linux build fix

Hello,

I tried to compile revision 182096 of compile-rt on linux. An include missied. I attached the patch and changelog.

by:
Antal Tatrai

changelog (53 Bytes)

missing_include.diff (481 Bytes)

Tátrai Antal <tatraian@deverto.com> writes:

I tried to compile revision 182096 of compile-rt on linux. An include
missied. I attached the patch and changelog.

by:
Antal Tatrai

Index: lib/sanitizer_common/sanitizer_platform_limits_posix.cc

--- lib/sanitizer_common/sanitizer_platform_limits_posix.cc (revision 182096)
+++ lib/sanitizer_common/sanitizer_platform_limits_posix.cc (working copy)
@@ -18,6 +18,7 @@

#include "sanitizer_internal_defs.h"
#include "sanitizer_platform_limits_posix.h"
+#include "sanitizer_platform_limits_posix.h"

This patch duplicates the previous #include.

Apart from that, the most correct place for submitting patches is either
the llvm-commits mailing list or the bug database.

We don't accept patches through the bug database. Patches should be sent to
llvm-commits for review.

-- Sean Silva

Sean Silva <silvas@purdue.edu> writes:

We don't accept patches through the bug database. Patches should be sent to
llvm-commits for review.

Not even patches for fixing bugs? For instance: creating a bug report
with "the build is broken in X, the attached patch fixes it."

No really, that relies on someone looking at the bug report who is
also a person that can and wants to review it and that it's perfect on
the first go.

Pretty small odds.

-eric

Attaching patches to bugs is fine when you want to share a potential or
incomplete patch.

If you want your patch to be considered for review and be committed, you
must email to either llvm-commits or cfe-commits (or both).

cheers,
--renato

Eric Christopher <echristo@gmail.com> writes:

Sean Silva <silvas@purdue.edu> writes:

We don't accept patches through the bug database. Patches should be sent to
llvm-commits for review.

Not even patches for fixing bugs? For instance: creating a bug report
with "the build is broken in X, the attached patch fixes it."

No really, that relies on someone looking at the bug report who is
also a person that can and wants to review it and that it's perfect on
the first go.

Pretty small odds.

I'll expect from component maintainers to review the bugzilla mailing
list watching for bugs on their area, and I'm pretty sure that that is
the case for the most part. A patch attached to a bug report can be
discussed there as any other technicalities related to the bug actually
are.

I've experienced submitting some fix for a simple bug on -commits just
to see it vanish in the stream of everyday's activity and had to keep an
eye on it, pinging until some charitable soul reviewed it. The bug
database, OTOH, never forgets.

IMHO it is reasonable to submit a patch to -commits when you are
determined to incorporate it and attach the patch to a bug report when
you found a fix but are not terribly interested on its fate.

But if the policy is to always send the patch to -commits, I'll keep it
in mind for pointing newcomers to the right direction.