summary's header records the prompt and facts it was written with:
- P:(the CRC32 of the system prompt) and- N:(of the fact counters), written only when set, so package and project headers and older summaries stay as they are. The run record takes both from the header for an unchanged file, so after a prompt change it shows which summaries the old prompt wrote (AiMdHeader.withHashes,p(),n();AiFactExtractor.factsHashreplaces the engine's own). -
Calibrate prints a YAML
calibration:block for the model's entry underaiDefinitionsof the configuration file, not an XML<calibration>element no configuration reads any more (CalibrationReport.renderPasteBlock()replacesrenderXml()). -
Unset,
modelDirectoryandprofileDirectoryare the configuration file's directory, so a relativemodelPath,templateFileorfactsFileresolves against the file's directory like every other relative path of the file, never against the working directory. -
The SQL fact counters count what their labels say.
INSERT rowscountedINSERT INTOstatements, so a multi-rowINSERTof 28 rows read as 1; it isINSERT statementsnow, and the newVALUES rows (approx)counts the row tuples ofVALUESlists.viewsalso countsMATERIALIZEDandOR ALTERviews,tablesTEMP/UNLOGGEDtables,indexesT-SQL'sCLUSTERED/NONCLUSTERED,functions/proceduresCREATE PROC,OR ALTERand MySQL'sDEFINER=;triggers,typesandALTER TABLEare new. Then, after the forms a research pass found missing:INSERT statementscounts MySQL'sINSERT IGNORE/LOW_PRIORITYandREPLACE INTO, views, routines and triggers Oracle'sEDITIONABLE/FORCE,indexesFULLTEXT/SPATIAL/BITMAP;sequences,extensions,schemas and databasesandpackages and package bodiesare new, and--comments, string literals and/* ... */block comments are ignored in one pass (MySQL's/*! */and Oracle's/*+ */are not, they hold SQL; an unclosed/*comments out the rest of the file). -
Dependencies and build plugins: logback-classic 1.6.6 (test scope), NullAway 0.14.3, maven-surefire-plugin 3.6.1, jacoco-maven-plugin 0.8.16, palantir-java-format 2.103.0 (no formatting change) and
anthropics/claude-code-actionv1.0.264. The core depends onjackson-databindandjackson-dataformat-yaml(with it SnakeYAML) forSrcMorphConfigurationFileLoader; up to 1.3.0 only the CLI did. -
The GPG signing-key preflight runs only when something is published. The two jobs
verify-signing-keyandverify-signing-key-gradleran on every trigger: red on every pull request, because themaven-centralenvironment admits onlymainandv*tags, and on a push tomainthey sat out the environment's wait timer for a run that publishes nothing. They run now on apublish_to_centraldispatch only, still at the start of the run, and both publish jobs need them (no longer an exemption of the release gate). -
examples/windows-fatjar/is nowexamples/build-your-own-jar/, for every platform. Itspom.xmltakessrcmorph-clifrom Maven Central, picks the natives jar of the build machine's OS/arch with an OS-activated profile (-Dnatives=<classifier>overrides it, and the enforcer names the property when no profile matches), adds one GPU module with-Dgpu=<classifier>, and packs everything with the logging defaults intotarget/srcmorph.jar.build.sh/build.batpass their arguments to Maven;run.sh/run.batstart the jar. The eightexamples/run_*scripts now run that jar instead ofsrcmorph-cli/target/*-jar-with-dependencies.jar. The README mentions JBang as a shortcut (srcmorph-cli's jar names its main class); Maven is the supported path. -
CI:
smoke-fatjaris nowsmoke-jarand tests the user's path. It installs the reactor, buildsexamples/build-your-own-jaragainst it (srcmorph.versionandllama.versionread from the reactor), checks the Java 8 bytecode floor over that jar and the plugin jar, and runsPlan. The example pom is therefore under test on every pull request.
-
The plugin knows
${project.basedir}and the other project values. A${name}in the configuration file was a user, project or system property only, so${project.basedir}stayed as it was and became a directory of that name.basedir,project.basedir,project.build.directory,project.build.sourceDirectory,project.groupId,project.artifactId,project.versionandproject.nameare known now, and a${name}that no property sets fails the load when it stands in a directory,runRecordFileor a subtree (CLI and plugin alike). -
Nested subtrees no longer fail the run. A subtree inside another one, or listed twice, collected its files twice, and the run failed with "X and X would both be summarized"; it is left out now, and every file is collected once. The plugin no longer takes a resource directory that holds the source directory (a POM that packs files of the project directory itself) as a subtree: it would index the whole project, build output included.
-
Javadoc and messages name the configuration file's keys, not the removed XML. Comments and error messages no longer speak of
<fieldGenerations>,<aiDefinitions>or the plugin configuration; they name the keys of the one configuration file (fieldGenerations,packageFieldGenerations,projectFieldGenerations,factDefinitions).maxChunksis documented as the chunk count ofsampleChunks(required there,0otherwise;mapReducereads every chunk), not as "0 = unbounded"; theonOversizeandonIncompletelists namesampleChunksandkeep;SrcMorphExceptionsays what it is thrown for. Changed messages: "Invalid configuration:", "No fieldGenerations configured (srcMorph.fieldGenerations of the configuration file)", "add a rule with fallback: true". -
Fact counters: a shift no longer opens a heredoc, and a long Java enum counts in linear time. In shell, a
<<inside(( )),$(( ))or$[ ], or whose line goes on with an unclosed quote, opens nothing; in Perl, a<<after a variable, a number,)or], in a trailing comment or in a string opens nothing (print $fh <<EOTand<< "EOT"are heredocs). An unclosed heredoc runs to the end of the file, as bash reads it. 20 000 such lines took 25 to 207 s in shell and 95 to 223 s in Perl, now at most 0.2 s.java-factsreads a parameter list possessively, one level of parentheses deep: the 6 000 constants of the large corpus's enum took 8.2 s inconstructors, now 0.1 s for the whole file.type declarationsanchors with^[ \t]*(40 000 blank lines took 65 s, now 0.02 s) and counts private and protected nested types and@interfaces. No count the tests pin changed. -
An ignored region within one line leaves a space. The
[ignore]counters blanked a region to its line breaks only, soFROM t/* x */JOIN ureadtJOINand counted noJOIN; a region within one line now leaves one space. No count the tests pin changed. -
The deterministic summary shows the real last lines. For a source ending with a line break, "Last N lines" showed N-1 lines and an empty one, and "Content" an empty last line.
-
A cut-off answer kept by
onIncomplete: keepsays so. It ends with_(Cut off at the output budget; kept by onIncomplete: keep.)_instead of stopping mid-sentence without a sign. -
What 1.3.0 wrote from an incomplete answer is written again. 1.3.0 wrote the facts line alone for an empty or cut-off file answer, and a project index without its
#### Overviewfor an empty overview answer, under the checksum of a complete one, so no later run touched them. A summary whose body is the facts line and nothing else now counts as missing, and a project index without its overview is written again when an overview rule is configured. -
The plan sizes a map-reduce's chunks as the run does: next to the longer of the rule's map and reduce prompts with a chunk's label reserve, not next to the rule's prompt alone, so its call count and time follow the run (
AiRouteEstimator.estimateRouteSecondswith a chunk prompt length). -
Benchmark: cut-off answers and unwritten files are counted, numbers with separators found.
run.pylooked for a provider line srcmorph no longer writes, so every meta'scutOffwas empty; it reads srcmorph's warnings and the run record's finish reasons now. A file phase that leaves some files unwritten (exit 1) no longer skips the package and project phases, which made one lost file cost a run 14 points of its format share.score.pycounts20,000and20 000as 20000 and a number before a sentence's period, and scores the large corpus's middle names (middleFound, a name or the prefix of one:LEDGER_LOCKinLEDGER_LOCK_02701). A relative--work/--out/--metaresolves against the working directory, and a NaN--chars-per-tokenis refused. -
The map-reduce drops, cuts and resends nothing behind the summary's back (seven reviews of the branch). When no two part summaries fitted one reduce batch, all of them went into one call cut from the head, and the final answer read 2 of 22 parts under a coverage line that said 22 of 22; now each part is shortened at a line end so two fit, and the line says how many (
3 part summaries shortened to fit the reduce). A chunk denser than its window was sized for is read in two halves instead of losing its tail, split at the line end nearest the middle that leaves each half a quarter of the chunk (not after a blank line at its very start). A window that leaves fewer than 1000 characters for a chunk next to the map and reduce prompts fails the file with that number, instead of cutting every chunk. The answer to a single chunk is the final one and has to be complete. A part cut at the output budget is kept at once, without the line the cut ended in and marked with...so the reduce sees it stops early, also underretryOnce, and counted underkeeptoo; a call cut to a smaller window counts once. In a round that is not the last, a batch of one part passes it on unchanged. A facts line longer than a quarter of a batch is cut for the reduce at a counter. A trimmed prompt is never longer than its window (the EOF marker took room from it, and a line break at the cut kept one character more); the token pre-check aims 2 % below the limit, leaves room for the 4 beginning and end tokens a count does not see, has its own budget of four smaller windows and then refuses the prompt instead of sending it. The density is measured on a head of four estimated windows, not on the whole file. A chunk's overlap starts at a line start. Window arithmetic no longer overflows above 28 M characters. A blankmapPromptKey/reducePromptKeyis unset, and the prompt hash tags the map and reduce prompts by role. -
A request the server may already be billing is never sent twice. A read timeout while the answer body was read still resent the POST, and the JDK itself resends a buffered POST once after a connection reset (
sun.net.http.retryPost); a POST now goes in streaming mode with a fixed length, and a 401, whose body that mode cannot read, says that the key was refused or none was sent. A timeout that ends the retries is the file's failure, not an interrupt of the run (aSocketTimeoutExceptionis anInterruptedIOException; the run ends only when the thread is interrupted). Gemini's per-minute 429 waits and retries (its message reads "exceeded your current quota" too; only the daily quota is final), and aRetry-Aftergiven as an HTTP date is honoured, in all three forms RFC 9110 has a recipient accept (IMF-fixdate, rfc850-date with its two-digit year, asctime-date). A source with a NUL character is not counted by the local tokenizer, which would stop at it. -
A run leaves the configuration as it was given.
