Test tools, verification and some bug fixes

[kdvalin]

User description

Description

This PR does a couple things

1. Moves test_tools to $HOME/test_tools which is a well known location we are standardizing on
2. Implements verification to ensure the resulting CSV makes sense, and exit with an appropriate return code
3. Fixes a bug with chameleon where a recent setuptools removed a module that it depended on. If the pyperformance version is affected (<=1.11.0), then we setup the venv separately and install the newest setuptools version that still has the module (v81.0.0).

Before/After Comparison

Before

Test tools would end up at the pwd of the calling shell.

CSVs would not be verified, resulting in possible garbage output

Chameleon would not run

After

Test tools will reside at $HOME/test_tools

CSVs are verified, ensuring that garbage does not land into the output

Chameleon runs again.

Clerical Stuff

Closes #65 and #63

Relates to JIRA: RPOPC-835
Relates to JIRA: RPOPC-823

Supporting files

pyperf.log
pyperf_out_2026.02.17-16.25.31.csv

---

PR Type

Enhancement, Bug fix

---

Description

• Standardize test_tools location to $HOME/test_tools environment variable

• Add CSV verification with schema validation and proper exit codes

• Fix chameleon benchmark by capping setuptools version for pyperformance <=1.11.0

• Refactor output handling to use exec redirection instead of recursive calls

---

Diagram Walkthrough

flowchart LR
  A["pyperf_run script"] -->|uses TOOLS_BIN| B["$HOME/test_tools"]
  A -->|calls setup_pyperformance| C["setuptools v81.0.0 fix"]
  A -->|generates| D["pyperf.json"]
  D -->|verified by| E["verify_results"]
  E -->|returns| F["exit code rtc"]
  A -->|exits with| F

Loading

File Walkthrough

	Relevant files
Enhancement, bug fix	pyperf_run Standardize paths, add verification, fix setuptools            pyperf/pyperf_run Replace hardcoded test_tools references with $TOOLS_BIN environment variable pointing to $HOME/test_tools Add setup_pyperformance() function to cap setuptools at v81.0.0 for pyperformance versions <=1.11.0 Implement CSV verification using csv_to_json and verify_results tools with schema validation Refactor output redirection from recursive call with tee to exec-based approach Propagate verification return code through rtc variable to final exit +36/-22
Enhancement	pyperf_schema.py Add pydantic schema for results validation                              pyperf_schema.py Create new schema file defining Pydantic model for pyperformance results validation Define TestNames enum with all supported benchmark names Define DurationUnits enum for time measurement units Define PyperfResults model with required fields: Test, Avg, Unit, Start_Date, End_Date Add validation constraints: Avg must be positive and not infinite/NaN +124/-0

---

[github-actions]

This relates to RPOPC-835

[qodo-code-review]

PR Compliance Guide 🔍

Below is a summary of compliance checks for this PR:

