This is the mail archive of the libc-alpha@sourceware.org mailing list for the glibc project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Avoid "inline" after return type in function definitions


On Fri, 15 Feb 2019, Paul Eggert wrote:

> Because we update manywarnings with each new GCC release, downstream projects
> that use the module don't need to track GCC versions all that carefully.

One thing we get from using -Wall is that we know within a day or two (via 
my build-many-glibcs.py bots) if any new warnings added to GCC mainline 
break the glibc (or testsuite) build on any platform, and so can either 
fix glibc immediately if the warning is pointing out a real problem with 
glibc, or give early feedback on issues with the warning while it's still 
fresh in the mind of whoever added it to GCC.  Anything only updated 
retrospectively after a release would lose that quick feedback and 
corresponding opportunity for the warning to be refined, while GCC is in a 
development stage where significant incompatible changes to a new warning 
are OK.

I.e., I'm considering the GNU toolchain as a whole as something developed 
together, where we get benefits for the toolchain by dogfooding new 
warnings as soon as they are added to it.

-- 
Joseph S. Myers
joseph@codesourcery.com


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]