check()followed byexecute(), or twoexecute()calls, failed on a rule withfactsKey("names its facts twice"), since the run resolved the facts into the caller's rules; it works on copies now. A map or reduce prompt without a template fails before a model loads, not in every file of the rule after it. An empty list item in the configuration file (-in YAML) is refused naming the list instead of reaching the run as a null.planOnlystops the package and project phases too, so the plugin's aggregate goals load no model under it.SourceFileIndexer.indexFilewithout the hashes is gone (it wrote headers withoutPandN), andRunRecordkeeps no value equality over its growing list of entries (CodeQL). -
A part summary cut at the output budget no longer fails its file. A small model that lists a long run of entries one by one (Granite 4.0 H-1B on a 6 000-constant enum: every constant as its own line until the 1536 tokens ran out) had its map answer refused as incomplete, and the whole file failed at the first such chunk. A part summary is an intermediate the reduce reads, so a cut one is kept now, with a WARN, and the coverage line counts it (
5 part summaries cut at the output budget); the final answer still has to be complete. -
A part that comes back without text is read again in two halves. A reasoning model can spend the whole output budget thinking about one chunk and answer nothing: hosted gpt-oss-20b did on chunks of the large corpus's 1.6 MB SQL dump (4093 of 4096 tokens reasoning, on fp4 and bf16 providers alike, and more often with a larger budget), and the whole file failed. A map answer that is empty or cut before any text now has its chunk read again as two halves at a line end near the middle, as a chunk too dense for its window is; a piece below twice the minimum chunk still fails the file. Measured: the dump is written in three of three runs, 93.0 % against 75.2 % on the large corpus (
docs/ai-index-benchmark/remeasure-2026-11-02/). -
A read timeout of a request is not sent again. The
openai-compatibleprovider resent aPOSTthat got no answer within 600 s up tomaxRetriestimes, each one possibly generated and billed; it now fails the file after the first (a connect timeout, before anything was sent, is still retried). A501, a505and a429whose body says the quota or credit is used up are no longer retried; aRetry-Afterof any number of seconds waits (capped at 300 s) instead of falling back to the short backoff; llama-server's "exceeds the available context size" and Gemini's "input token count ... exceeds the maximum number of tokens" count as a context overflow. -
An edit of a file that is not UTF-8 is noticed. The checksum was taken over the decoded text, so two Latin-1 files that differ only in an umlaut decoded alike and the edit counted as unchanged; it is taken over the bytes with only the line endings normalized (for valid UTF-8 the same value as before, so no summary is regenerated by the change).
-
The package and project rules are checked like file rules. Their
onOversize,onIncompleteandmaxChunkswere never validated (amaxChunksundermapReducewas dropped silently, a typo failed after the model had loaded), and theirfactsKeywas never resolved although the overview honours its rule's facts;checkand the run now refuse such a rule, and askiprule, naming the phase, and resolve the facts. The rule is checked before the missing-output early return, so a wrong one fails on a fresh tree too. -
A rule with both
factsandfactsKeyis refused instead of losing the inline facts. -
An unknown provider name fails before the run, in
checkas well, instead of at its model's group after the earlier groups have generated. -
AllwithplanOnly: trueonly plans: the package and project phases no longer run. -
A key set to nothing keeps its default.
srcMorph:with no value ended in a bareNullPointerException, andpluginVersion:putnullinto a non-null field; both mappers now skip null values (Nulls.SKIP). -
The echoed configuration reads back.
getOversizeStrategy()was a bean getter, so the echo carriedoversizeStrategy, which the strict reader refuses; it isoversizeStrategy()now, likeincompleteStrategy(). -
One failed file never ends the run, and the run record survives an abort. The per-file catch took
IOExceptiononly, while the llama.cpp binding throws unchecked exceptions, so one native error ended the whole run; an unchecked error now fails its file like a failed request (an interrupt still ends the run), and the log names the exception's class. The run record is written however the run ends; a failed write is added to the run's own failure instead of replacing it. -
A line longer than the chunk budget no longer multiplies the chunks. Behind such a line (a dump with one
INSERTper line, minified code) the chunker cut at the previous line end again and moved on by one character, so a 50-line header before an 80 000-character line made 504 chunks, 499 of them the same overlap, each one a model call; it now cuts such a line hard (6 chunks). -
A context overflow no longer cuts a source behind the rule's back. A source that fits by the character estimate but not in tokens was cut by the overflow retry and written as a complete summary, also under
onOversize: fail. Now the rule'sonOversizedecides, as for an oversized file:failrefuses it,samplesays in its coverage line how much of the head the last call read,mapReduce/sampleChunks/deterministictake their path. A map-reduce whose calls had to be cut says so too:**Coverage:** 9 of 9 chunks summarized (2 calls cut to fit the window)(AiMapReduceSupport.chunkCoverage(summarized, total, cutCalls, cutParts, shortened)). -
The benchmark's
score.pytook the facts line for the model's text. With--factsa summary starts with srcmorph's exact counts, so the lead line was reported missing and the counts were scored as the model's. Every measure is now taken on the body without that line; the model card's runs had no facts lines, their scores do not change. -
The self-test's ~275 KB .. 1 MB rules (then
big-window-java/big-window-sql, nowbig-java/big-sql) wrote no facts line, although the smaller and the larger files had one; they namejava-facts/sql-factsnow, insrcmorph-maven-plugin/srcmorph-selftest.yamland inexamples/build-your-own-jar/srcmorph.yaml. -
A local model reasoned with thinking switched off. The llama.cpp provider sent
reasoningEffort(defaultlow, meant for gpt-oss) to every chat template as thereasoning_effortkwarg, also withchatTemplateEnableThinking: false. A template that reads it first, as Aleph Alpha's Kolibri-1 does, then reasoned atlowand spent the output budget on it (the model-card run's first file came back empty). Thinking off is now sent the way llama.cpp's server handles it,enable_thinkingfalseand noreasoning_effort, reached through either knob: withchatTemplateEnableThinking: falsethe effort is not sent, andreasoningEffort: nonesendsenable_thinkingfalseinstead of the effort (before, only the template sawnone, and llama.cpp itself took thinking for on). gpt-oss, which has no thinking switch, keeps its effort. -
A summary that quotes a thinking-block marker lost most of its text.
AiCompletionParserdropped everything up to the last<channel|>anywhere in a response, so a model describing a Gemma-4 parser, a prompt or anything else that names the marker had its answer cut down to the tail (seen onAiCompletionParser.javaitself in a run against OpenRouter: lead, purpose and type were gone). A thinking block is now recognized only at the start of the response, where the model emits it, and closes at the first<channel|>; a marker later in the text stays part of the answer, and a quoted<|channel>thoughtno longer reports an exhausted token budget.
1.3.0 - 2026-10-26
-
examples/windows-fatjar/: build your own fat jar, then run it. Four files:pom.xmlnames srcmorph-cli and the llama classifiers of the machine, withllama-platformexcluded;build.batis onemvn packageon a JDK 8;srcmorph.yamlis a commented configuration that starts withPlan;run.batruns it.
docs/windows-intel-core-ultra.mdwalks through them for an Intel Core Ultra (CPU and Vulkan classifiers, checking what loaded, threads). Tested on Windows: built on JDK 8 (32 MiB instead of the release fat jar's 111 MiB),Planon JDK 8, and a real index with the Vulkan module throughrun.bat.ExamplesConfigBindingTestbinds its configuration and checks the pom's main class.
- The default thread count is now 4, or fewer on a machine with fewer cores. It was 8 for a model
definition (
threads) and 2 for the fallback parameters (llamaThreads,srcmorph.llama.threads). Both now useAiGenerationConfig.DEFAULT_THREADS = min(4, cores). Eight threads on a four-core machine is not a little slower but orders of magnitude slower, because ggml's threads wait on each other at every step: on four visible cores the same real-model test class took 644.9 s with 8 threads and 7.8 s with 4. A configuration that setsthreadsexplicitly is unchanged. On a machine with more cores, raise it after measuring (Calibrate). bernardladenthin/actionspinned at v1.1.0. v1.1.0 addsjava/audit-publication, an audit of a published release from the outside (not used here yet); nothing changes for the actions in use. v1.0.3 before it made the snapshot publication survive GitHub: the Central snapshot transfer lists the jobs of older runs 50 per page and repeats GitHub reads on a server error, and a refused move of thesnapshottag deletes and re-creates it (both failed in java-llama.cpp's first snapshot dispatch on v1.0.1). v1.0.2 before it carried a documentation change only: the license collector'spython3 -cguard must use single-quoted Python string literals, becauseexec-maven-plugindrops a double quote out of an argument on Windows. It records the trap this repository's own POM just stepped out of (see the Fixed entry); the shared CI behaviour is unchanged.- The shared CI code moved to bernardladenthin/actions.
The build-check library, the publication gate and its consumers, the signing-key checks and the
shared scripts and workflows are actions and reusable workflows there, pinned by commit; the
shared-filesjob and.github/shared-files.sha256are replaced by therepository-checksjob, configured in.github/actions-config.json: the files this repository must carry are compared with the templates there, and the publication wiring and a dry run of the release asset contract are checked in every pull request. Packaging srcmorph-cli runs the license collector from there:ACTIONS_DIRin CI, locally a checkout at../actions. - The CLI fat jar carries a license text for every runtime jar, and the build checks it. Runtime
jars without legal files of their own (JSpecify, SnakeYAML) get their upstream texts from
third-party-notices/; a runtime jar with neither fails the build. The dependencies' legal files go into the fat jar only, no longer also into the publishedsrcmorph-clijar. - Releases and snapshots are published through one local gate. The publish jobs stage and sign
everything,
publication.py preparechecks the exact file set, every signature and the version against the tag (v<version>, no-SNAPSHOT;check-tag/check-snapshotcheck it first), and freezes the result; the GitHub release, the Maven Central transfer and the provenance attestation are independent jobs on that frozen package, so a Central outage no longer hides the GitHub files and nothing is attached before the gate has passed. Builds run on JDK 25 (.java-version); the published bytecode target is unchanged. Every published jar carriesLICENSEandTHIRD_PARTY.mdunderMETA-INF/third-party/.
modelDirectory: relative model paths resolve against one directory, absolute ones stay as they are. A new configuration field (plugin parametermodelDirectory, propertysrcmorph.modelDirectory;"modelDirectory"in a CLI configuration; a relative value is relative to the module's${basedir}in Maven and to the working directory in the CLI) is the directory every relativemodelPathof theaiDefinitionsand thellamaModelPathfallback is resolved against. An absolutemodelPath(on Windows also a root- or drive-relative one) is used unchanged, so one model can live elsewhere while the others are named by file name only. Unset, a relative path stays relative to the working directory, as before. Thesrcmorph-selftestprofile uses it:<modelDirectory>${ai.modelDir}</modelDirectory>once, and its 36<modelPath>entries are bare file names; before, each entry was the text${ai.modelDir}/<file>, which put every model in that one directory unless the entry was edited.
