[PATCH] D53994: Fixing lower bound regression in certain situations.

[Denis Yaroshevskiy via Phabricator]

dyaroshev created this revision.
dyaroshev added reviewers: EricWF, mclow.lists.
Herald added subscribers: libcxx-commits, ldionne, christof.

Patch https://reviews.llvm.org/D52697 introduced a regression on certain platforms. (for example on quick-bench: http://quick-bench.com/kguTLHndCQWFuzlxJQT4tJQa_p8)

This patch should fix that regression: http://quick-bench.com/fcV9gaAWPeqB_3-AePOw9H2wxXU

I do not know yet why it behaves like it is.
I do think that these measurements are correct since I've tried multiple variations on quick-bench and didn't stumble into the opposite.
On my machine this regression does not reproduce.

As part of the review, please check that you do not run into this issue on your machine.

Related bug: https://bugs.llvm.org/show_bug.cgi?id=39129


Repository:
  rCXX libc++

https://reviews.llvm.org/D53994

Files:
  include/algorithm
  test/libcxx/algorithms/half_positive.pass.cpp
  test/libcxx/algorithms/unsigned_len_type.pass.cpp

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D53994.172206.patch
Type: text/x-patch
Size: 7256 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/libcxx-commits/attachments/20181101/8c32ecc6/attachment.bin>