-
Notifications
You must be signed in to change notification settings - Fork 5
Expand file tree
/
Copy pathpom.xml
More file actions
757 lines (742 loc) · 37.3 KB
/
Copy pathpom.xml
File metadata and controls
757 lines (742 loc) · 37.3 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
<?xml version="1.0" encoding="UTF-8"?>
<project xmlns="http://maven.apache.org/POM/4.0.0"
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xsi:schemaLocation="http://maven.apache.org/POM/4.0.0 http://maven.apache.org/xsd/maven-4.0.0.xsd">
<modelVersion>4.0.0</modelVersion>
<groupId>tech.streamfusion</groupId>
<artifactId>StreamFusion</artifactId>
<version>${revision}</version>
<packaging>pom</packaging>
<name>StreamFusion</name>
<description>Native Rust and Apache Arrow/DataFusion acceleration for Apache Flink SQL.</description>
<url>https://streamfusion.tech/</url>
<modules>
<module>streamfusion-runtime</module>
<module>streamfusion-core</module>
<module>streamfusion-kafka</module>
<module>streamfusion-json</module>
<module>streamfusion-csv</module>
<module>streamfusion-raw</module>
<module>streamfusion-avro</module>
<module>streamfusion-avro-confluent-registry</module>
<module>streamfusion-protobuf</module>
<module>streamfusion-fluss</module>
<module>streamfusion-parquet</module>
<module>streamfusion-loader</module>
</modules>
<licenses>
<license>
<name>Apache License, Version 2.0</name>
<url>https://www.apache.org/licenses/LICENSE-2.0</url>
<distribution>repo</distribution>
</license>
</licenses>
<developers>
<developer>
<name>Jordan Epstein</name>
<email>32082339+jordepic@users.noreply.github.com</email>
<organization>datafusion-contrib</organization>
<organizationUrl>https://github.com/datafusion-contrib</organizationUrl>
</developer>
</developers>
<scm>
<connection>scm:git:https://github.com/datafusion-contrib/StreamFusion.git</connection>
<developerConnection>scm:git:ssh://git@github.com/datafusion-contrib/StreamFusion.git</developerConnection>
<url>https://github.com/datafusion-contrib/StreamFusion</url>
<tag>HEAD</tag>
</scm>
<properties>
<revision>0.1.0-rc1</revision>
<maven.compiler.source>17</maven.compiler.source>
<maven.compiler.target>17</maven.compiler.target>
<project.build.sourceEncoding>UTF-8</project.build.sourceEncoding>
<junit.version>5.10.2</junit.version>
<arrow.version>18.3.0</arrow.version>
<flink.version>2.2.1</flink.version>
<!-- Keep the runtime Kafka client aligned with flink-connector-kafka 5.0.0-2.2. A direct
dependency is necessary for source compilation, but pinning an older version here
silently overrides the connector's transitive client and changes producer semantics. -->
<kafka.clients.version>4.2.0</kafka.clients.version>
<scala.binary.version>2.12</scala.binary.version>
<fluss.version>0.9.1-incubating</fluss.version>
<zookeeper.version>3.8.3</zookeeper.version>
<!-- Matches the protobuf runtime Flink's protobuf format generates code against. -->
<protobuf.version>4.32.1</protobuf.version>
<delta.version>4.4.0-SNAPSHOT</delta.version>
<!-- Native build: debug by default for a fast correctness loop. The `bench` profile flips
both of these to a release build, which benchmarks must use (debug Rust is far slower). -->
<native.profile>debug</native.profile>
<!-- The full Java test suite exercises Kafka. Fluss source tests are optional local-cluster
integration tests, enabled explicitly by adding the fluss feature to native.cargo.args.
Deployment packaging builds a
connector-free core and feature-specific extension DSOs separately in
bin/build-release.sh. -->
<native.cargo.args>build --features kafka,parquet,json,csv,raw,avro,protobuf,rocksdb-state</native.cargo.args>
<native.build.skip>false</native.build.skip>
<native.test.resource.skip>false</native.test.resource.skip>
<native.resource.directory>${project.basedir}/../native/target/${native.profile}</native.resource.directory>
<!-- Extra JVM args appended to the Surefire fork's argLine; empty by default. Used to attach a
sampler to a profile-loop test, e.g. -Dsf.extraJvmArgs=-agentpath:.../libasyncProfiler.dylib=... -->
<sf.extraJvmArgs></sf.extraJvmArgs>
<sf.benchmarkJvmArgs></sf.benchmarkJvmArgs>
<!-- Test JVM forks per core fraction: the suite is ~140 independent test classes run
sequentially by default; class-level fork parallelism cuts wall time ~linearly.