- PIT ran for hours on CI: the real-model tests used eight threads on a four-vCPU runner. The first
mainrun to pass its tests since llama 6.0.0 (40316163607) sat in the PIT step for over two hours until it was cancelled. Two tests built their configuration withthreads(8). ggml synchronises its threads on spinning barriers, so more threads than cores is not a little slower but orders of magnitude slower: one generation took 431 s on the runner, PIT's coverage pass 904 s against 41 s locally, and PIT re-runs these tests for every mutant they cover. All real-model tests now useNativeLlamaAvailability.THREADS, the new default (see Changed). The PIT step has a 60-minute timeout, so a regression cannot burn the six-hour job limit again. - The one PIT mutant without a test:
AiModelDefinitionSupport.hasRootreturnstruefor a path the platform cannot parse, and nothing tested that. A model path with a NUL character now does; it stays unchanged instead of being joined belowmodelDirectory. That bringssrcmorphback from 842/843 to the 100 % threshold. - Stale prompt-template documentation. The plugin README described a template as taking two
%splaceholders, and its examples, like the README Quickstart, ended inFile: %s/Source: %s. Since the core extraction (b141d20) a template is the system prompt, sent verbatim, and the file name and source go to the model as a separate user message. A%stherefore reached the model literally. The text now says so, and the examples carry no placeholders. - The plugin README's GPU section described the llama natives before 6.0.0. It said a GPU natives
jar holds
jllama.dll, thatlib.pathpoints at "the GPUjllama.dll", and that a GPU library whose runtime is missing fails over to the CPU one. Since 6.0.0 a GPU jar holds one ggml backend module that is put next to the CPU library, andlib.pathnames a natives directory. The section now describes that, addscublasLt64_13.dllto the CUDA runtime list, and links the Windows guide. - Two CI defects that the unresolvable llama 6.0.0 had hidden. Until 6.0.0 reached Central, every
run stopped at the first Maven step. The first run that resolved it (srcmorph#243) went further and
found two more:
- The
Testjob runsverify, which packagessrcmorph-cliand runs its license collector from bernardladenthin/actions. Unlikebuild, the job had nocommon/setup, soACTIONS_DIRwas unset and the collector stopped the build. The job now runscommon/setuptoo. - The plugin integration test compares the descriptor's
srcmorph.*properties both ways against a deliberate list.srcmorph.modelDirectory, added with the model-directory resolution, was never added to that list. It is listed now, and all six steps pass locally (20 properties).
- The
srcmorph-clicould not be built on Windows at all. Thecollect-runtime-licensesexecution passes a one-line Python program topython3 -c, andexec-maven-plugindrops the double quotes out of that argument on Windows:find_spec("actionlib")reached Python asfind_spec(actionlib), so everymvn package/installdied withSyntaxError: '(' was never closedbefore the module was packaged. The string literals are single-quoted now, which is valid Python and survives the argument handling on both platforms. Measured with this plugin version:print("x")arrives asprint(x)and fails,print('x')arrives intact. Escaping as\"would fix Windows and break Linux, where the argument reachesexecveverbatim, so it is not an option; a comment at the execution says so. The Linux CI never showed this.- The
calibrategoal's paste block mixed line endings on Windows.CalibrationReportwrote its three value lines withString.format("...%n"), the platform line separator, between lines ending in\n, so on Windows the<calibration>block meant for apom.xmlcarried 3x CRLF and 3x LF. All lines end in\nnow.CalibrationReportTestrenders the block in a child JVM started with-Dline.separator=\r\n, so the Linux CI sees what only a Windows build showed before (red without the fix). - ...and that fix left
spotbugs:checkred. SpotBugs reportsVA_FORMAT_STRING_USES_NEWLINEfor everyString.formatpattern containing\nand asks for%n— the platform separator this block must not carry, so the rule and the requirement are in direct conflict. Nothing could have shown it: the reactor still died earlier on the then unresolvablenet.ladenthin:llama:6.0.0, so no CI run reached SpotBugs, andmainwas red on a check that never executed.appendPasteBlockappends the newline as a literal through a newappendXmlNumberhelper now — the shapeappendJsonNumber/appendYamlNumberalready use for the same reason — and formats the numbers at the call site through theFORMAT_*constants the JSON and YAML renderers share. No pattern carries a newline, the block keeps its LF endings, and no suppression was added.
-
The
srcmorph-selftestprofile names its model directory once, asai.modelDir. All 36<modelPath>entries ofsrcmorph-maven-plugin/pom.xmlcarried the absoluteX:/Modelle/..., a drive that no longer exists on the owner's machine, so the profile could not run anywhere as committed — and a machine-specific absolute path sat in a published POM. They read${ai.modelDir}/<file>.ggufnow, the property defaults to the owner's model drive, and another machine runs the profile with-Dai.modelDir=/path/to/models.docs/ai-index-benchmark/tools/calibrate-bench-pom.xmlkeeps its single overridablebench.modeland only follows the drive. -
LlamaCppJniAiGenerationProviderno longer explainsgpuLayers = -1in terms of a "CPU build" and a "GPU build". That was jllama's pre-6.0.0 artifact layout, where the classifier chose the backend. Since 6.0.0 the library jar is CPU-only and a GPU backend joins it as an additive module jar, so the comment's premise was gone while its conclusion stayed right:-1leaves llama.cpp's own default, which offloads everything once a GPU device is registered. Measured on an RTX 3070 Laptop (Qwen3-4B-Instruct-2507-Q4_K_M, identical config, only the classpath differing): 63 t/s prefill and 7.1 t/s decode without the module jar, 3474 / 86.8 with it, andgpuLayers-1 vs 99 within noise either way. -
Removed four unused constants and one string concatenation from
LlamaCppJniAiGenerationProviderTest(six Error ProneUnusedVariable/StringConcatToTextBlockwarnings). The wire-key assertions they belonged to moved to the model-freeLlamaCppJniProviderSupportTestwith the core-module extraction, where the same four constants are declared and used; these were fossils, so no coverage is lost. -
Build tooling bumped in step with the sibling repositories:
anthropics/claude-code-actionv1.0.258 andgithub/codeql-actionv4.39.1. maven-clean-plugin 3.5.0, maven-resources-plugin 3.5.0, maven-install-plugin 3.2.0, maven-deploy-plugin 3.2.0 and maven-dependency-plugin 3.11.0 are now declared, in a new pluginManagement of the parent POM: no POM named them, so their versions came from whichever Maven ran the build. Maven 3.9 runs clean 3.2.0, resources 3.3.1, install and deploy 3.1.x and dependency 3.7.0; the Ubuntu runners' Maven 3.10.0 already matches, except that it leaves the dependency plugin to the newest release. Resource output is byte-identical under resources 3.3.1 and 3.5.0. -
Dependencies and build tooling bumped in step with the sibling repositories: Jackson 2.22.3 -> 2.22.4 (the CLI's JSON/YAML configuration binding) and Gradle 9.8.2 in the shared
verify-signing-key-gradlejob. -
Build tooling: spotbugs-maven-plugin 4.10.4.1 -> 4.10.5.0 in all three modules, in step with the sibling repositories, and
checker-qual4.3.1 managed in the parent (it had stayed at 4.3.0 next to the 4.3.1 Checker Framework processor). With the SpotBugs bump, Lombok-generated code carries JSpecify nullness (lombok.addNullAnnotations = jspecifyin the sharedlombok.config): under@NullMarkeda generatedequals(Object)parameter read as non-null whileObject.equalsaccepts null, which SpotBugs 4.10.5 -- the first release that reads JSpecify -- reports asNP_METHOD_PARAMETER_TIGHTENS_ANNOTATIONper class; Lombok now writes@Nullablethere (and@NonNullwhere a generated member cannot return or take null), so a consumer's nullness checker sees the real contract too. -
Build and test tooling bumped in step with the sibling repositories: Spotless 3.10.4 and Checker Framework 4.3.1 in all three modules. The shared
check-versions.pynow also compares the Kotlin Gradle plugins of a repository with its POMs' Kotlin (nothing to compare here) and reads java-llama.cpp'sexamples/server/pom.xml. -
CI prints the runner's CPU at the top of
testandvmlens(.github/print-host-info.sh, a shared file of the four repositories), so the knob sweep's and PIT's timings can be read against the hardware the job drew. -
CI: snapshot deploy fails with HTTP 401 on Maven 3.10 (seen in java-llama.cpp). The runners moved to Maven 3.10, which sends a
<server>'s credentials only to the origins declared for it; for the idcentralthat ishttps://repo.maven.apache.org, so the upload tocentral.sonatype.com/repository/maven-snapshots/goes out without credentials ("Not using credentials of server 'central'"). Everydeploystep now passes-Dmaven.repository.credentialScope=id, Maven's own switch back to id-only matching, until a setup-java release can write<repositoryOrigins>(mvn-server-repository-origins, merged upstream but not yet released). -
CI: shared files and the release gate are checked. The files kept byte-identical with the sibling repositories are listed with their SHA-256 in
.github/shared-files.sha256; a newshared-filesjob fails on a copy changed here alone and warns on a sibling's differing copy. The same job runs the shared build-check library's tests andcheck-release-gate.py: every job must gate both publish jobs unless.github/release-gate-exemptions.txtsays why (vmlensnow gates). The crash-log step and the signing-key preflight are shared scripts (print-crash-logs.sh,verify-signing-key.sh) instead of copies pasted into the workflow. -
Workflow jobs kept identical across the repositories are checked too: a
.github/shared-files.sha256entry.github/workflows/publish.yml#<job>hashes one job (startgate,shared-files,verify-signing-key,check-snapshot,check-tag, and where presentverify-signing-key-gradle,github-snapshot,github-release). -
Own scripts are licensed
