[PATCH] D92357: clang/darwin: Don't use response files with ld64, do use them with ld64.lld.darwinnew

[Nico Weber via Phabricator via cfe-commits]

thakis added a comment.

In D92357#2424484 <https://reviews.llvm.org/D92357#2424484>, @keith wrote:

> It looks like at least in the `-Wl` case clang passes the file through directly,

Right, that's my point: We see crashes around response files even if we let clang write it.

> I _don't_ actually see this reproing when using `clang @shell_dialogs_unittests.rsp`:

Yes, it's much more fickle to repro . See https://bugs.chromium.org/p/chromium/issues/detail?id=1147968 (linked to from the openrdar): We tried passing response files to clang instead of `-Wl,-filelist` like we currently do in Chromium and after a while (within about a day after the change landing; after hours after 2nd attempt) ld64 started crashing. I failed to repro this locally. With `-Wl, at foo.rsp` the crash reprod locally immediately (and appeared on the bots immediately), so that's what I used for filing the rdar. But even with clang writing the response file, we found ld64 to be unusuably crashy.


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D92357/new/

https://reviews.llvm.org/D92357