Security Compliance
⚪	Supply-chain code execution Description: The script git clones external test_tools into $HOME/test_tools (line 129-133) and then executes/sources tools from it (e.g., source $TOOLS_BIN/general_setup at line 287 and $TOOLS_BIN/* invocations at lines 268, 419-420), and also installs setuptools==81.0.0 from PyPI (line 93) without integrity pinning (hash/signature), creating a potential supply-chain path to arbitrary code execution if the repo, transport, or package source is compromised. pyperf_run [129-421] Referred Code if [ ! -d "$TOOLS_BIN" ]; then git clone $tools_git "$TOOLS_BIN" if [ $? -ne 0 ]; then exit_out "pulling git $tools_git failed." 1 fi fi if [ $show_usage -eq 1 ]; then usage $1 fi } generate_csv_file() { instance=0 float=0 ivalue=0 fvalue=0.0 test_name="" unit="" reduce=0 ... (clipped 272 lines)
Ticket Compliance
🟡	🎫 #65 🟢 Fix the chameleon subbenchmark failure caused by missing pkg_resources (setuptools change), so the benchmark runs again across affected environments. Ensure compatibility for affected pyperformance versions (workaround noted: install later pyperformance). ⚪ Confirm chameleon now runs successfully in the listed environments (Ubuntu 24.04, UBI9, AL2023, RHEL 9/10) and that the setuptools cap does not break other benchmarks.
🟡	🎫 #63 🟢 Use pyantics-based verification to validate run results and ensure the resulting CSV/converted output “makes sense”. Handle test_tools properly (implied: consistent/standardized tool location and invocation). ⚪ Confirm csv_to_json + verify_results correctly validate real produced outputs and that exit codes propagate as intended in CI/automation.
Codebase Duplication Compliance
🟢	No codebase code duplication found New Components Detected: - TestNames - PyperfResults
Custom Compliance
🟢	Generic: Meaningful Naming and Self-Documenting Code Objective: Ensure all identifiers clearly express their purpose and intent, making code self-documenting Status: Passed Learn more about managing compliance generic rules or creating your own custom rules
	Generic: Secure Error Handling Objective: To prevent the leakage of sensitive system information through error messages while providing sufficient detail for internal debugging. Status: Passed Learn more about managing compliance generic rules or creating your own custom rules
🔴	Generic: Robust Error Handling and Edge Case Management Objective: Ensure comprehensive error handling that provides meaningful context and graceful degradation Status: Unchecked command failures: Several newly added external commands (e.g., gather_data, csv_to_json, and pip install) are invoked without validating prerequisites or handling failures with actionable context beyond propagating $?. Referred Code $TOOLS_BIN/gather_data ${curdir} exec &> >(tee /tmp/pyperf.out) if [ -d "workloads" ]; then # # If running from zathras, workloads will be symlinked to # to /mnt. Which is done due to azure having a very small # user space. # start_dir=`pwd` cd workloads for file in `ls ${start_dir}`; do if [[ ! -f $file ]] && [[ ! -d $file ]]; then ln -s $start_dir/$file . fi done fi source $TOOLS_BIN/general_setup "$@" ... (clipped 135 lines) Learn more about managing compliance generic rules or creating your own custom rules
⚪	Generic: Comprehensive Audit Trails Objective: To create a detailed and reliable record of critical system actions for security analysis and compliance. Status: Missing audit context: New actions like cloning tools, creating virtualenvs, and installing packages are executed without structured audit logging that includes timestamp/user/action/outcome, making later reconstruction difficult. Referred Code setup_pyperformance() { major=$(echo $PYPERF_VERSION | cut -d. -f1) minor=$(echo $PYPERF_VERSION | cut -d. -f2) # Revert to setuptools v81.0.0 due to `pkg_resources` removal at v82.0.0 # https://setuptools.pypa.io/en/stable/history.html#deprecations-and-removals # Skipping using pkg_tool to ensure that all relevant venv PATHs are honored if [[ "$major" -le 1 ]] && [[ "$minor" -le 11 ]]; then echo "Detected <=1.11.0, applying setuptools fix" $python_exec -m pyperformance venv create venv_path=$($python_exec -m pyperformance venv show | cut -d: -f2 | awk '{ print $1 }') source $venv_path/bin/activate python3 -m pip install setuptools==81.0.0 #Operating within venv, so $python_exec is not needed deactivate fi Learn more about managing compliance generic rules or creating your own custom rules
	Generic: Secure Logging Practices Objective: To ensure logs are useful for debugging and auditing without exposing sensitive information like PII, PHI, or cardholder data. Status: Potential PII in logs: The new flow captures full stdout/stderr to /tmp/pyperf.out and later persists it via save_results, which may include usernames (--user $to_user) or other environment details depending on tool output. Referred Code exec &> >(tee /tmp/pyperf.out) if [ -d "workloads" ]; then # # If running from zathras, workloads will be symlinked to # to /mnt. Which is done due to azure having a very small # user space. # start_dir=`pwd` cd workloads for file in `ls ${start_dir}`; do if [[ ! -f $file ]] && [[ ! -d $file ]]; then ln -s $start_dir/$file . fi done fi source $TOOLS_BIN/general_setup "$@" ARGUMENT_LIST=( "pyperf_version" ... (clipped 148 lines) Learn more about managing compliance generic rules or creating your own custom rules
	Generic: Security-First Input Validation and Data Handling Objective: Ensure all data inputs are validated, sanitized, and handled securely to prevent vulnerabilities Status: Unvalidated external sourcing: The script executes code via source $TOOLS_BIN/general_setup "$@" from a cloned tools directory without validating the source, integrity, or ensuring tools_git/TOOLS_BIN cannot be influenced to run unintended code. Referred Code if [ ! -d "$TOOLS_BIN" ]; then git clone $tools_git "$TOOLS_BIN" if [ $? -ne 0 ]; then exit_out "pulling git $tools_git failed." 1 fi fi if [ $show_usage -eq 1 ]; then usage $1 fi } generate_csv_file() { instance=0 float=0 ivalue=0 fvalue=0.0 test_name="" unit="" reduce=0 ... (clipped 138 lines) Learn more about managing compliance generic rules or creating your own custom rules
Update

