László Kiss Kollár

Anyone else running into this issue, make sure the `binutils` package path does not come before the system tools, as otherwise those could be picked up during build and they...

Indeed, the modified library seems to suffer from the same dynamic symbol number mismatch. I tried the workaround posted in #239 to use `lief.ELF.DYNSYM_COUNT_METHODS.SECTION` and that yields the correct number...

> Another side effect of generating the hash before grafting is that it is impossible to validate the hash, if one wanted to make sure no modifications have been done...

I think we should do this, perhaps under a new `validate` or similar subcommand. This could help automated validations, like what we plan to do for PyPI uploads. `validate` could...

Shouldn't we update the PEPs as well? `auditwheel` will start reporting previously compliant wheels as non-compliant and given that `libcrypt.so.1` is still on the list of allowed libraries in the...

I submitted PRs for the manylinux images. GCC 8.2 failed to build `libxcrypt` on the manylinux2010 image so I will have to look into that.

Is there any way you can upload the wheel somewhere so that I can try to debug this? The missing `libjvm.so` in the lddtree output looks odd. Can you run...

Let's reopen this as I intend to keep for the rest of the unit tests.

I wanted to address @takluyver's [request](https://github.com/pypa/auditwheel/pull/169#issuecomment-503127058) on the overall plan for testing in this repo as part of this issue. I should have probably better explained what my intention was...

I'll try to sum up my conclusions here from the discussion: - 100% coverage is not useful enough to offset the churn due to the extra tests which need constant...