MIT OR Apache-2.0like the CI files:examples/run_*anddocs/ai-index-benchmark/tools/(the Java sources and the rest of the project stay Apache-2.0; the headergenerate-fixture.shwrites into its generated Java fixtures is unchanged). -
More shared files, and files identical up to the repository name: a shared-files entry ending in
?repois hashed with the repository's name replaced by{repo}. Added:.editorconfig,.gitattributes(now with*.gguf binaryeverywhere),FUNDING.yml,CODEOWNERS, the license texts,SUPPORT.md,ISSUE_TEMPLATE/config.ymland further files listed in.github/shared-files.sha256; the signing self-test now runs on Gradle 9.8.0 in all four repositories..mvn/settings.xmllost the unusedgithub-java-llamaserver. -
The JDK is named once, in
.java-version: every workflow reads it through setup-java'sjava-version-file(theJAVA_VERSIONenv and the literal21s are gone);.java-versionandcodeql.ymlare now byte-identical in all four sibling repositories and in the shared-files manifest. -
CI files are licensed
MIT OR Apache-2.0: every.githubfile carrying only the owner's copyright now has the same license header in all four sibling repositories, so the shared ones are byte-identical.claude.yml,claude-code-review.yml,scorecard.yml,reuse.yml,osv-scanner.yml,dependabot.ymlandCODE_OF_CONDUCT.mdjoined the shared-files manifest;osv-scanner.ymlnow grantscontents: readinstead ofread-all(as java-llama.cpp already did). -
Workflow run scripts are parsed in the
shared-filesjob:check-run-scripts.pyrunsbash -nover everyrun:script of the workflows and composite actions that runs in bash (shell decided as the runner does), so a broken script fails within minutes instead of in the job that runs it. -
Maven versions are compared with the sibling repositories:
check-versions.py(in theshared-filesjob) warns where a dependency or plugin -- incl. annotation-processor paths and the Spotless formatter version -- is used in another version than in a sibling's default branch. -
net.ladenthin:llama6.0.0's natives jars. The binding is now the Java classes only, and every native build is its own jar of the same artifact (classifier<backend>-<os>-<arch>).srcmorphdepends onnet.ladenthin:llama-platform(a pom naming the CPU natives of every desktop platform), so the plugin and the CLI keep running everywhere with no change for users. A GPU natives jar is now added next to the CPU natives rather than swapped in:-P gpu-cuda/-P gpu-vulkanand the new-Dllama.classifier=<classifier>(profilegpu-natives) add one, and the loader falls back to the CPU natives when the GPU runtime is missing. The GPU classifier fat jars of the CLI are therefore the CPU fat jar plus one backend — they now also run on a machine without that GPU..github/verify-classifier-fatjars.shasserts exactly that (the promised<OS>/<ARCH>/<backend>/directory, and no CPU native of the default jar lost), and the CI fat-jar set follows llama'snatives.csv(the twomsvc-windows-*jars added). -
The CLI fat jar now ships
slf4j-simpleinstead of logback, and nochecker-qualat all. Production code here targets Java 8, but every logback release from 1.4.0 on is Java 11 bytecode: SLF4J'sServiceLoaderfindsLogbackServiceProviderat startup, so a Java 8 JVM died withUnsupportedClassVersionErrorbefore a single line was logged. The Java 8 logback line (1.3.x) is end-of-life and its CVEs are fixed only in 1.5.x/1.6.x with no backport, so downgrading was not an option either.slf4j-simpleis six classes from the same release train asslf4j-api, with no configuration or socket layer for a CVE to live in.checker-qualmoved toprovidedscope. Its annotations are major 55 from 4.0.0 on and@Retention(RUNTIME), so anything reflecting over an annotated element — Jackson, which binds the CLI's whole configuration — loads them.providedkeeps 4.2.2 on the compile classpath, where the Checker Framework processor needs it, while shipping none of it:jar-with-dependenciesfilters on scope, which is also why<optional>true</optional>alone would not have been enough.Consequences for users:
examples/logbackConfiguration.xmlis replaced byexamples/simplelogger.properties; the fat jar carries its own defaults (INFO, stdout, timestamps) which any-Dorg.slf4j.simpleLogger.*system property or a classpathsimplelogger.propertiesoverrides. The publishedsrcmorphandsrcmorph-clilibrary jars are unchanged in this respect — they impose no binding and carry no logging configuration; a binding is an application concern and the fat jar is the application.net.ladenthin:llama's own binding is now excluded transitively, so embeddingsrcmorphno longer forces one on the caller and the Maven plugin no longer ends up with two providers alongside Maven'smaven-slf4j-provider.
- CI gates every built jar on the Java 8 class-file floor.
release 8governs only the code we compile; a dependency built for a newer Java lands in the jar untouched and surfaces asUnsupportedClassVersionErroron a consumer's JVM. The new.github/verify-bytecode-version.sh(kept byte-identical across the four sibling repos) runs in thesmoke-fatjarjob with--max-major 52and fails on any class above it, skipping onlymodule-info.classandMETA-INF/versions/**, which a classpath JVM never loads.
-
srcmorph:calibratenow writes a machine-readable report. The goal built aCalibrationReportand printed it asINFOlines, and that was the only output — so the numbers a calibration run produces (prefill / decode throughput, chars per token per model) could not be diffed across runs, committed as a baseline, or fed back intoaiDefinitionsby anything other than a human re-reading the console, which is most of the point of measuring them.execute()now also writessrcmorph-calibration.jsonandsrcmorph-calibration.yamlinto the configuredoutputDirectory, creating it if needed. Both carry the same keys in the same order and addloadSeconds,midPrefillTokensPerSecondandcachedPromptTokens, which the pasteable<calibration>block does not. The three figures the XML does carry are formatted identically, so the JSON and the paste block can never disagree about what was measured.The renderers are hand-rolled in
CalibrationReportrather than delegated to Jackson: the core module is framework-free and carries no JSON dependency (onlysrcmorph-clidoes), andrenderXml()already set that precedent. Model keys come from user configuration, so they are escaped for both formats — an unescaped quote would produce a file neither parser accepts.
-
flashAttnworks. The knob was documented and settable since it was introduced, but could not be forwarded:--flash-attntakes a mandatoryon|off|autoandnet.ladenthin:llamaoffered only a bare-flag setter, which emitted the key alone and made llama.cpp's parser consume the following argv token — the load then died naming a flag the user never set. 1.2.0 shipped a refusal at plan time rather than that diagnostic.net.ladenthin:llama6.0.0 addsModelParameters.setFlashAttn(FlashAttn), so the provider now emits--flash-attn onwhen the knob is set.falsestill emits nothing, which leaves llama.cpp's ownautodefault in force — that is the correct behaviour for an unconfigured run, and it is what the knob's javadoc now says.The knob is exercised against a real model for the first time.
flashAttnmoves out ofLlamaCppJniKnobSweepTest'sNOT_SWEPTlist into a real sweep case; until now only the refusal had ever been executed.
- The plan-time and provider-side refusals (
EngineSupport.validateFlashAttnIsNotRequested,LlamaCppJniAiGenerationProvider.FLASH_ATTN_UNSUPPORTED_MESSAGE) and the three tests that pinned them.
-
lazyModereplacestensorReadLazy— breaking. llama.cpp b10731 renamed--tensor-read-lazyto-lzm/--lazy-modewith no alias, andnet.ladenthin:llamafollowed the rename rather than papering over it, so the model-definition field, itssrcmorph.llama.*surface, theLlamaCppJniConfigaccessor and the sweep case move with it. Carrying the old name would leave the configuration describing a flag that no longer exists.Worth knowing why this is a real hazard and not cosmetics: a new Java name paired with an old native produces
Failed to parse model parametersat load time and nothing earlier catches it. The knob sweep did — it is what surfaced the mismatch here. -
net.ladenthin:llama5.1.0 → 6.0.0 (the pin said 5.2.0 until 2026-10-15; jllama never published that number and ships the breaking natives split as 6.0.0). Deliberately the release version, not6.0.0-SNAPSHOT: this repo declares no<repositories>element (thedistributionManagement/snapshotRepositoryis deploy-only), so the Central snapshot channel is not resolvable here at all — and a-SNAPSHOTdependency onmainwould block every srcmorph release, since Central rejects one. The build therefore stays red untilnet.ladenthin:llama:6.0.0is published and goes green with no further edit the moment it is — recorded here so nobody mistakes it for a regression. -
BREAKING:
chatTemplateEnableThinkingis a tri-state; theenable_thinkingkwarg is only sent when it was actually set. It used to be a plainbooleandefaulting totrue, so every run putenable_thinkinginto the chat-template kwargs -- including runs whose chat template has never heard of the kwarg, which llama.cpp's Jinja layer has been moving from "silently ignored" toward "warned about". The only way to stop that noise was to set the knob tofalse, which means something else entirely.The obvious phrasing -- "send it only when it differs from the template's default" -- is not implementable: srcmorph would have to parse and evaluate the template to know that default, which is exactly the work it delegates to the binding. So the rule is "send it only when the user actually set it".
AiGenerationConfig,AiModelDefinitionandLlamaCppJniConfig(and its builder) now carry@Nullable Booleaninstead ofboolean,DEFAULT_CHAT_TEMPLATE_ENABLE_THINKINGisnull(unset) instead oftrue, and the twoisChatTemplateEnableThinking()getters were renamed togetChatTemplateEnableThinking()to match the wrapper type.trueandfalseare both still forwarded verbatim -- unset is the only value that omits the kwarg. A configuration that never mentioned the knob keeps working and simply stops sending it; one that set it explicitly is unaffected.Guarded by four model-free tests over the extracted
buildChatTemplateKwargs()-- including thetruecase, so a future "send it only when false" shortcut fails rather than silently swallowing a configured value.
- CI: a fat jar without any native library is named in the release check again.