Compliance status legend

🟢 - Fully Compliant
🟡 - Partial Compliant
🔴 - Not Compliant
⚪ - Requires Further Human Verification
🏷️ - Compliance label

[qodo-code-review]

PR Code Suggestions ✨

Latest suggestions up to f676b26

Category	Suggestion	Impact
Incremental [*]	Abort when pip install fails Add an error check after the pip install command to ensure it was successful and exit with an informative message if it failed. pyperf/pyperf_run [100] "$venv_path/bin/python3" -m pip install --upgrade setuptools==81.0.0 # Execute venv pip instead + if [ $? -ne 0 ]; then + exit_out "Error: Failed to install pinned setuptools into pyperformance venv" 1 + fi Apply / Chat Suggestion importance[1-10]: 7 __ Why: The suggestion correctly points out that the script does not check the exit code of the pip install command, which could lead to running benchmarks in a broken environment; adding an explicit error check is a significant improvement for robustness.	Medium
	Validate venv path robustly Sanitize the venv_path variable by trimming whitespace and taking the first line, and add a check to ensure it is a valid directory before proceeding. pyperf/pyperf_run [88-90] - if [[ -z "$venv_path" ]]; then - exit_out "Error: Could not determine pyperformance venv path, exiting" 1 + venv_path="$(printf '%s' "$venv_path" | head -n1 | xargs)" + if [[ -z "$venv_path" ]] || [[ ! -d "$venv_path" ]]; then + exit_out "Error: Could not determine valid pyperformance venv path, exiting" 1 fi Apply / Chat Suggestion importance[1-10]: 6 __ Why: The suggestion correctly identifies a potential robustness issue where the venv_path variable might contain unexpected whitespace or multiple lines, and it proposes a solid fix to sanitize the path and validate it's a directory, improving script reliability.	Low
Possible issue	Ensure tools exist before use Add a check to ensure test_tools are installed before they are used, and quote path variables to prevent word splitting issues. pyperf/pyperf_run [274-293] -$TOOLS_BIN/gather_data ${curdir} +if [[ ! -x "$TOOLS_BIN/gather_data" ]] || [[ ! -f "$TOOLS_BIN/general_setup" ]]; then + install_tools +fi + +"$TOOLS_BIN/gather_data" "${curdir}" exec &> >(tee /tmp/pyperf.out) ... -source $TOOLS_BIN/general_setup "$@" +source "$TOOLS_BIN/general_setup" "$@" [To ensure code accuracy, apply this suggestion manually] Suggestion importance[1-10]: 7 __ Why: The suggestion correctly identifies that the script may fail if run on a fresh machine, as it attempts to use tools from $TOOLS_BIN before ensuring they are installed, thus improving the script's robustness.	Medium
	Use string-based enums Inherit from str in the TestNames and DurationUnits enums to ensure they are handled as strings by Pydantic for stricter validation and serialization. pyperf_schema.py [5-116] -class TestNames(Enum): +class TestNames(str, Enum): two_to_three = "2to3" async_generators = "async_generators" async_tree_none = "async_tree_none" ... -class DurationUnits(Enum): +class DurationUnits(str, Enum): second = "sec" millisecond = "ms" microsecond = "us" nanosecond = "ns" [To ensure code accuracy, apply this suggestion manually] Suggestion importance[1-10]: 6 __ Why: The suggestion correctly points out that using str, Enum is a best practice with Pydantic for string-based enumerations, which ensures stricter validation and consistent serialization behavior.	Low
More

