On 2014-02-28, 11:06 AM, Boris Zbarsky wrote:
On 2/28/14 10:49 AM, Gregory Szorc wrote:
Speaking of compiler warnings, do people commonly run into "compiler
warning mismatch" with warnings-as-errors due to running separate
versions of Clang/GCC/MSVC locally than what runs in automation?
I did, to the point where I locally don't --enable-warnings-as-errors
anymore because my build kept breaking. But it's been a few months
since I last checked whether this was still a problem, so maybe it's
gotten better.
FWIW I do have --enable-warnings-as-errors enabled locally because I was
getting bitten more often by my patches causing a fatal warning when I
pushed them out than by my tree failing to build locally. I don't
remember the last time that I got a burnt tree because of a warning
which only happened on my system -- it must have happened at least in 2013.
This was even a problem when I used the exact same clang revision as in
automation but just used slightly different compile options from
automation (e.g. "--enable-debug --disable-optimize" locally, so
debugging is actually sane, vs the "--enable-debug --enable-optimize"
automation uses)...
I believe many (all?) of clang warnings do not depend on the
optimization level so I find this a bit weird. (The same is not true
about gcc!)
i.e. do you find yourself building things fine locally only to run
into try
failures or backouts when the patch runs on automation?
This is very rare, in my case. The reverse issue, of stuff compiling in
automation but not locally, was happening to me about 2-3 times a week
before I went back to not enabling warnings as errors.
My experience is *very* different than Boris' it seems. I guess we can
conclude that no one size fits all here!
Cheers,
Ehsan
_______________________________________________
dev-platform mailing list
dev-platform@lists.mozilla.org
https://lists.mozilla.org/listinfo/dev-platform