.github/verify-classifier-fatjars.shruns underset -euo pipefail, and itsnatives_ofhelper pipedunzip -Z1throughgrep: for a jar with no native entrygrepexits 1, the assignment of the helper's output failed and the script ended right there -- red, but with no::error::line and no jar name, so its two "carries no native library at all" checks (default jar and classifier jar) could never run. The helper now tolerates only grep's "no match"; anunzipfailure or a grep error still fails it. The script had no test at all: the new srcmorph-local.github/buildcheck/tests/test_verify_classifier_fatjars.pybuilds fixture jars and asserts exit code and::error::message for every check the script makes itself (a matching classifier jar passes; CPU-only under a GPU name, a lost default native, an unmapped classifier shape, a jar without natives -- classifier and default -- and a single-OS default jar fail), and for a corrupt jar unzip's own exit code 9 without the script's diagnosis (a helper whose failure were masked,|| trueon its call or a lost pipefail, would report the file as a jar without natives and is caught; a|| trueon the grep branch alone is not observable through a corrupt jar, pipefail carries unzip's status either way). The two no-natives cases were red before the fix. - CI: the GitHub
snapshotpre-release holds exactly the current build, and the release assets include the pom.gh release upload --clobberreplaces same-named assets and never removes any other, so thesnapshotrelease accumulated: all 16 of its assets were stale -- the 0.1.0-SNAPSHOT and 1.0.1-SNAPSHOT jars of the oldllamacpp-ai-index-maven-plugincoordinates and four SpotBugs plugin jars (fb-contrib-7.6.4.jar,fb-contrib-7.7.4.jar,findsecbugs-plugin-1.13.0.jar,findsecbugs-plugin-1.14.0.jar) an old collect step uploaded fromtarget/spotbugs/in 2026-05/06.github-snapshotnow deletes every asset the run did not upload, after the upload and each printed by name; it refuses an empty upload, treats an asset that is already gone as removed and fails only after trying every deletion (asset by asset on purpose: the job's first form deleted and recreated the release, and failed in every burst of main pushes in May 2026); versioned releases are never pruned, and the next snapshot run clears the backlog by itself. The collect steps copy<artifactId>-<version>.pom+.pom.ascas well (Central had them, the release did not), and the signature check counts every attachable type (*.jar,*.aar,*.pom) instead of jars alone. Shared job text with the two sibling repositories (shared-files.sha256), the convention inworkspace/policies/fat-jar-release-assets.md.
1.2.0 - 2026-09-01
-
The plan phase now checks the configuration against the model itself, still without loading it.
net.ladenthin:llamashipsGgufInspector, which parses only a GGUF's header key/value table — no native library, no tensor data — so it is usable from a phase whose whole promise is that it loads no model. srcmorph was not using it; its check wasnew File(modelPath).isFile(), which misses two things, both silently:- The file is not a GGUF at all. A Git LFS pointer, a truncated download, or simply the wrong file passes an existence check and dies much later inside the native loader — in a multi-model run, after the earlier model groups have already generated. Now a plan-time failure naming the file.
contextSizeexceeds what the model declares. This is the one that costs the most and shows the least: every number the plan produces is derived fromcontextSize—maxInputChars, the oversize/chunking decision, the time estimate. Point the default 32768 at a 4096-context model and the plan is wrong by 8× before a token is generated. A warning rather than an error, because llama.cpp will deliberately run past a model's trained context with RoPE scaling.
New
provider.GgufModelInspector(thejniConfinedToProviderrule keeps binding types out ofengine) and the framework-freeprovider.GgufModelInfoit returns, the latter on the PIT gate. -
A truncated summary is now reported instead of being written as if it were complete.
maxOutputTokensdefaults to 128; a model that hits that ceiling stops mid-sentence, and llama.cpp says so — the OpenAI finish reason islengthrather thanstop. The provider was callingchatCompleteText(...), which returns only the text, so the signal was not merely ignored but unavailable. The generate path now parses the whole response, which costs no extra inference:chatCompleteTextis literallyextractChoiceContent(chatComplete(...)), the same native call.Worth recording for whoever touches this next: the check compares against the literal
"length", deliberately not againstStopReason. Those are two different vocabularies —getFinishReason()is OpenAI's (stop/length/tool_calls), whileStopReasonmaps llama.cpp's ownstop_type(eos/word/limit).StopReason.fromStopType("length")returnsNONE: a silent wrong answer, not a compile error. -
Prompt-cache reuse is now visible during an indexing run, not only during
calibrate, closing the follow-up that release left open. Parsing the full response also makesUsageavailable, so theDEBUGline reports cached / total prompt tokens and tokens generated, per file — the run that actually paysswaFull's KV-memory surcharge had no visibility into whether it was paying off. -
srcmorph-cliandsrcmorph-maven-pluginare PIT-gated, at the samemutationThreshold100 as the core. They were the last two modules without one, and the gap was not theoretical: PIT measured the plugin at 53% (33/62) and the CLI at 84% (16/19) before any test was written.The plugin's hole was the interesting one.
AbstractAiIndexMojoTestcovers the sharedbuildConfiguration(), but each goal adds its own mapping step on top — and every one ofbuildGenerateConfiguration/buildAggregatePackagesConfiguration/buildAggregateProjectConfiguration, plus all eightgetLlamaContextSize/getLlamaThreadsaccessors, had no coverage at all. DropsetExcludes(...)fromGenerateMojoand the plugin quietly indexes files the user excluded, with nothing failing.MojoConfigurationMappingTestnow pins each goal's own parameters with values distinct within their type, so a transposition fails rather than cancelling out — verified by swappingminFileSizeBytes/maxFileSizeBytes, which reds the test. The three mapping methods went fromprivateto package-private for it.Two survivors that the new gate exposed were weaknesses in existing tests, not missing ones:
AbstractAiIndexMojoTestassertedgenerationProviderwas"mock", which isSrcMorphConfiguration's own default — so dropping the setter call was invisible; it now uses a non-default provider name. AndCalibrateMojo.execute()'s deliberate blank separator line could not be seen by acontainscheck, so it is asserted by position.The CLI gate excludes exactly one method,
Main.main(String[]), with the reason recorded in the pom: thesmoke-fatjarjob — a release gate for both publish jobs — already runs the realjava -jarartifact and assertsMain#run end., which onlyrun()emits and onlymain()calls. Its one true survivor, theSystem.out.printlnthat makes the<calibration>block paste-ready rather than a log line, is now pinned by a stdout-capturing test.CI runs the goal reactor-wide instead of
-pl srcmorph -am, and the survivor extraction and report upload cover every module'starget/pit-reports. -
Four new model-definition knobs, wired through to
net.ladenthin:llama5.1.0's newModelParameterssetters. They are configured exactly like the existinggpuLayers/mainGpu/devicesknobs — as elements of an<aiDefinition>in the Maven plugin, or as keys undersrcMorph.aiDefinitions[]in a CLI JSON/YAML config — and each is only forwarded to the binding when explicitly set, so an unconfigured build behaves exactly as before:cpuMoeLayers(--n-cpu-moe/-ncmoe, default-1) — keep the MoE expert weights of the first n layers on the CPU. Usually the better trade than loweringgpuLayerson a MoE model: it moves only the expert weights (the class that dominates such a model's size), so a substantially larger model fits the same VRAM at a smaller speed cost.0is a valid, meaningful value, so the forwarding guard is>= 0, not> 0.cpuFfnLayers(--n-cpu-ffn/-ncffn, default-1) — the dense-model counterpart; same>= 0guard for the same reason.kvUnifiedPerSlot(--kv-unified-per-slot, default-1) — per-slot unified KV context cap. The binding rejects0and negatives, so only a positive value is forwarded.tensorReadLazy(--tensor-read-lazy, default empty) —off/auto/on. Shortens model load time, which matters here because a run loads one model per model group andcalibratepreflights every model in turn. Carried as aStringthrough theconfigpackage (thejniConfinedToProviderArchUnit rule keepsnet.ladenthin.llamatypes out of it) and resolved to the binding'sTensorReadLazyModeinside the provider, matched case-insensitively against the CLI strings the enum itself declares. An unrecognised value is rejected, not silently dropped — a typo there would otherwise hand the user a run that quietly did not do what was asked.
-
seedmodel-definition knob (InferenceParameters.withSeed, default-1). Upstream draws a random seed per request, so with a non-zero temperature every generation samples differently: a re-index of an unchanged file, or anyforce=truerun, produced a different.ai.mdbody than the one already committed. That sat awkwardly next to the project's own "deterministic indexing" principle. Setting a seed makes the body stable for a given machine and configuration, which turns a re-index into a reviewable diff. Forwarded only when>= 0, so an unconfigured run is byte-identical to before. Documented deliberately as not bit-reproducibility: llama.cpp results move with thread count, batch size and backend, so the seed pins the sampling, not the arithmetic. -
The prompt cache is now measured instead of assumed —
cache_nreaches Java. The binding has always reported how many leading prompt tokens llama.cpp served straight from the KV cache; the provider read theTimingsobject and dropped that number. It is the only observable the whole prefix-reuse stack produces (cachePrompt,cacheReuse,swaFull), so with it discarded a run paidswaFull's KV-memory surcharge on the assumption that it was working, with no way to tell whether it was.AiGenerationTimingsnow carries it ascachedPromptTokens()next to the evaluatedpromptTokens(), plus their sum astotalPromptTokens();AiCalibrationMeasurementcarries the near-window run's count, andsrcmorph:calibratelogs it per model — anINFOline naming the reused token count, or aWARNwhen nothing was reused, which is the case where those settings cost memory and buy nothing. The real-model provider test asserts that a second request against the same system prompt finds it cached, so the reuse is proven end to end and not just plumbed. -
Seven further model-definition knobs, wired through to
net.ladenthin:llama5.1.0 exactly like the existinggpuLayers/cpuMoeLayersset — an<aiDefinition>element in the Maven plugin, or a key undersrcMorph.aiDefinitions[]in a CLI JSON/YAML config — and each forwarded to the binding only when explicitly set, so an unconfigured build behaves exactly as before. They are added now rather than later because configuration surface is the part of this library that cannot be extended without a release.repeatLastN(--repeat-last-n, default-1) — the gap this closes is a pair that was half wired:repeatPenaltywas configurable, the window it acts on was not, so the strength was adjustable while the reach stayed on llama.cpp's default of 64 tokens.0disables the penalty and is a meaningful value, so the guard is>= 0; the binding rejects negatives. Per-request, next torepeatPenaltyitself.cacheTypeK/cacheTypeV(--cache-type-k/--cache-type-v, default empty) — KV-cache quantization, the most direct trade of quality for context length: atq8_0the cache costs about half off16, so a largercontextSize(orswaFull) fits the same VRAM. Carried as aStringthrough theconfigpackage — thejniConfinedToProviderArchUnit rule keepsnet.ladenthin.llamatypes out of it — and resolved to the binding'sCacheTypeinside the provider, matched case-insensitively against the CLI strings the enum itself declares. An unrecognised value is rejected, and the message names which of the two knobs was wrong, since one resolver serves both.flashAttn(--flash-attn, defaultfalse) — lower KV memory and faster attention where the backend supports it, and the precondition for quantizing the V cache.batchSize/ubatchSize(--batch-size/--ubatch-size, default-1) — prefill sizing, and prefill is what dominates an indexing run: every file is one large prompt with a short answer. The binding's own default for both is0, which llama.cpp reads as "decide for me", so0is not a meaningful user value and the guard is> 0.threadsBatch(--threads-batch, default-1) — prompt-processing threads, separate from the decodethreads. The two optima differ on machines with efficiency cores;-1keeps llama.cpp's behaviour of reusingthreads.