---

Previous suggestions

✅ Suggestions up to commit ef093e4

Category	Suggestion	Impact
Incremental [*]	✅ Fix setuptools installation command Suggestion Impact: Replaced the incorrect `python3 install setuptools==81.0.0` invocation with `python3 -m pip install --upgrade setuptools==81.0.0` inside the venv. code diff: - "$venv_path/bin/python3" install setuptools==81.0.0 # Execute venv pip instead + "$venv_path/bin/python3" -m pip install --upgrade setuptools==81.0.0 # Execute venv pip instead Fix the command for installing setuptools by using python -m pip install instead of calling python install. pyperf/pyperf_run [97] -"$venv_path/bin/python3" install setuptools==81.0.0 # Execute venv pip instead +"$venv_path/bin/python3" -m pip install --upgrade setuptools==81.0.0 # Execute venv pip instead [Suggestion processed] Suggestion importance[1-10]: 9 __ Why: The suggestion correctly identifies a bug where the command to install setuptools is syntactically incorrect and would fail, and provides the correct command to fix it.	High
	✅ Validate virtualenv path before use Suggestion Impact: The commit adds an explicit check for an empty venv_path and exits with an error before attempting to use $venv_path/bin/python3. code diff: + if [[ -z "$venv_path" ]]; then + exit_out "Error: Could not determine pyperformance venv path, exiting" 1 + fi if [[ ! -x "$venv_path/bin/python3" ]]; then exit_out "Error: Could not find interpreter at $venv_path/bin/python3, exiting" 1 fi Add a check to ensure the venv_path variable is not empty before using it to prevent accidentally using the system's Python interpreter. pyperf/pyperf_run [88-90] +if [[ -z "$venv_path" ]]; then + exit_out "Error: Could not determine pyperformance venv path, exiting" 1 +fi if [[ ! -x "$venv_path/bin/python3" ]]; then exit_out "Error: Could not find interpreter at $venv_path/bin/python3, exiting" 1 fi [Suggestion processed] Suggestion importance[1-10]: 7 __ Why: The suggestion correctly identifies that an empty venv_path could cause the script to fall back to the system's Python interpreter, and proposes adding a check to prevent this potential bug.	Medium
Possible issue	Propagate tool failures correctly Add an explicit exit code check after the csv_to_json command to ensure any failures are caught and prevent the script from continuing with potentially invalid data. pyperf/pyperf_run [422-424] -$TOOLS_BIN/csv_to_json $to_json_flags --csv_file ${pyresults}.csv --output_file pyperf.json -$TOOLS_BIN/verify_results $to_verify_flags --file pyperf.json --schema_file $to_script_dir/../pyperf_schema.py --class_name PyperfResults +"$TOOLS_BIN/csv_to_json" $to_json_flags --csv_file "${pyresults}.csv" --output_file pyperf.json +if [ $? -ne 0 ]; then + exit_out "Failed: csv_to_json" 1 +fi + +"$TOOLS_BIN/verify_results" $to_verify_flags --file pyperf.json --schema_file "$to_script_dir/../pyperf_schema.py" --class_name PyperfResults rtc=$? Suggestion importance[1-10]: 8 __ Why: The suggestion correctly identifies a bug where a failure in csv_to_json is ignored, which could lead to the script reporting success based on stale data. This improves the script's robustness and correctness.	Medium

