[Lldb-commits] [PATCH] D68980: [LLDB] [test] Pass --target=<arch>-windows-msvc to clang-cl

Tue Oct 15 05:29:31 PDT 2019

labath added a comment.

In D68980#1709346 <https://reviews.llvm.org/D68980#1709346>, @mstorsjo wrote:

> In D68980#1709314 <https://reviews.llvm.org/D68980#1709314>, @labath wrote:
>
> > Yeah, I raised this issue back when build.py was first introduced, but it did not really get resolved properly. (Back then, windows arm wasn't considered, so I guess the issue wasn't seen as pressing.)
>
>
> Well, windows arm doesn't really affect this aspect at all. But if windows arm didn't exist, then clang-cl would probably error out more directly when run on arm linux, unless given an explicit architecture option.

Ah, so clang-cl takes the host triple, and just sets the "os" field to "windows"? That's fun. :)

In D68980#1709341 <https://reviews.llvm.org/D68980#1709341>, @mstorsjo wrote:

> At least these tests run with `--nodefaultlib` so they shouldn't actually need a full winsdk installation, and as they do run and succeed with this change, it should be possible to rewrite them to use clang-cl directly.

Cool. Can we try that instead? (Maybe don't charge for it straight away, but wait to see if anyone comes up with a reason to not do that.). That would make it consistent with what we do for non-windows targets, where we invoke "clang --target=XXX" directly when we want a specific target hardcoded. If there's a substantial amount of repeated arguments involved, you can make a lit substitution to house them...

Repository:
  rLLDB LLDB

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D68980/new/

https://reviews.llvm.org/D68980