srcmorph:calibrateis what makes these measurable rather than guesswork: it reports prefill and decode throughput, and — since this release — the prompt-cache reuse the extra KV room buys.Note the cost this exposed:
LlamaCppJniConfig's positional constructor reached 37 arguments. It was covered field-by-field byLlamaCppJniConfigFactoryTest, which gives every field a distinct value so a mis-ordered pair fails — but the shape was past its limit, and this same release replaces it with a builder (see Changed below). The constructor is private now; nothing outside the class can call it. -
Fail-fast check on a routed model's
<modelPath>. A typo previously survived the whole plan phase — walk, classify, rendered plan — and only died inside the native loader, which with several model groups means after the earlier groups had already generated. The check joins the fail-fast block the engines already run.Two placement constraints, both load-bearing. It is gated on
generationProvider == llamacpp-jni, because only that provider loads a GGUF and every shipped example deliberately points at a non-existentunused-with-mock-provider.ggufwhile running the mock; an ungated check would red the examples, their binding tests, the CLI end-to-end test and the fat-jar release smoke. And it runs after theplanOnlyearly-out, becauseplanOnlyis documented to "stop before loading any model" andPlanis the CLI's default command — the workflow it serves is configuring routing on a machine where the GGUFs are not present and running for real on the one that has them. Stat'ing a file is not loading it, but failing the plan on a missing model would break that workflow all the same.The ordering was wrong in the first cut and no test saw it: every other
GenerateEngineTestfixture uses themockprovider, so the check never ran through the engine at all. Both halves are now pinned — a plan-only run with the real provider and a missing model still plans; the same configuration withoutplanOnlystill fails fast. -
Every model knob is now exercised against the real model, one knob per test case.
LlamaCppJniKnobSweepTestsets each of the 34 sweepableLlamaCppJniConfigknobs to a non-default value in turn and runs a real generation with the committed 135M model. The failure class it exists for is precisely the one that produced this release's two provider fixes: a knob is not exercised by being set, it is exercised by the binding accepting the value derived from it, and those two came apart twice —dryPenaltyLastNat its own-1default (rejected outright, so every generation threw before a token) andflashAttn(mapped onto a setter that cannot express what llama.cpp now demands). Neither is visible to a mapping unit test, and neither is visible to the rest of the suite, which runs on themockprovider and never loads the native library.Per-knob rather than all-at-once on purpose: an all-knobs config catches the same breakage but names no culprit, and one rejected value masks every knob behind it. 36 cases in ~35 s. Three knobs are excluded with stated reasons (
modelPath,devices,flashAttn), and a reflective completeness check fails the class when a knob is neither swept nor excluded — in both directions, plus a count assertion so an empty sweep list cannot satisfy it. A knob added later therefore reds this test until somebody decides how it is covered. Verified by falsification: a deliberately invalid value reds exactly the case carrying it and names the knob. -
The Maven plugin is now run as a plugin in CI (
plugin-itjob,.github/plugin-it.sh, fixture in.github/plugin-it/). Until now every check onsrcmorph-maven-plugincalled Java methods directly --PluginArchitectureTest,MojoPhaseSkipTest,MojoConfigurationMappingTest, PIT at 62/62 -- so nothing covered the parts only Maven performs: plexus binding of the<configuration>XML onto the@Parameterfields (the CLI's Jackson binding is a different code path), goal-prefix resolution, thesrcmorph.*property strings, lifecycle-phase binding, and the descriptormaven-plugin-plugingenerates. A renamed@Parameterproperty would have shipped green.Six checks: the full lifecycle with all three goals configured entirely from the fixture's pom XML (the nested
<condition><extensions>routing tree, both<subtrees>entries,<excludes>, the per-execution<configuration>override, and both readonly${project.*}injections); the generated descriptor read out of the packaged jar (goal prefix, the four goal names, and a two-way diff of all 19srcmorph.*property strings, so a rename fails and a new property fails until it is listed deliberately);mvn srcmorph:generate -Dsrcmorph.planOnly=truevia the goal prefix;-Dsrcmorph.aiVersion=9.9.9reaching the written document header; all four skip properties including one off-diagonal case (srcmorph.file.skipmust not skip the packages goal -- the diagonal alone cannot catch a copy-pasted property); and thecalibrategoal, the one goal the lifecycle run does not reach. The fixture uses the defaultmockprovider, so no GGUF, no GPU and no network.The fixture's
<configuration>deliberately mirrors the worked example insrcmorph-maven-plugin/README.md, so it also fails when the documented XML stops being the XML that works.Every assertion was falsified rather than assumed, and that pass rewrote three of them. A first version passed a
settings.xmlwhose<pluginGroups>entry was believed necessary formvn srcmorph:generateto resolve; it is not (Maven matches the prefix against the descriptor of the plugin the fixture declares), and adding it opened a second resolution path through repository metadata that masked a changed goal prefix entirely -- on a pristine runner too, since Central serves that metadata. Removing the file turned that check into a real second guard. Separately,<subtrees>,<excludes>and all three execution-level<configuration>blocks could each be deleted with the test staying green -- the first because the fixture's only value equalled the engine's own fallback, the second because the pattern matched no file that existed, the third because the overrides changed nothing observable. All three are now pinned by a fixture that makes them observable. -
The classifier fat jars are verified before they are attached (
.github/verify-classifier-fatjars.sh). The publish jobs build seventeen fat jars — one pernet.ladenthin:llamanative classifier plus the default — and exactly one of them was ever checked:smoke-fatjarlaunches the default. The sixteen GPU jars still cannot be launched meaningfully (a GitHub-hosted runner has no such device, and the only command that loads the native library is a real generation), so the script asserts instead that each is the artifact its name claims: one jar per requested classifier and no unexpected one, a native library present, the native for the OS/arch the name promises, a default jar spanning more than one OS, and — the load-bearing check — a native set that differs from the default jar's. If-Dllama.classifier=ever stops being wired through, Maven resolves the default artifact and the loop ships sixteen copies of the CPU build under GPU names; nothing else in the pipeline would notice. A classifier shape nobody has mapped fails the script rather than passing unchecked. -
A release can no longer attach an unsigned asset quietly. The attach jobs collect
target/*.jar.ascwith|| true, so a signing step that produced nothing yielded an attach that looked complete and was not. The check deliberately does not gate the upload: both attach jobs run even when the publish job failed, because when Central is unreachable the GitHub assets are the only way to get the build output at all. So it reports before the upload, uploads unconditionally, and fails the job afterwards — the assets always land, an unsigned release is loudly red instead of quietly wrong. The same two steps are now byte-identical in all four sibling repositories.
-
Two further public constructors gained a parameter, and a mojo parameter was removed. Neither was listed as breaking when the changes landed; both are, so they are recorded here rather than left for a consumer to discover at compile time.
AiGenerationTimingsandAiCalibrationMeasurementeach went from five parameters to six, both gainingcachedPromptTokens— thecache_nvalue this release surfaces. Anyone constructing either directly must add the argument; both are value types a downstream extension could plausibly build.srcmorph.llama.libraryPathis gone from the plugin (see Removed), and Maven does not ignore an unknown<configuration>element: a consumer POM still declaring it fails the goal outright rather than warning. That is the intended outcome — a parameter that never worked should not keep looking like it does — but it is a hard break on first invocation, not a silent one. CLI users hit the same wall at parse time, because both JSON and YAML are read withFAIL_ON_UNKNOWN_PROPERTIESenabled. -
LlamaCppJniConfigis built through a builder; its positional constructor is private. Breaking for anyone who constructed one directly — deliberately taken in this release rather than later, because the class is already breaking here and a second break for the same type is worse than one.It had grown to a 37-argument constructor, which is past the point where a call site can be read or a swapped pair of same-typed neighbours is noticeable: the compiler accepts
…, batchSize, ubatchSize, …in either order.LlamaCppJniConfig.builder(modelPath)replaces it, andmodelPath— the one value with no meaningful default — is required at the entry point rather than checked inbuild(), so a half-formed builder cannot be constructed at all.The builder is also where the defaults now live, and that removes a real duplication:
LlamaCppJniConfigFactory.fromFallbackParametersused to restate all 30+AiGenerationConfig.DEFAULT_*constants by hand, so a knob added toAiGenerationConfigcould end up with a different default there by simple omission. It is now five lines. Each builder field starts at the matchingAiGenerationConfig.DEFAULT_*, and anullon a String or list setter restores that knob's default instead of storingnull— note this is "the default", not "empty":reasoningEffort(null)gives back"low".Two tests carry this, and both were verified to fail against a deliberately broken build rather than assumed to:
builder_leavesEveryUnsetValueAtItsAiGenerationConfigDefault(every unset value equals its config default — reds when one builder field is initialised to0instead) andbuilder_everySetterWritesItsOwnField(36 distinct values — reds whenubatchSize(…)is made to writebatchSize). The old positional-constructor tests they supersede are removed. -
net.ladenthin:llama5.0.6 → 5.1.0 (llama.cpp b10456 → b10682). Unlike the earlier bumps in this series, 5.1.0 is not purely additive: it deprecates sixInferenceParametersmethods that were always silent no-ops, and this provider called one of them —withUseChatTemplate(true). Becausesrcmorphcompiles with-Xlint:all -Werror, that deprecation is a compile error, so the bump would have broken the build. The call was removed rather than suppressed: it never had an effect (upstream reads no such key from a request body), and llama.cpp defaultsuse_jinjato true, so chat templating and tool calling are unchanged. The rest of the surface this provider uses —LlamaModel,InferenceParameters,ModelParameters,ChatResponse/Timings/Pair,ChatResponseParser,ReasoningFormat— is untouched by 5.1.0. Verified by a full reactorclean test: 639 / 39 / 32 tests, 0 failures, 0 skipped, all four modules SUCCESS. -
CI actions bumped to latest:
actions/setup-javav5 → v6. -
Build tooling bumped and NullAway aligned with the sibling repos:
nullaway0.13.8 → 0.14.0,spotless-maven-plugin3.10.0 → 3.10.1,palantir-java-format2.96.0 → 2.97.0,pitest-maven1.25.9 → 1.30.0. All three of the first are declared per module here (srcmorph,srcmorph-cli,srcmorph-maven-plugin), not in the reactor parent, andpitest-mavenhas two separate version declarations — every one was moved. nullaway is an alignment rather than a plain bump: streambuffer had already merged a Dependabot bump to 0.14.0, so the four repos had silently stopped being identical. Deliberately not taken:jqwik1.9.3 → 1.10.1, forbidden byworkspace/policies/jqwik-prompt-injection.md. -
AiGenerationConfig.getStopStrings()no longer declares a@Nullablereturn. The field is initialised to an empty list and the setter normalisesnull, so the documented "ornullif not set" case was unreachable; the getter now matches its siblinggetDrySequenceBreakers(). The two null-guards this made dead inLlamaCppJniConfigFactorywere dropped with it — the real guard lives inLlamaCppJniConfig's constructor, and is now pinned directly by a newLlamaCppJniConfigTest. -
PIT gate widened from 632 to 775 mutations in the core module, all killed at
mutationThreshold100. Newly gated:provider.LlamaCppJniConfig,config.AiConditionGroup(both already at 100% with no new test —TODO.mdhad listed them as needing "careful fixtures", which was stale),document.AiMdDocumentCodecandindexer.AiIndexPlan(survivors killed here).document.AiMdHeaderCodecis documented as permanently out: its last two survivors are equivalent mutants in the colon-position guard, unkillable through the public API. Coverage was also added for the CLI's.js/.ymlextension aliases and all sixCCommanddispatch arms, the plugin'sbuildConfiguration()/messageOf(), andGenerateEngine's missing-subtree and unknown-factsKeypaths. The count reached 775 by release: later work in this cycle addedprovider.GgufModelInfoto the list and grew the gated classes to 52. Measured onmainat release time: 775 / 16 / 62, one hundred percent in each of the three modules. -
Line-based routing is now covered end to end.