✅ Suggestions up to commit 7c193c8

Category	Suggestion	Impact
Incremental [*]	✅ Fix benchmark name typo Suggestion Impact: Updated the TestNames enum entry for dask from "dasks" to "dask", fixing the typo as suggested. code diff: @@ -32,7 +32,7 @@ coroutines = "coroutines" coverage = "coverage" crypto_pyaes = "crypto_pyaes" - dask = "dasks" + dask = "dask" Correct the typo in the TestNames enum from "dasks" to "dask" to ensure correct schema validation for the dask benchmark. pyperf_schema.py [35] -dask = "dasks" +dask = "dask" [Suggestion processed] Suggestion importance[1-10]: 9 __ Why: This suggestion identifies a clear typo ("dasks" instead of "dask") in the schema definition that would cause validation to fail for valid dask benchmark results, which is a critical bug in the new verification logic.	High
	✅ Use venv Python for pip Suggestion Impact: The commit changed the installation command from using the venv's pip3 to invoking the venv's python3 (and added a check that the venv python3 exists), aligning with the intent to rely on the venv interpreter, though it did not implement the recommended `-m pip` usage. code diff: + if [[ ! -x "$venv_path/bin/python3" ]]; then + exit_out "Error: Could not find interpreter at $venv_path/bin/python3, exiting" 1 + fi + # Revert to setuptools v81.0.0 due to `pkg_resources` removal at v82.0.0 # https://setuptools.pypa.io/en/stable/history.html#deprecations-and-removals # Skipping using pkg_tool to ensure that all relevant venv PATHs are honored if [[ "$major" -le 1 ]] && [[ "$minor" -le 11 ]]; then echo "Detected <=1.11.0, applying setuptools fix" - $venv_path/bin/pip3 install setuptools==81.0.0 # Execute venv pip instead + "$venv_path/bin/python3" install setuptools==81.0.0 # Execute venv pip instead fi Use the virtual environment's Python interpreter to run pip (i.e., python -m pip) to ensure the correct pip instance is used for package installation. pyperf/pyperf_run [93] -$venv_path/bin/pip3 install setuptools==81.0.0 # Execute venv pip instead +"$venv_path/bin/python" -m pip install setuptools==81.0.0 # Execute venv pip instead [Suggestion processed] Suggestion importance[1-10]: 7 __ Why: This suggestion correctly recommends using python -m pip which is the canonical and more robust way to ensure the correct pip from the virtual environment is used, preventing potential installation issues.	Medium
	Validate virtualenv path extraction Improve the robustness of virtual environment path extraction by using a more reliable parsing method and adding a validation check to ensure the path is valid. pyperf/pyperf_run [85-86] $python_exec -m pyperformance venv create -venv_path=$($python_exec -m pyperformance venv show | sed -e "s/Virtual environment path: //g" -e "s/ (already created)//" ) +venv_path="$($python_exec -m pyperformance venv show | awk -F': ' '/Virtual environment path:/ {print $2}' | sed -e 's/ (already created)//' -e 's/[[:space:]]*$//')" +if [[ -z "$venv_path" || ! -d "$venv_path" ]]; then + exit_out "Error: Unable to determine pyperformance venv path" 1 +fi Suggestion importance[1-10]: 6 __ Why: The suggestion improves the script's robustness by making the path extraction more reliable and adding an explicit check to ensure the virtual environment path was correctly identified, preventing potential downstream failures.	Low
Possible issue	Quote tool and directory paths Quote usages of $TOOLS_BIN and $curdir to prevent script failures from word-splitting or glob expansion when these paths contain whitespace. pyperf/pyperf_run [267-286] -$TOOLS_BIN/gather_data ${curdir} +"$TOOLS_BIN/gather_data" "${curdir}" exec &> >(tee /tmp/pyperf.out) ... -source $TOOLS_BIN/general_setup "$@" +source "$TOOLS_BIN/general_setup" "$@" Suggestion importance[1-10]: 7 __ Why: The suggestion correctly points out that unquoted variables $TOOLS_BIN and $curdir can lead to script failures if their paths contain spaces, and quoting them is a crucial robustness improvement.	Medium
	✅ Quote venv path executions Suggestion Impact: The change added quoting around the venv_path-based executable path (now "$venv_path/bin/python3" ...) which addresses the word-splitting concern, though it did not implement the exact suggested pip3 quoting. code diff: - $venv_path/bin/pip3 install setuptools==81.0.0 # Execute venv pip instead + "$venv_path/bin/python3" install setuptools==81.0.0 # Execute venv pip instead fi Quote the $venv_path variable when executing pip3 to prevent word-splitting issues if the path contains spaces. pyperf/pyperf_run [85-94] $python_exec -m pyperformance venv create venv_path=$($python_exec -m pyperformance venv show | sed -e "s/Virtual environment path: //g" -e "s/ (already created)//" ) # Revert to setuptools v81.0.0 due to `pkg_resources` removal at v82.0.0 # https://setuptools.pypa.io/en/stable/history.html#deprecations-and-removals # Skipping using pkg_tool to ensure that all relevant venv PATHs are honored if [[ "$major" -le 1 ]] && [[ "$minor" -le 11 ]]; then echo "Detected <=1.11.0, applying setuptools fix" - $venv_path/bin/pip3 install setuptools==81.0.0 # Execute venv pip instead + "$venv_path/bin/pip3" install setuptools==81.0.0 # Execute venv pip instead fi [Suggestion processed] Suggestion importance[1-10]: 6 __ Why: The suggestion correctly identifies that the unquoted $venv_path could cause script failure if the path contains spaces, and proposes a valid fix to improve robustness.	Low