MiniClusters bind random ports, temp dirs come from JUnit/Files.createTempDirectory,
and the native library loads per fork, so classes are isolated across JVMs. The
bench profile pins this to 1: timing measurements must not share the machine. -->
<sf.testForks>0.5C</sf.testForks>
</properties>
<dependencies>
<dependency>
<groupId>org.junit.jupiter</groupId>
<artifactId>junit-jupiter</artifactId>
<version>${junit.version}</version>
<scope>test</scope>
</dependency>
<!-- Arrow on the JVM is the marshalling glue this library owns: it fills the C Data
Interface structs handed to native code and reads results back, so it ships with the
library rather than being provided by the host. -->
<dependency>
<groupId>org.apache.arrow</groupId>
<artifactId>arrow-vector</artifactId>
<version>${arrow.version}</version>
</dependency>
<dependency>
<groupId>org.apache.arrow</groupId>
<artifactId>arrow-c-data</artifactId>
<version>${arrow.version}</version>
</dependency>
<dependency>
<groupId>org.apache.arrow</groupId>
<artifactId>arrow-memory-unsafe</artifactId>
<version>${arrow.version}</version>
<scope>runtime</scope>
</dependency>
<!-- The host engine supplies coordination and SQL planning; this library plugs into a
Flink the user already runs, so Flink is provided rather than bundled. Provided
dependencies remain on the test classpath, so the harness still executes SQL. -->
<dependency>
<groupId>org.apache.flink</groupId>
<artifactId>flink-table-api-java-bridge</artifactId>
<version>${flink.version}</version>
<scope>provided</scope>
</dependency>
<dependency>
<groupId>org.apache.flink</groupId>
<artifactId>flink-table-planner_${scala.binary.version}</artifactId>
<version>${flink.version}</version>
<scope>provided</scope>
</dependency>
<dependency>
<groupId>org.apache.flink</groupId>
<artifactId>flink-clients</artifactId>
<version>${flink.version}</version>
<scope>provided</scope>
</dependency>
<!-- The operator test harness ships in the runtime's test artifact, which leans on shared
test utilities from the core test artifact. -->
<dependency>
<groupId>org.apache.flink</groupId>
<artifactId>flink-runtime</artifactId>
<version>${flink.version}</version>
<type>test-jar</type>
<scope>test</scope>
</dependency>
<dependency>
<groupId>org.apache.flink</groupId>
<artifactId>flink-test-utils-junit</artifactId>
<version>${flink.version}</version>
<scope>test</scope>
</dependency>
<!-- StreamFusion consumes Flink's public RocksDB configuration keys and delegates JVM
fallback state to Flink's backend. The distribution provides this artifact. -->
<dependency>
<groupId>org.apache.flink</groupId>
<artifactId>flink-statebackend-rocksdb</artifactId>
<version>${flink.version}</version>
<scope>provided</scope>
</dependency>
<!-- Provides the JUnit5 MiniClusterExtension: the SQL harness tests share one Flink mini-cluster
per test class (redirecting getExecutionEnvironment to it) instead of booting and tearing
down a fresh local cluster on every executeSql().collect() — the suite's dominant cost. -->
<dependency>
<groupId>org.apache.flink</groupId>
<artifactId>flink-test-utils</artifactId>
<version>${flink.version}</version>
<scope>test</scope>
</dependency>
<!-- The native file sources plug into Flink's file source for enumeration/split assignment/
checkpointing, so the connector is a compile-time dependency (provided at runtime by the
Flink distribution, which bundles it). The format jars and Hadoop below stay test-only,
used to write the baselines the native readers/writer are compared against. -->
<dependency>
<groupId>org.apache.flink</groupId>