<lines>is documented in the plugin README and the condition layer was tested directly, but no test ever made the indexer'sanyRuleUsesLinesreturn true, socountLinesnever ran during planning. A silent failure there would send every file to the fallback rule with nothing failing. Verified by inverting theusesLinescheck: the new test goes red. Writing it also surfaced a semantic worth knowing — a condition node evaluates exactly one leaf and returns on the first present, withextensionsbeforelines, so combining the two needs an<and>group; the first version of the fixture set both on one node and matched on extension alone. -
AiSourceChunker's observable boundaries are pinned (28/34 mutants, up from 25):maxChars == 1, the guard that stops the last chunk being trimmed and re-split, and the guard that stops a chunk beginning on a newline collapsing to a single character — the last two are silent-content-change shapes. The six remaining mutants are equivalent (capacity hint, empty-list return, two arithmetic terms an enclosingMath.maxabsorbs, a loop head an earlierbreakalready guarantees, and a subset branch that at equality reproduces its input); they are enumerated inTODO.mdso nobody re-hunts them.
-
The
llamacpp-jniprovider produced nothing at all, and had done so since 1.1.0. Every real generation threwIllegalArgumentException: Invalid dry_penalty_last_n value: -1before a single token, becausebuildInferenceParametersforwarded the DRY penalty window unconditionally while the window's default is the-1sentinel and the binding rejects any negative value — llama.cpp b10273 dropped the old "-1= context size" meaning, andnet.ladenthin:llamahas enforced that since 5.0.5. The neighbouringrepeatLastNwas already guarded with>= 0, and its comment even spells out the rule; DRY's window was written earlier and never caught up. Both are guarded now, so-1means "send nothing, llama.cpp keeps its own window", which is what the javadoc always claimed. That javadoc was wrong too, on five sites, and is corrected.What let it survive two releases is worth recording. The belief behind the unguarded call is in the code: "multiplier 0.0 (default) = off, so the base/allowed-length/penalty-last-n knobs have no effect unless opted in." True of the window's effect — and irrelevant, because the wither validates whatever it is handed, DRY enabled or not. Everything that runs in CI uses the
mockprovider (the test suite,Plan, the fat-jar smoke), so no gate ever touched the failing path, and every one of them stayed green throughout. -
The three real-model tests now run instead of being disabled by a flag nobody set.
-DrunNativeLlamaTests=truedates from the initial commit, when the repository carried no model and a developer had to supply one — a correct opt-in at the time. The model was committed six minutes later, and the flag was never revisited: it appears in no workflow and no POM in the repository's entire history, so those tests had never executed anywhere while the ~90 MB GGUF they need was checked out on every job. Turning them on found the defect above in ten seconds. The gate is now a capability check (NativeLlamaAvailability): the tests run by default and skip only where the model or a loadable native library is genuinely absent, each skip naming its cause. -
warnOnTruncatedAnswerandlogPromptCacheReuse, both added in this release, had no executed coverage at all — they sit on the same provider path. Five model-free tests now drive them from a hand-builtChatResponse, pinning the"length"-versus-"stop"decision (the vocabulary trap the code documents), the empty-choicesguard, and theDEBUGline's three counts. Four further tests pin the penalty-window guards without a model, so the regression is caught on every platform rather than only where a GGUF is present; removing either guard reds them. -
publish.ymlconcurrency group: every non-PR run now gets its own group. The block previously claimed a push tomainor av*tag "always runs to completion" becausecancel-in-progressis scoped topull_request. GitHub cancels a pending run whenever a newer run joins the same group behind an in-progress one, and that rule is independent ofcancel-in-progress— so a queued release run onmaincould be dropped silently by a later push. The group expression now appends the uniquegithub.run_idfor non-PR runs; PR runs still share a group per ref and supersede each other as intended. -
Bumped
jackson.version2.22.0 → 2.22.2 (jackson-databind/jackson-dataformat-yaml, pinned in the parentpom.xml) to close GHSA-5jmj-h7xm-6q6v (CVSS 5.3, Medium), flagged by OSV-Scanner againstsrcmorph/pom.xmlandsrcmorph-cli/pom.xmlafter themainmerge of the relocation-stub removal. -
AiMdHeaderSupport.shouldWrite's change-detection chain was not covered by a single test. The seven-way header comparison that decides whether a.ai.mdis regenerated could be replaced wholesale withreturn false;and the class's tests stayed green — verified, not inferred. Two of them wrote the fixture document with a header but no body, so the method returned at its blank-body guard before the comparison ever ran. Since checksum-driven regeneration is a stated design principle, dropping any disjunct would have meant stale files silently never regenerating, with nothing failing. Both tests now write a body, and a parameterized case covers each compared field (h/x/title/c/d/g/a) plus the unchanged-header negative. Re-running the same mutation now fails 9 of 15 tests, and removing a single disjunct fails 2. -
srcmorph:calibrate's chars-per-token silently depended on a cache hit nothing checked. It divided the near-window run's source characters by the prompt tokens the model had evaluated — a figure that excludes whatever the KV cache served. That is only the source's own token count while the base prompt is a cache hit; with no reuse the same divisor also covers the base prompt, so the ratio came out too low and the plan's time estimate with it. Nothing enforced the assumption and, untilcache_nwas surfaced, nothing could even observe whether it held. The ratio is now read from the mid→near size differential overtotalPromptTokens()(evaluated plus cache-served): both runs share the same base prompt, so subtracting them cancels it and the result describes the source text alone, identically whether or not the prefix was reused. A regression test measures the same prompt twice — once fully reused, once not — and requires both to agree; it fails on the previous formula. -
No
execute()verified that it honours the skip flag.shouldSkip()itself was well tested, but a mojo that never called it would have passed every one of those tests — and would have loaded a model and written files under-Dsrcmorph.skip=true. All four goals now assert it,CalibrateMojoincluded (it previously had no test constructing it at all).
-
Relocation-stub module (
llamacpp-ai-index-maven-plugin/,net.ladenthin:llamacpp-ai-index-maven-plugin) removed from the active reactor. It was published once at1.0.4as part of the1.1.1release and the redirect verified working end-to-end (a clean-environment Maven resolution of the old coordinates correctly follows the relocation through tosrcmorph-maven-plugin:1.1.1and its full dependency graph). The published1.0.4artifact is permanent on Maven Central regardless of this repo's module list and will never need another release, so there is no reason to keep carrying the module (and theversions:set -Dexcludes=...caveat it required) in ongoing development..github/workflows/publish.ymlupdated accordingly (no more per-module handling for a fourth artifact). -
llamaLibraryPath/srcmorph.llama.libraryPath— a parameter that never worked.It was declared on every mojo and threaded through
SrcMorphConfiguration,EngineSupport, bothLlamaCppJniConfigFactorymethods andLlamaCppJniConfig— and no code read it, so setting it was a silent no-op. Git history shows this was never a working feature a refactor broke: ina1df3e0 "Initial version."it is already the first field of the then six-fieldLlamaCppJniConfigrecord, next tomodelPath, while that commit's provider built itsModelParametersfrommodelPath/contextSize/threadsalone.It is removed rather than implemented, because the binding already offers the same knob and offers it better.
net.ladenthin.llama.lib.pathis an ordinary JVM system property and the loader's highest-precedence source;MAVEN_OPTSand.mvn/jvm.configset it before the JVM starts, so it is in place before any class loads. A plugin parameter can only act once the provider builds a model — by thenLlamaModel's static initializer may already have resolved the library, and every later value is silently ignored. Implementing it would have added a weaker path with a silent-failure mode alongside one that always works. Verified empirically that the property reaches the build JVM:mvn -Dnet.ladenthin.llama.lib.path=…is readable viaSystem.getPropertyfrom inside the build.Migration: replace
<llamaLibraryPath>DIR</llamaLibraryPath>with-Dnet.ladenthin.llama.lib.path=DIRon the command line, inMAVEN_OPTS, or in.mvn/jvm.config. A POM that still sets the element will fail with an unknown-parameter error; it was doing nothing before, so no behaviour changes with it gone. The plugin README's GPU section already documented the property as the runtime override. -
config.AiGenerationKind— dead code from a superseded design. The enum (FILE_SUMMARY,PACKAGE_SUMMARY,FILE_KEYWORDS,PACKAGE_KEYWORDS) is the original fixed generation taxonomy from when the goals wereSummarizeFilesMojo/SummarizePackagesMojo, and8d1be51and0a6f462deliberately replaced that with configurable prompt keys and rule routing — which is design principle 6, "prompt templates and routing rules are data, never hardcoded in Java". No production class referenced it in any commit. Its two apparent users were concurrency-infrastructure examples that had borrowed the name: the Lincheck one never touched the enum at all (it exercises anAtomicInteger) and is nowAtomicCounterLincheckTest; the jcstress one now races a real enum and isAiOversizeStrategyRace. Both examples are kept — only their misleading names and the dead enum are gone. It waspublicin a published artifact, so formally a breaking change; nothing could construct or receive it, so practically not.
