[PATCH] D144084: [LLD] [COFF] Don't try to detect MSVC installations in mingw mode

Martin Storsjö via Phabricator via llvm-commits llvm-commits at lists.llvm.org
Mon Apr 24 23:36:14 PDT 2023


mstorsjo added a comment.

In D144084#4293934 <https://reviews.llvm.org/D144084#4293934>, @mati865 wrote:

> In D144084#4140722 <https://reviews.llvm.org/D144084#4140722>, @mstorsjo wrote:
>
>> I'm not sure what would be best here - add a more specific `-tolerate-duplicate-weak-sybols` option or something like that? Or add a more specific flag, always passed by default by the mingw driver frontend, e.g. `-detect-msvc:no`?
>
> Sorry if you have been waiting for me, this has slipped under my radar.

No problem - I wasn't waiting specifically for you, I just didn't have time to take much further action on this at the time, and it's not a high priority thing.

> I don't know MSVC at all so I cannot provide any suggestions there but for MinGW any way that disables MSVC search by the default probably would be beneficial.
>
> In D144084#4134321 <https://reviews.llvm.org/D144084#4134321>, @mstorsjo wrote:
>
>> I had to revert this, as it broke one test in compiler-rt/test/profile/Windows/coverage-weak-lld.cpp, which relied on linking in an MSVC environment (finding the lib path from the environment) with `-lldmingw` to opt in to mingw style behaviours.
>
> I was under impression lld-link should follow MSVC closely (so this behaviour would be surprising) but that might just a wrong impression from a bystander.

Well, the starting point is that it should follow MSVC style behaviours, but overall I do think lld-link offers a few extensions outside of what's possible in the regular MSVC ecosystem, and that's probably fine in itself. But using `-lldmingw` is indeed quite loosely specified...


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D144084/new/

https://reviews.llvm.org/D144084



More information about the llvm-commits mailing list