Mike Hommey

e14s?

Note, webextensions don't allow to register about: urls, and AIUI, the only way to have this work would be to go through a toolbar button, which is kind of an...

FWIW, with `--symbols-section .rodata` and a patch to binfarce for it not to filter out `STT_FUNC` symbols in https://github.com/Shnatsel/binfarce/blob/ff619127bd1400d3589fec0ff067b4edb3b45d9c/src/elf64.rs#L247-L252, you can get useful results, like in https://github.com/unicode-rs/unicode-normalization/pull/86#issuecomment-1164910121

It's not enough, but it's already helpful. I've seen cases where [unknown] stays high, though.

Would it be worth renaming the feature to make its status clear? (something like box_syntax_deprecated). Or adding warnings?

I actually failed to reproduce on a x86_64 mac, despite the signature not being valid there too. I did originally hit this on an arm64 DTK, and got someone else...

Editing `create/via_global_ref/builtin/cpython/mac_os.py` to always fallback to `install_name_tool` makes it emit a warning about the first half of the issue: ``` /Library/Developer/CommandLineTools/usr/bin/install_name_tool: warning: changes being made to the file will invalidate...

FWIW, I confirmed both are independent: - After disabling SIP (`csrutil disable`), the codesign issue doesn't appear (no need to replace the signature), but the arm64e problem still does -...

As mentioned in the second comment, the codesign validity preventing python from running actually doesn't happen on x86_64 based Big Sur, for some reason, although the code signature *is* invalid.

The problem is virtualenv is editing the executable to change its dependencies to use `@executable_path`, invalidating the signature.