1.1.1 - 2026-07-15
- Reactor split: the former single-module
llamacpp-ai-index-maven-pluginis now a 3-module Maven reactor under a new parent,net.ladenthin:srcmorph-parent—srcmorph(new core library,net.ladenthin:srcmorph, framework-free — no Maven Plugin API dependency),srcmorph-cli(new standalone CLI,net.ladenthin:srcmorph-cli, driven by a single JSON/YAML configuration file, ships as ajava -jar-ready fat jar), andsrcmorph-maven-plugin(the original plugin, now a thin wrapper depending onsrcmorph). All three (plus the parent pom) release together at one shared version. - Plugin renamed from
net.ladenthin:llamacpp-ai-index-maven-plugintonet.ladenthin:srcmorph-maven-pluginin this same release (goal prefixai-index→srcmorph; packagenet.ladenthin.maven.llamacpp.aiindex.mojo→net.ladenthin.maven.srcmorph.mojo; every@ParameterpropertyaiIndex.*→srcmorph.*). A new, independently-versioned relocation-stub module/POM (net.ladenthin:llamacpp-ai-index-maven-plugin:1.0.4, no source, no dependencies, only<distributionManagement><relocation>) keeps the old coordinates resolvable on Maven Central, redirecting tonet.ladenthin:srcmorph-maven-plugin:1.1.1. - New engine layer in
srcmorph(GenerateEngine,AggregatePackagesEngine,AggregateProjectEngine,CalibrateEngine) extracted from what used to be each mojo'sexecute()body, plus a new shared root configuration object,net.ladenthin.srcmorph.config.SrcMorphConfiguration, bindable identically from Maven plexus XML, Jackson JSON/YAML (the new CLI), or plain Java code. - New
examples/directory at the repo root: pairedconfig_*.json/.yamlfixtures for everysrcmorph-clicommand (Plan,GenerateFileIndex,All,Calibrate), pairedrun_*.sh/.batlauncher scripts, and an examplelogbackConfiguration.xml— all runnable out of the box with themockprovider (no GGUF model required). - Per-module
README.mdfiles (srcmorph/README.md,srcmorph-cli/README.md) and a rewritten, product-level rootREADME.md/CLAUDE.mddescribing the reactor.
- Logging in the extracted core/CLI layers moved from a constructor-injected Maven
Logtoorg.slf4j.Logger(see the1.0.xentries below for the indexer-layer half of this change, already shipped before the reactor split). .github/workflows/publish.ymladapted to the 4-module reactor: per-module jar upload/release globs, a repo-wide crash-dump glob, the PIT step scoped tosrcmorph(the only module with a mutation-testing gate), thevmlensjob scoped tosrcmorph-maven-plugin(where its test actually lives), and Coveralls/Codecov pointed atsrcmorph's jacoco report.
- This release renames the Maven plugin's coordinates, package, goal prefix, and
@Parameterproperty names.net.ladenthin:llamacpp-ai-index-maven-plugin→net.ladenthin:srcmorph-maven-plugin; packagenet.ladenthin.maven.llamacpp.aiindex.mojo→net.ladenthin.maven.srcmorph.mojo; goal prefixai-index→srcmorph; everyaiIndex.*property →srcmorph.*. Existing consumers of the old coordinates are not broken: a new, independently-versioned relocation-stub artifact (net.ladenthin:llamacpp-ai-index-maven-plugin:1.0.4, POM-only, no source/dependencies) is published with a<distributionManagement><relocation>pointing atnet.ladenthin:srcmorph-maven-plugin:1.1.1, so Maven transparently redirects any build still declaring the old artifactId.
- Sources-jar signing:
maven-source-plugin'sattach-sourcesexecution was bound to theverifyphase instead ofpackagein all three real modules.maven-gpg-plugin's signing execution is also bound toverify, and within the same phase execution order follows declaration/inheritance order — the inherited gpg execution ran before the sources jar was built, so it was silently omitted from every signed bundle. Reboundattach-sourcestopackage(its own goal default). Also gave the relocation-stub module a<parent>so it inherits the release profile's signing/publishing plugins at all (it previously had none), while keeping its own version pinned independently.
1.1.0 - 2026-07-11
Reconstructed on 2026-08-29 from
git log v1.0.2..v1.1.0(49 commits). This section was missing entirely:v1.1.0was tagged and published, but neither a heading nor a compare-link was ever added, so the chain jumped1.0.2 → 1.1.1and a shipped version was undocumented. The omission is recorded in../workspace/workflows/release-process.mdas the reason the CHANGELOG footer now has a mechanical headings/links/tags check.
srcmorph-climodule — a standalone JSON/YAML-driven CLI (net.ladenthin:srcmorph-cli), with the fat jar as its deliverable, bound unconditionally to thepackagephase.
- Restructured into a 3-module Maven reactor (migration steps 3–9): a parent pom plus the
framework-free core library
srcmorph, the newsrcmorph-cli, and the Maven plugin. The engine layer was extracted out of the mojos so the core carries no dependency on the Maven API. - The Maven plugin was renamed to
net.ladenthin:srcmorph-maven-plugin(packagenet.ladenthin.maven.srcmorph.mojo, goal prefixsrcmorph,@Parameterpropertiessrcmorph.*). The old coordinates stayed resolvable via a one-time relocation-stub POM published at1.0.4. - Maven
Logreplaced with SLF4J throughout the indexer layer, which is what makes the core module Maven-free. - CI adapted to the reactor; a Gradle/BouncyCastle signing-key preflight added (cross-repo sync);
net.ladenthin:llama5.0.4 → 5.0.6;pitest-mavenandjunit-jupiterbumped.
module-infoforsrcmorphwas missingrequires org.slf4j.- The code-style job used bare
spotless/spotbugsprefix goals, which do not resolve in a reactor. - REUSE compliance: SPDX headers / license sidecars added to the benchmark files.
1.0.2 - 2026-07-02
- Split the big-window size-routing rule per source kind in the
ai-index-selftestexample POM: the former singlebig-windowrule is nowbig-window-java(promptfile-body-java), and a matchingbig-window-sqlrule (promptfile-body-sql) routes oversized.sqlsources to the same large-context model — so an oversized.sqlfile keeps the SQL prompt instead of being misrouted/uncovered. - Bumped
net.ladenthin:llama5.0.3 → 5.0.4.
1.0.1 - 2026-06-29
- Third-level project index (
aggregate-projectgoal): a singleproject.ai.mdtable of contents harvesting each package's lead, with an optional one-call AI#### Overviewparagraph. - Rule-based file routing for the
generategoal: each<fieldGeneration>routes a file to a(model, prompt)via a composable<condition>tree (<and>/<or>/<not>overextensions/size/lines/modifiedAfter/modifiedBefore/pathGlob), with<priority>,<skip>, and exactly one explicit<fallback>. Plan-then-execute loads each model once;aiIndex.planOnly=trueprints the routing plan tree without loading a model. - Plan-time context-window fit check: oversized files fail the build up front; new big-window fallback preset (IBM Granite 4.0-H-Tiny, Apache-2.0) covers source files up to ~1 MB.
- Independently switchable phases via
aiIndex.file.skip/aiIndex.package.skip/aiIndex.project.skip(plus the globalaiIndex.skip). - File-size band filter (
minFileSizeBytes/maxFileSizeBytes) for size-tiered indexing, andexcludesglobs to skip generated/trivial sources. - Deterministic child-link list (
Fheader field) for project → package → file navigation. - Per-file progress bar with ETA and measured (actual vs estimated) per-file generation time.
- GPU support: opt-in
gpu-cuda/gpu-vulkanprofiles, parameterized native classifier, andgpuLayers/mainGpu/devicesknobs. - Opt-in sampling controls (default off):
min_p,top_n_sigma, DRY repetition suppression, reasoning/think budget, and model-levelswa-full+ cache-reuse. - Extension-selected file-body prompts (java / sql / fallback).
- Default model preset is now
gpt-oss-20B-mxfp4; all gpt-oss presets are GPU-ready. - Pinned
net.ladenthin:llamato the released5.0.3(dropped the SNAPSHOT dependency and snapshot repository). - Bumped JUnit 6.1.0 → 6.1.1 and palantir-java-format 2.92.0 → 2.94.0.
- The empty-body generation retry mechanism.
1.0.0 - 2026-06-08
First public release on Maven Central. Pre-OpenSSF history themes (March–May 2026): Java 8 compatibility, key-indexed aiDefinitions (PR #21), Sonatype Central Portal migration, JaCoCo+Coveralls+Codecov, GH Actions major-version bumps, CodeQL v3→v4, model catalogue (Qwen2.5-Coder, Ministral 8B/14B, Gemma 4 MoE).
- OpenSSF Best Practices badge and passing-level artifacts (CONTRIBUTING.md, SECURITY.md, CHANGELOG.md).
- Switched runtime dependency to
net.ladenthin:llama5.0.2 (official Maven Central release). - CI: added
startgateabort-window environment before publish pipeline. - CI: separated snapshot and release publish paths; added
check-snapshot/check-taggate jobs. - CI: bumped
softprops/action-gh-releasev2 → v3 (Node 24 compatibility). - CI: added JaCoCo coverage reporting with Coveralls and Codecov integration.
- README: grouped badges by category (Build / Coverage / Package / License / Community); added Maven Central dependency section.
- CI: quoted gate job names to avoid YAML colon-in-scalar parsing error.
- CI: use
GITHUB_TOKENfor Coverallsgithub-tokenparameter instead ofCOVERALLS_TOKEN.