✅ Suggestions up to commit a456bd3

Category	Suggestion	Impact
High-level	Consider a more robust dependency management Instead of manually pinning setuptools inside a shell script for the 'chameleon' benchmark fix, use a dependency constraints file. This makes the dependency management more declarative and easier to maintain. Examples: pyperf/pyperf_run [80-96] setup_pyperformance() { major=$(echo $PYPERF_VERSION | cut -d. -f1) minor=$(echo $PYPERF_VERSION | cut -d. -f2) # Revert to setuptools v81.0.0 due to `pkg_resources` removal at v82.0.0 # https://setuptools.pypa.io/en/stable/history.html#deprecations-and-removals # Skipping using pkg_tool to ensure that all relevant venv PATHs are honored if [[ "$major" -le 1 ]] && [[ "$minor" -le 11 ]]; then echo "Detected <=1.11.0, applying setuptools fix" ... (clipped 7 lines) Solution Walkthrough: Before: # pyperf/pyperf_run setup_pyperformance() { if [[ "$PYPERF_VERSION" <= "1.11.0" ]]; then echo "Applying setuptools fix" $python_exec -m pyperformance venv create venv_path=$($python_exec -m pyperformance venv show ...) source $venv_path/bin/activate python3 -m pip install setuptools==81.0.0 deactivate fi } ... setup_pyperformance $python_exec -m pyperformance run ... After: # pyperf/pyperf_run setup_pyperformance() { if [[ "$PYPERF_VERSION" <= "1.11.0" ]]; then # pyperformance would need to be invoked in a way # that respects a constraints file. # This is a conceptual change. PYPERF_PIP_INSTALL_ARGS="--constraint /path/to/constraints.txt" fi } # /path/to/constraints.txt setuptools==81.0.0 ... setup_pyperformance # The pyperformance tool would internally use the constraints $python_exec -m pyperformance run ... Suggestion importance[1-10]: 8 __ Why: The suggestion correctly identifies that the shell-based manipulation of the virtual environment in setup_pyperformance is a brittle solution and proposes a more robust, standard approach using a dependency constraints file, which would significantly improve maintainability.	Medium
Possible issue	✅ Improve venv path parsing robustness Suggestion Impact: The commit replaced the previous `cut ... | awk ...` parsing of `pyperformance venv show` output with a `sed`-based extraction to obtain `venv_path`, improving robustness for paths with spaces (though with a slightly different sed pattern than suggested). code diff: + $python_exec -m pyperformance venv create + venv_path=$($python_exec -m pyperformance venv show | sed -e "s/Virtual environment path: //g" -e "s/ (already created)//" ) + # Revert to setuptools v81.0.0 due to `pkg_resources` removal at v82.0.0 # https://setuptools.pypa.io/en/stable/history.html#deprecations-and-removals # Skipping using pkg_tool to ensure that all relevant venv PATHs are honored if [[ "$major" -le 1 ]] && [[ "$minor" -le 11 ]]; then echo "Detected <=1.11.0, applying setuptools fix" - $python_exec -m pyperformance venv create - venv_path=$($python_exec -m pyperformance venv show | cut -d: -f2 | awk '{ print $1 }') - source $venv_path/bin/activate - python3 -m pip install setuptools==81.0.0 #Operating