<artifactId>flink-connector-files</artifactId>
<version>${flink.version}</version>
<scope>provided</scope>
</dependency>
<!-- Flink owns Kafka consumption and production; StreamFusion wraps its byte-array source
and sink with native format operators. The 5.0.0-2.2 release targets Flink 2.2.x and
is provided at runtime by the deployment. -->
<dependency>
<groupId>org.apache.flink</groupId>
<artifactId>flink-connector-kafka</artifactId>
<version>5.0.0-2.2</version>
<scope>provided</scope>
</dependency>
<!-- The JSON format factory. Flink instantiates the original kafka+json table source during SQL
validation (the native node replaces it only at physical-plan time), and the shallow
fallback uses it at runtime. Provided at runtime by the Flink distribution. -->
<dependency>
<groupId>org.apache.flink</groupId>
<artifactId>flink-json</artifactId>
<version>${flink.version}</version>
<scope>provided</scope>
</dependency>
<!-- Avro format: the Confluent-Avro benchmark's row-materializing baseline
(AvroRowDataDeserializationSchema) and its AvroSchemaConverter + Avro writer for generating
the test data. Provided at runtime by the Flink distribution. -->
<dependency>
<groupId>org.apache.flink</groupId>
<artifactId>flink-avro</artifactId>
<version>${flink.version}</version>
<scope>provided</scope>
</dependency>
<!-- Confluent-registry Avro format: the host baseline the native avro-confluent decode is
compared against, and the format factory Flink needs to validate an avro-confluent table
at DDL time. Test-only — the native path fetches registry schemas over plain HTTP, so the
shipped jar has no Confluent client dependency. Its registry client resolves from the
Confluent repository below. -->
<dependency>
<groupId>org.apache.flink</groupId>
<artifactId>flink-avro-confluent-registry</artifactId>
<version>${flink.version}</version>
<scope>test</scope>
</dependency>
<!-- CSV format factory. Flink validates+builds the kafka+csv table source during SQL validation
(the native decode node replaces it at physical-plan time); the format factory must be on
the classpath for that, like flink-json. Provided at runtime by the Flink distribution. -->
<dependency>
<groupId>org.apache.flink</groupId>
<artifactId>flink-csv</artifactId>
<version>${flink.version}</version>
<scope>provided</scope>
</dependency>
<dependency>
<groupId>org.apache.flink</groupId>
<artifactId>flink-parquet</artifactId>
<version>${flink.version}</version>
<scope>test</scope>
</dependency>
<!-- Flink's protobuf format: the host baseline the native protobuf decode is compared against,
and its protoc (below) generates the message class a protobuf table names. It needs
protobuf-java 4.x on the test classpath. -->
<dependency>
<groupId>org.apache.flink</groupId>
<artifactId>flink-protobuf</artifactId>
<version>${flink.version}</version>
<scope>test</scope>
</dependency>
<dependency>
<groupId>com.google.protobuf</groupId>
<artifactId>protobuf-java</artifactId>
<version>${protobuf.version}</version>
<scope>test</scope>
</dependency>
<!-- Hadoop is the filesystem/IO layer Flink's Parquet format reads and writes through; the
Parquet baseline tests need it at runtime. (Hadoop 3.3.4 runs fine on protobuf-java 4.x.) -->
<dependency>
<groupId>org.apache.hadoop</groupId>
<artifactId>hadoop-common</artifactId>
<version>3.3.4</version>
<scope>test</scope>
</dependency>
<dependency>
<groupId>org.apache.hadoop</groupId>
<artifactId>hadoop-mapreduce-client-core</artifactId>
<version>3.3.4</version>
<scope>test</scope>
</dependency>
<!-- Kafka client classes (TopicPartition etc.) referenced by the native source's split reader.
