Rok Mandeljc

> @rokm how do I resign a .app after modifying it? I don't have apple dev account, so users have to right click open to get past security warning. You...

> > So perhaps implementing an specific `--osx-bundle-version` akin to existing `--osx-bundle-identifier` would be better approach for this. Or an option to provide an existing `Info.plist` and parse entries from...

> Any news on this? Did somebody find a solution for this issue? Have you tried adding `--bootloader-ignore-signals` to your PyInstaller command? This should prevent the parent process in onefile...

Hmm, that doesn't look like "`KeyboardInterrupt` while handling `KeyboardInterrupt`" that was originally reported (and which is caused by child process of onefile program receiving the signal twice; once from process...

The UPX check has been reworked in #7707 (PyInstaller 5.13.0), so this should not be a problem on contemporary PyInstaller versions anymore.

@danyeaw Are you sure this is related to the issue you linked? If I recall correctly, the problem with this particular issue was that incompatible GObject DLLs were collected due...

I suppose it could help, if AV's pattern matcher fixated on patterns belonging to .pyc files for collected stdlib modules (assuming zlib output is deterministic, which I think it is...)....

> I did a bit more testing and if I add the library to the `binaries` on the Analysis, the app succeeds, but I'd rather not be bundling libcrypt. It's...

What about the `_crypt` extension (`/usr/local/cpython-exe-builder/lib/python3.9/lib-dynload/_crypt.cpython-39-x86_64-linux-gnu.so`)? That should be linked against `libcrypt.so.1` now... and I think that's the part of python that actually uses `libcrypt`. As per https://bugs.python.org/issue45433, the python...

Maybe we could check if resolved `dot` still has the same basename, and if not, use the original path. This symlinked setup will result in unnecessary duplication of executables, but...