feat: auto-generate ruleset cache on source change #2133
Conversation
![github-actions[bot]](https://avatars.githubusercontent.com/in/15368?s=60&v=4){kind=small}
There was a problem hiding this comment.
Please add bug fixes, new features, breaking changes and anything else you think is worthwhile mentioning to the master (unreleased) section of CHANGELOG.md. If no CHANGELOG update is needed add the following to the PR description: [x] No CHANGELOG update needed
There was a problem hiding this comment.
required changes marked with 👀
Great first pass!
I don't love this logic living somewhat deep in the library code, since it depends on the environment of the program (capa). For example, another project that uses capa as a library may not want to have this behavior. Do you think it's possible to cleanly move the dev/git logic into capa.main?
github-actions
bot
dismissed
their stale review
CHANGELOG updated or no update needed, thanks! 😄
agree here to move this to capa.main or a utils file. |
Moving the dev environment checks to capa.main and importing them from capa.main would create circular dependencies when I tested it. |
I mean to call it from there not to implement them in main. |
This commits groups the computing cache identifier logic into a a single function: compute_cache_identifier. Initially, this the cache id computing was split into two functions compute_ruleset_cache_identifier and compute_cache_identifier.
This commits groups the computing cache identifier logic into a a single function: compute_cache_identifier. Initially, this the cache id computing was split into two functions compute_ruleset_cache_identifier and compute_cache_identifier.
mr-tz
force-pushed
the
autogen-ruleset-cache
branch
from
98916be to
077d4bb
Compare
mr-tz
force-pushed
the
autogen-ruleset-cache
branch
4 times, most recently
from
4f26e07 to
8b89a70
Compare
mr-tz
force-pushed
the
autogen-ruleset-cache
branch
from
8b89a70 to
5b2c5ea
Compare
| if "site-packages" in __file__: | ||
| # running from a site-packages installation | ||
| return False |
There was a problem hiding this comment.
Should we also consider dist-packages here?
There was a problem hiding this comment.
doesn't seem important but if you can demonstrate it's relevant this shouldn't be hard
|
Any concerns or blockers before merging this? |
|
ping @s-ff, @williballenthin, @mike-hunhoff |
Co-authored-by: Willi Ballenthin <wballenthin@google.com>
Closes #1898.
This is a proposal for the issue discussed in #1898.
Let's test out its feasibility and make sure it does not introduce an overhead.