Provided at runtime by the Flink Kafka connector distribution, like flink-connector-kafka
above; also used by the opt-in ingest benchmark and integration tests. -->
<dependency>
<groupId>org.apache.kafka</groupId>
<artifactId>kafka-clients</artifactId>
<version>${kafka.clients.version}</version>
<scope>provided</scope>
</dependency>
<dependency>
<groupId>org.testcontainers</groupId>
<artifactId>kafka</artifactId>
<version>1.21.4</version>
<scope>test</scope>
</dependency>
<dependency>
<groupId>org.testcontainers</groupId>
<artifactId>minio</artifactId>
<version>1.21.4</version>
<scope>test</scope>
</dependency>
<!-- The remote-filesystem sink test writes through Flink's own s3 filesystem (the native sink
only encodes bytes). Discovered from the test classpath via ServiceLoader, standing in for
the plugins/ directory of a real deployment. -->
<dependency>
<groupId>org.apache.flink</groupId>
<artifactId>flink-s3-fs-hadoop</artifactId>
<version>${flink.version}</version>
<scope>test</scope>
</dependency>
<!-- Apache Fluss connector classes. Main code references the public split types for the native
proof seam; the dependency stays provided so StreamFusion does not bundle Fluss at runtime. -->
<dependency>
<groupId>org.apache.fluss</groupId>
<artifactId>fluss-flink-2.2</artifactId>
<version>${fluss.version}</version>
<scope>provided</scope>
</dependency>
<dependency>
<groupId>org.apache.fluss</groupId>
<artifactId>fluss-server</artifactId>
<version>${fluss.version}</version>
<scope>test</scope>
</dependency>
<dependency>
<groupId>org.apache.fluss</groupId>
<artifactId>fluss-server</artifactId>
<version>${fluss.version}</version>
<type>test-jar</type>
<scope>test</scope>
</dependency>
<dependency>
<groupId>org.apache.fluss</groupId>
<artifactId>fluss-common</artifactId>
<version>${fluss.version}</version>
<type>test-jar</type>
<scope>test</scope>
</dependency>
<dependency>
<groupId>org.apache.fluss</groupId>
<artifactId>fluss-rpc</artifactId>
<version>${fluss.version}</version>
<type>test-jar</type>
<scope>test</scope>
</dependency>
<dependency>
<groupId>org.apache.fluss</groupId>
<artifactId>fluss-test-utils</artifactId>
<version>${fluss.version}</version>
<scope>test</scope>
</dependency>
<dependency>
<groupId>org.apache.curator</groupId>
<artifactId>curator-test</artifactId>
<version>5.4.0</version>
<scope>test</scope>
<exclusions>
<!-- Fluss 0.9.1 runs its ZooKeeper client stack at 3.8.3. Hadoop's test-only
transitive otherwise wins with 3.5.6 and the embedded server drops Fluss's
shaded client during coordinator startup. -->
<exclusion>
<groupId>org.apache.zookeeper</groupId>
<artifactId>zookeeper</artifactId>
</exclusion>
</exclusions>
</dependency>
<dependency>
<groupId>org.apache.zookeeper</groupId>
<artifactId>zookeeper</artifactId>
<version>${zookeeper.version}</version>
<scope>test</scope>
</dependency>
<dependency>
<groupId>org.apache.zookeeper</groupId>
<artifactId>zookeeper-jute</artifactId>
<version>${zookeeper.version}</version>
<scope>test</scope>
</dependency>
</dependencies>
<repositories>
<!-- Hosts io.confluent:kafka-schema-registry-client, the transitive of the test-only
flink-avro-confluent-registry baseline above (the shipped jar doesn't touch it). -->
<repository>
<id>confluent</id>
<url>https://packages.confluent.io/maven/</url>
</repository>
</repositories>
<build>
<resources>
<resource>
<directory>${project.basedir}/../src/main/resources</directory>
</resource>
<!-- The build stamp: the project version filtered into a classpath resource, compared
against the loaded native library's own stamp so a stale library on
java.library.path fails loudly instead of silently shadowing the bundled one. -->
<resource>
<directory>${project.basedir}/../src/main/resources-filtered</directory>
<filtering>true</filtering>
</resource>
<!-- The local-development resource. The universal release profile replaces this with
its OS/architecture tree, while Native keeps this location as a development fallback. -->
<resource>
<directory>${native.resource.directory}</directory>
<targetPath>tech/streamfusion/native</targetPath>
<includes>
