[PATCH] D77605: [FileCheck] Fix --dump-input implicit pattern location

[Thomas Preud'homme via Phabricator via llvm-commits]

thopre added a comment.

In D77605#1970014 <https://reviews.llvm.org/D77605#1970014>, @jdenny wrote:

> I'm new to working with Harbormaster.  Do I manually force it to re-run after updating a patch?


I've seen it do it automatically in other phabricator instances but that doesn't seem to be the case here. Might be configuration-dependent. Better force a rebuild to be sure.



================
Comment at: llvm/utils/FileCheck/FileCheck.cpp:320-322
+      // The implicit buffer ID is the buffer ID minus the input file buffer
+      // and the check file buffer plus one for one-origin indexing.
+      Label << "imp" << (CheckBufferID - 1);
----------------
Is this going to be stable if a new option that require a buffer gets added? I know the tests would catch this if that's the case but I'd prefer to not have to update this formula.

I'm also confused at the comment. Does it match the formula? All I see in the formula is that the implicit buffer ID is the check file buffer minus one. Does the comment need updating? Actually repeating the formula does not seem useful, I'd rather have an explanation as to why the implicit buffer ID can be obtained with this formula.


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D77605/new/

https://reviews.llvm.org/D77605