[PATCH] D108332: Have lit preserve SOURCE_DATE_EPOCH

[David Blaikie via Phabricator via llvm-commits]

dblaikie added a comment.

In D108332#2957408 <https://reviews.llvm.org/D108332#2957408>, @serge-sans-paille wrote:

> @dblaikie are you suggesting we should drop this change and  update the tests instead? Although I tend to agree that the tests above are over-constrained, I still think that keeping `SOURCE_DAT_EPOCH` is a good idea.

Mixed feelings - if someone's got a super pedantic build system that requires tests to produce identical intermediate files, I could live with it. But without that requirement, I wonder if it's better to allow the variation - since that means that running tests will look more like the real world. (well, I guess this is only an issue for people who specifically set `SOURCE_DATE_EPOCH` and by setting it they're specifically asking for that behavior)

So, I guess: passing/respecting `SOURCE_DATE_EPOCH` seems like the right thing to do, but for those who are passing it I wonder if they'd be better off/it'd be more suitable not to do so, and to make the tests more robust & continue testing the way the compiler's likely to be run by users on that platform.


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D108332/new/

https://reviews.llvm.org/D108332