tech-toolchain archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: Patch to switch all ports to common_ansi.h



On 28.04.2019 03:10, Christos Zoulas wrote:
> In article <7f33d277-e16e-5623-a998-d0a9617f420b%gmx.com@localhost>,
> Kamil Rytarowski  <n54%gmx.com@localhost> wrote:
>> -=-=-=-=-=-
>> -=-=-=-=-=-
>>
>> I propose a patch unifying all ports to switch to common_ansi.h.
>>
>> http://netbsd.org/~kamil/patch-00107-common_ansi.h.txt
>>
>> It's a follow up of the previous thread on this mailing list:
>>
>> https://mail-index.netbsd.org/tech-toolchain/2019/02/02/msg003393.html
>>
>> It was commonly agreed to switch to this approach. I was waiting for
>> global switch of all ports to GCC7 before this move and this has been
>> accomplished.
>>
>> This patch will unbreak building packages like ovmf (user of
>> -fshort-wchar) with Clang headers:
>>
>> http://pkgsrc.se/sysutils/ovmf
>>
>> The distribution build passes fine on amd64 and all tests pass. If there
>> are problems on any of the ports we will catch real bugs.
>>
>> I intend to get this merged into -9.
> 
> Doesn't this break the c++ ABI?

Not that I'm aware of.

libstdc++.so with and without this patch are the same for me.

> But I would like to see that happen if
> it does not break things.

Maybe, Joerg?

> We can also merge int_fmtio.h, int_limits.h
> int_types.h, etc. using #ifdef _LP64 where needed. Perhaps all that should
> be done together.
> 

ansi.h is correctness switch to stop miscompiling some software (at
least -fshort-wchar one).

Other headers could be unified, but I prefer to perform that independently.

> christos
> 


Attachment: signature.asc
Description: OpenPGP digital signature



Home | Main Index | Thread Index | Old Index