-
Notifications
You must be signed in to change notification settings - Fork 377
Comparing changes
Open a pull request
base repository: bloomberg/memray
base: v1.9.0
head repository: bloomberg/memray
compare: v1.9.1
- 7 commits
- 13 files changed
- 2 contributors
Commits on Jul 29, 2023
-
Set our recursion guard in all hooks
This prevents reentrancy if one hooked function calls into another. For instance, the jemalloc implementation of `malloc` calls into `mmap` while holding a lock, and our `mmap` hook calls `malloc`, so we must prevent our `mmap` hook from running beneath our `malloc` hook. Signed-off-by: Matt Wozniski <godlygeek@gmail.com>
Configuration menu - View commit details
-
Copy full SHA for 9ac0989 - Browse repository at this point
Copy the full SHA 9ac0989View commit details -
tests: Exercise greenlet for Python 3.11
The latest versions of greenlet now support 3.11, though 3.12 is not yet supported. Update our tests to exercise our greenlet handling on 3.11. Signed-off-by: Matt Wozniski <mwozniski@bloomberg.net>
Configuration menu - View commit details
-
Copy full SHA for 9ac9045 - Browse repository at this point
Copy the full SHA 9ac9045View commit details -
Support old versions of greenlet
Prior to greenlet 1.0, the "greenlet" module wasn't a Python package containing a private extension module, it was just a public extension module. Handle this case by updating our hacky detection for greenlet being imported after tracking has begun to recognize "greenlet.*.so" in addition to "_greenlet.*.so", and update our trace hook injection to fall back to `greenlet.settrace` if `greenlet._greenlet` hasn't been imported but `greenlet` has. Note that we can't unconditionally call `greenlet.set_trace`, because at the point where our hooks are being installed, it's possible that the `greenlet._greenlet` extension module has been imported but the `greenlet/__init__.py` module doesn't yet have a `settrace` attribute because it hasn't yet done `from ._greenlet import settrace`. Signed-off-by: Matt Wozniski <mwozniski@bloomberg.net>
Configuration menu - View commit details
-
Copy full SHA for 99896f5 - Browse repository at this point
Copy the full SHA 99896f5View commit details -
tests: Fix lint errors in greenlet test scripts
These don't show up as lint errors since the scripts are given as string literals, but when I copied these test programs out into standalone scripts to try to investigate the greenlet 0.X issue, Black and flake8 caught these issues. Signed-off-by: Matt Wozniski <mwozniski@bloomberg.net>
Configuration menu - View commit details
-
Copy full SHA for 28b0bea - Browse repository at this point
Copy the full SHA 28b0beaView commit details
Commits on Jul 31, 2023
-
Build macOS wheels from the sdist
We switched from building our Linux wheels from a repo clone to building them from a build sdist, but we missed making the same change for building our macOS wheels. Signed-off-by: Matt Wozniski <mwozniski@bloomberg.net>
Configuration menu - View commit details
-
Copy full SHA for d853623 - Browse repository at this point
Copy the full SHA d853623View commit details
Commits on Aug 1, 2023
-
Fix a bug causing empty
summary
reportsOur TUI's `update_snapshot` method declares that it accepts any iterable, but it expects to be able to iterate over the snapshot repeatedly, which means that it does not accept generators or other iterators that can only be iterated once. Update its type annotation to require a sequence instead of an iterable, and update the caller that was providing a generator to instead provide a sequence. Signed-off-by: Matt Wozniski <mwozniski@bloomberg.net>
Configuration menu - View commit details
-
Copy full SHA for b0d5dc1 - Browse repository at this point
Copy the full SHA b0d5dc1View commit details -
See changelog for more details. Signed-off-by: Matt Wozniski <mwozniski@bloomberg.net>
Configuration menu - View commit details
-
Copy full SHA for 214debf - Browse repository at this point
Copy the full SHA 214debfView commit details
This comparison is taking too long to generate.
Unfortunately it looks like we can’t render this comparison for you right now. It might be too big, or there might be something weird with your repository.
You can try running this command locally to see the comparison on your machine:
git diff v1.9.0...v1.9.1