This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: [PATCH 00/25] Public header file hygiene improvements for 2.30
- From: Adhemerval Zanella <adhemerval dot zanella at linaro dot org>
- To: libc-alpha at sourceware dot org
- Date: Fri, 28 Jun 2019 08:27:19 -0300
- Subject: Re: [PATCH 00/25] Public header file hygiene improvements for 2.30
- References: <20190626174954.8009-1-zackw@panix.com>
On 26/06/2019 14:49, Zack Weinberg wrote:
> I would like to propose this set of hygiene improvements to public
> header files for 2.30. Its principal goal is to minimize the number
> of cases where standardized public headers include other public
> headers (standardized or not) and therefore define a bunch of symbols
> that they’re not formally specified to define. It also removes a fair
> amount of cruft from sys/types.h and bits/types.h, obsoletes four
> additional typedefs that should not be used in public interfaces
> anymore (caddr_t, daddr_t, loff_t, register_t), and restricts
> sys/types.h’s definition of u_intN_t and register_t to __USE_MISC
> (i.e. they are no longer present in strict conformance modes).
>
> These changes should not break portable programs, but they do have
> fairly high source compatibility risk; I regret not having been able
> to complete this patch series earlier in the development cycle.
>
> This patch series is a superset of the “sys/types.h cleanup and
> further typedef obsoletion” patch series I posted in March, which was
> never fully reviewed. I have rolled in all of the changes requested
> from what reviews did happen.
I will try to catch up with the patchset, but I am not sure I will be
able to review in full for 2.30. I did a sanity check and looks that
build-many-glibcs.py tools can produce workable toolchains for most
ABIs (I have not checked all multiple variants for every one though,
such the myriad of mips abis).