nickdesaulniers wrote:

> I know there is the detailed issue, but a summary of the "why" would be good 
> to have in the PR description.

Updated the commit messages and PR description, PTAL.

> Moving the defines closer to use makes it easier to reason about, but they 
> are still defined, so don't you have the same problems?

No, and _that's not the problem_.

Take this comment 
https://github.com/llvm/llvm-project/blob/340fd2eb513bc7cfe042a8d6d11934c2093dc995/lldb/source/Plugins/Process/Linux/NativeRegisterContextLinux_arm64dbreg.cpp#L14-L15.

That is the problem; the same pattern of "boy this kernel UAPI headers sure are 
shit, and now I need to do stupid stuff with my include orders to work around 
it."  glibc just doesn't include kernel UAPI headers for ptrace at all for this 
reason (it has its own copy, with all kinds of curious spelling 
differences...).  bionic _does_ copy the UAPI headers for ptrace and thus is 
affected.  I suspect all libc maintainers can't be bothered to fix the UAPI 
headers, and no one on the kernel side cares about UAPI headers _at all_.  I 
suspect [my kernel patch trying to fix 
them](https://lore.kernel.org/llvm/[email protected]/)
 will just be ignored.  Should I put all that in the commit message, or keep 
the toned down version?

https://github.com/llvm/llvm-project/pull/218038
_______________________________________________
lldb-commits mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits

Reply via email to