<include>libstreamfusion.so</include>
<include>libstreamfusion.dylib</include>
</includes>
</resource>
</resources>
<extensions>
<!-- Resolves the OS classifier for the protoc binary artifact below. -->
<extension>
<groupId>kr.motd.maven</groupId>
<artifactId>os-maven-plugin</artifactId>
<version>1.7.1</version>
</extension>
</extensions>
<pluginManagement>
<plugins>
<!-- Generates the protobuf message class a `protobuf`-format table names (and that the host
baseline parses with), from a test .proto, using the same protoc Flink's format uses. -->
<plugin>
<groupId>org.xolstice.maven.plugins</groupId>
<artifactId>protobuf-maven-plugin</artifactId>
<version>0.6.1</version>
<configuration>
<protocArtifact>com.google.protobuf:protoc:${protobuf.version}:exe:${os.detected.classifier}</protocArtifact>
</configuration>
<executions>
<execution>
<id>generate-test-protobuf</id>
<phase>generate-test-sources</phase>
<goals>
<goal>test-compile</goal>
</goals>
</execution>
</executions>
</plugin>
<!-- Keep normal Java edit/compile loops free of Cargo. Native code is built only
for test execution, while the explicit release script builds the portable
optimized payloads. -->
<plugin>
<groupId>org.codehaus.mojo</groupId>
<artifactId>exec-maven-plugin</artifactId>
<version>3.2.0</version>
<executions>
<execution>
<id>build-native</id>
<phase>test-compile</phase>
<goals>
<goal>exec</goal>
</goals>
<configuration>
<skip>${native.build.skip}</skip>
<executable>cargo</executable>
<workingDirectory>${project.basedir}/../native</workingDirectory>
<commandlineArgs>${native.cargo.args}</commandlineArgs>
</configuration>
</execution>
</executions>
</plugin>
<!-- test-compile builds the debug DSO after Maven's normal resource phase. Copy it only
then, so `mvn compile` stays Java-only while tests exercise the same bundled fallback
that a packaged debug artifact will contain. -->
<plugin>
<groupId>org.apache.maven.plugins</groupId>
<artifactId>maven-resources-plugin</artifactId>
<version>3.3.1</version>
<executions>
<execution>
<id>copy-test-native-resource</id>
<phase>test-compile</phase>
<goals>
<goal>copy-resources</goal>
</goals>
<configuration>
<skip>${native.test.resource.skip}</skip>
<outputDirectory>${project.build.outputDirectory}</outputDirectory>
<resources>
<resource>
<directory>${native.resource.directory}</directory>
<targetPath>tech/streamfusion/native</targetPath>
<includes>
<include>libstreamfusion.so</include>
<include>libstreamfusion.dylib</include>
</includes>
</resource>
</resources>
</configuration>
</execution>
</executions>
</plugin>
<!-- Tests load the library by name, so point the loader at cargo's output dir
rather than packaging it into a jar at this stage. -->
<plugin>
<groupId>org.apache.maven.plugins</groupId>
<artifactId>maven-surefire-plugin</artifactId>
<version>3.2.5</version>
<configuration>
<!-- Arrow's off-heap allocator and Flink's runtime both reach into JDK
internals that are sealed by default on JDK 17.
arrow.enable_unsafe_memory_access / enable_null_check_for_get turn off
Arrow's per-accessor bounds + refcount checks (Comet/Spark run with these):
profiling the row<->Arrow transpose showed ~1/3 of native CPU was those
checks (ArrowBuf.checkIndex/ensureAccessible/refCnt on every setSafe). -->
<forkCount>${sf.testForks}</forkCount>
<reuseForks>true</reuseForks>
<argLine>-Xmx8g -Djava.library.path=${project.basedir}/../native/target/${native.profile} -Dstreamfusion.native.development=true -Darrow.enable_unsafe_memory_access=true -Darrow.enable_null_check_for_get=false --add-opens=java.base/java.nio=ALL-UNNAMED --add-opens=java.base/java.lang=ALL-UNNAMED --add-opens=java.base/java.util=ALL-UNNAMED --add-opens=java.base/java.io=ALL-UNNAMED --add-opens=java.base/java.net=ALL-UNNAMED --add-opens=java.base/java.util.concurrent=ALL-UNNAMED --add-opens=java.base/java.util.concurrent.atomic=ALL-UNNAMED --add-opens=java.base/java.lang.reflect=ALL-UNNAMED --add-opens=java.base/sun.nio.ch=ALL-UNNAMED ${sf.extraJvmArgs} ${sf.benchmarkJvmArgs}</argLine>
</configuration>
</plugin>
<!-- The thin artifact remains useful to applications that manage dependencies themselves.
