Open khanhnt2 opened 1 year ago
Hi! It looks like you're facing a problem with inconsistent analysis results when running Pysa with different orders of source directories. This variability could be due to how Pysa manages its internal state or caching mechanisms. To get consistent results regardless of the order in which directories are specified, you might try running Pysa with a clean state each time using the --no-saved-state
option. Additionally, checking for any issues in the configuration file that might affect how directories are parsed and handled could also help. If these steps don't resolve the problem, consider running Pysa with a single worker to eliminate any issues related to concurrent processing. If the issue persists, reaching out to the Pyre/Pysa community or filing a detailed bug report with the developers could be beneficial, as they might provide more specific guidance or a fix.
Pysa Bug
Pre-submission checklist
Bug description With the same Pysa configuration, when I use multiple source directories (at least 2 or 3 source directories), sometimes the tool spots many issues in these directories, and sometimes it can't spot any issues.
Reproduction steps Suppose we are working in this folder structure:
.pyre_configuration
's content:rules/sources_sinks.pysa
's content:rules/taint.config
's content:types/flask
is Flask stub files from https://pypi.org/project/types-Flask/types/werkzeug
Werkzeug stub files from https://pypi.org/project/types-Werkzeug/vulpy
is a test repo from https://github.com/fportantier/vulpyRun 2 commands, they are in different
--source-directory
options orderExpected behavior The results of the 2 above commands are the same
Logs Please run your reproduction steps with
--noninteractive
(eg.pyre --noninteractive analyze
) and paste the output here:The correct result:
The unreliable result:
Additional context