within venv, so $python_exec is not needed - deactivate + $venv_path/bin/pip3 install setuptools==81.0.0 # Execute venv pip instead Replace the cut and awk pipeline with a sed command to robustly parse the virtual environment path, ensuring it correctly handles paths that contain spaces. pyperf/pyperf_run [91] -venv_path=$($python_exec -m pyperformance venv show | cut -d: -f2 | awk '{ print $1 }') +venv_path=$($python_exec -m pyperformance venv show | sed 's/Virtual environment: //') [Suggestion processed] Suggestion importance[1-10]: 7 __ Why: The suggestion correctly identifies a potential bug where paths with spaces would be parsed incorrectly and provides a more robust sed-based solution, improving script reliability.	Medium
General	✅ Use direct path for pip install Suggestion Impact: The commit removed `source .../activate` and `deactivate` and switched to using the venv’s own executable directly (`$venv_path/bin/pip3 install ...`) to install setuptools. code diff: + $python_exec -m pyperformance venv create + venv_path=$($python_exec -m pyperformance venv show | sed -e "s/Virtual environment path: //g" -e "s/ (already created)//" ) + # Revert to setuptools v81.0.0 due to `pkg_resources` removal at v82.0.0 # https://setuptools.pypa.io/en/stable/history.html#deprecations-and-removals # Skipping using pkg_tool to ensure that all relevant venv PATHs are honored if [[ "$major" -le 1 ]] && [[ "$minor" -le 11 ]]; then echo "Detected <=1.11.0, applying setuptools fix" - $python_exec -m pyperformance venv create - venv_path=$($python_exec -m pyperformance venv show | cut -d: -f2 | awk '{ print $1 }') - source $venv_path/bin/activate - python3 -m pip install setuptools==81.0.0 #Operating within venv, so $python_exec is not needed - deactivate + $venv_path/bin/pip3 install setuptools==81.0.0 # Execute venv pip instead fi Instead of sourcing the virtual environment's activate script, call the python3 executable directly from the venv's bin directory to install setuptools. pyperf/pyperf_run [92-94] -source $venv_path/bin/activate -python3 -m pip install setuptools==81.0.0 #Operating within venv, so $python_exec is not needed -deactivate +$venv_path/bin/python3 -m pip install setuptools==81.0.0 [Suggestion processed] Suggestion importance[1-10]: 7 __ Why: The suggestion correctly identifies that sourcing activate in a script is not ideal and proposes a more robust, self-contained method by calling the venv's python executable directly.	Medium
	Quote commands and paths Add quotes around the command path and variable expansions in the csv_to_json call to prevent issues with word splitting or globbing. pyperf/pyperf_run [419] -$TOOLS_BIN/csv_to_json $to_json_flags --csv_file ${pyresults}.csv --output_file pyperf.json +"$TOOLS_BIN/csv_to_json" $to_json_flags --csv_file "${pyresults}.csv" --output_file "pyperf.json" Suggestion importance[1-10]: 5 __ Why: The suggestion correctly recommends quoting variables and paths to prevent word splitting and globbing, which is a good practice for improving script robustness.	Low

[dvalinrh]

LGTM