The attached runtime artifact is the self-contained payload embedded by the Flink loader. -->
<plugin>
<groupId>org.apache.maven.plugins</groupId>
<artifactId>maven-shade-plugin</artifactId>
<version>3.5.3</version>
<executions>
<execution>
<id>attach-runtime-payload</id>
<phase>package</phase>
<goals>
<goal>shade</goal>
</goals>
<configuration>
<shadedArtifactAttached>true</shadedArtifactAttached>
<shadedClassifierName>runtime</shadedClassifierName>
<createDependencyReducedPom>false</createDependencyReducedPom>
<filters>
<filter>
<artifact>*:*</artifact>
<excludes>
<exclude>META-INF/*.SF</exclude>
<exclude>META-INF/*.DSA</exclude>
<exclude>META-INF/*.RSA</exclude>
</excludes>
</filter>
</filters>
</configuration>
</execution>
</executions>
</plugin>
</plugins>
</pluginManagement>
</build>
<profiles>
<!-- Central releases are deliberately opt-in. The tag-driven release workflow activates
this profile after assembling the universal native payload. -->
<profile>
<id>release</id>
<properties>
<!-- A dry-run tag overrides these defaults so Central validates the identical
signed bundle but leaves it unpublished. -->
<central.autoPublish>true</central.autoPublish>
<central.waitUntil>published</central.waitUntil>
</properties>
<build>
<plugins>
<plugin>
<groupId>org.codehaus.mojo</groupId>
<artifactId>flatten-maven-plugin</artifactId>
<version>1.6.0</version>
<configuration>
<flattenMode>resolveCiFriendliesOnly</flattenMode>
<updatePomFile>true</updatePomFile>
</configuration>
<executions>
<execution>
<id>flatten-release-pom</id>
<phase>process-resources</phase>
<goals><goal>flatten</goal></goals>
</execution>
</executions>
</plugin>
<plugin>
<groupId>org.apache.maven.plugins</groupId>
<artifactId>maven-source-plugin</artifactId>
<version>3.3.1</version>
<configuration>
<!-- Native libraries belong only in the executable JARs. Maven otherwise
copies the universal resource tree into every sources attachment. -->
<excludeResources>true</excludeResources>
</configuration>
<executions>
<execution>
<id>attach-sources</id>
<goals><goal>jar-no-fork</goal></goals>
</execution>
</executions>
</plugin>
<plugin>
<groupId>org.apache.maven.plugins</groupId>
<artifactId>maven-javadoc-plugin</artifactId>
<version>3.11.2</version>
<configuration>
<doclint>none</doclint>
<quiet>true</quiet>
</configuration>
<executions>
<execution>
<id>attach-javadocs</id>
<goals><goal>jar</goal></goals>
</execution>
</executions>
</plugin>
<plugin>
<groupId>org.apache.maven.plugins</groupId>
<artifactId>maven-gpg-plugin</artifactId>
<version>3.2.8</version>
<configuration>
<gpgArguments>
<arg>--pinentry-mode</arg>
<arg>loopback</arg>
</gpgArguments>
</configuration>
<executions>
<execution>
<id>sign-release-artifacts</id>
<phase>verify</phase>
<goals><goal>sign</goal></goals>
</execution>
</executions>
</plugin>
<plugin>
<groupId>org.sonatype.central</groupId>
<artifactId>central-publishing-maven-plugin</artifactId>
<version>0.11.0</version>
<extensions>true</extensions>
<configuration>
<publishingServerId>central</publishingServerId>
<autoPublish>${central.autoPublish}</autoPublish>
<waitUntil>${central.waitUntil}</waitUntil>
</configuration>
</plugin>
</plugins>
</build>
</profile>
<!--
Flink's source tree compiles the table planner against unshaded Calcite, while the
published planner artifact bundles Calcite under relocated package names. The upstream
suite harness replaces only that planner artifact in an isolated Maven repository and
activates this profile so StreamFusion is compiled against the same source-tree ABI.
-->
<profile>
<id>flink-source-suite</id>
<activation>
<property>
<name>streamfusion.flink-source-suite</name>
</property>
</activation>
<dependencies>
<!-- Flink's parser carries its patched Calcite classes (for example VARIANT).
Keep it before vanilla calcite-core on the source-suite classpath. -->
<dependency>
<groupId>org.apache.flink</groupId>
<artifactId>flink-sql-parser</artifactId>
<version>${flink.version}</version>
<scope>provided</scope>
</dependency>
<dependency>
<groupId>org.apache.flink</groupId>
<artifactId>flink-table-calcite-bridge</artifactId>
<version>${flink.version}</version>
<scope>provided</scope>
</dependency>
<dependency>
<groupId>org.apache.calcite</groupId>
<artifactId>calcite-core</artifactId>
<version>1.36.0</version>
<scope>provided</scope>
</dependency>
</dependencies>
<build>
<plugins>
<!-- The source-suite harness installs StreamFusion into an isolated Maven
repository, then resolves it from a standalone classpath POM. Publish
resolved CI-friendly versions there so ${revision} does not leak into
the installed parent relationship. -->
<plugin>
<groupId>org.codehaus.mojo</groupId>
<artifactId>flatten-maven-plugin</artifactId>
<version>1.6.0</version>
<configuration>
<flattenMode>resolveCiFriendliesOnly</flattenMode>
<updatePomFile>true</updatePomFile>
</configuration>
<executions>
<execution>
<id>flatten-flink-source-suite-pom</id>
<phase>process-resources</phase>
<goals>
<goal>flatten</goal>
</goals>
</execution>
</executions>
</plugin>
</plugins>
</build>
</profile>
<!-- Build and load the production native library. The allocator feature is part of the
measured default: it removes the cross-thread allocation bottleneck in the native
Kafka path and also reduces native state churn. -->
<profile>
<id>bench</id>
<properties>
<sf.testForks>1</sf.testForks>
<!-- Model a deployable shuffle: Flink serializes records even across a local
network edge, so StreamFusion must pay Arrow IPC rather than replacing the
record with a same-JVM process-global handle. -->
<sf.benchmarkJvmArgs>-Dstreamfusion.exchange.zeroCopyLocal=false</sf.benchmarkJvmArgs>
<native.profile>release</native.profile>
<native.cargo.args>build --release --features mimalloc,kafka,parquet,json,csv,raw,avro,protobuf,rocksdb-state</native.cargo.args>
</properties>
</profile>
<!-- Produces the deployable, performance-oriented runtime payload. Callers may override
native.cargo.args for a deliberately smaller development build or to add test features. -->
<profile>
<id>dist</id>
<properties>
<native.profile>release</native.profile>
<native.cargo.args>build --release --no-default-features --features mimalloc,rocksdb-state</native.cargo.args>
</properties>
</profile>
<!-- Packages the resource tree staged by bin/build-release.sh. It deliberately does not
invoke cargo: the script has already built each supported target in its own environment. -->
<profile>
<id>universal</id>
<properties>
<native.build.skip>true</native.build.skip>
<native.resource.directory>${project.basedir}/../native/target/no-local-payload</native.resource.directory>
</properties>
<build>
<resources>
<resource>
<directory>${project.basedir}/../native/target/universal</directory>
<targetPath>tech/streamfusion/native</targetPath>
<includes>
<include>**/libstreamfusion.so</include>
<include>**/libstreamfusion.dylib</include>
</includes>
</resource>
</resources>
</build>
</profile>
<!-- The image test creates Docker containers, so keep its fixture module out of the normal
unit-test reactor. It is selected explicitly with -Pimage-it. -->
<profile>
<id>image-it</id>
<modules>
<module>streamfusion-image-it</module>
</modules>
</profile>
<!-- Delta's Flink connector is still an unpublished private build. Keep its optional
acceleration module out of the default reactor until the normal connector artifact is
publicly resolvable; local Delta development enables it with -Pdelta after publishM2. -->
<profile>
<id>delta</id>
<modules>
<module>streamfusion-delta</module>
</modules>
</profile>
</profiles>
</project>