[cfe-dev] [llvm-dev] 3.8.1 Release Schedule

[Martin J. O'Riordan via cfe-dev]

Thanks Renato,

	MartinO

-----Original Message-----
From: Renato Golin [mailto:renato.golin at linaro.org] 
Sent: 23 April 2016 14:18
To: Martin.ORiordan at movidius.com
Cc: Tom Stellard <tom at stellard.net>; LLVM Dev <llvm-dev at lists.llvm.org>; Release-testers <release-testers at lists.llvm.org>; Clang Dev <cfe-dev at lists.llvm.org>
Subject: Re: [llvm-dev] [cfe-dev] 3.8.1 Release Schedule

On 22 April 2016 at 23:33, Martin J. O'Riordan via llvm-dev <llvm-dev at lists.llvm.org> wrote:
> What fixes are being proposed for 3.8.1 versus 3.8?

The biggest change is possibly not going in fully, so I'm CC'ing cfe-dev for a clear picture.

This bug https://llvm.org/bugs/show_bug.cgi?id=23529 reported an issue with the new GCC ABI for C++11 mangling that wasn't implemented in Clang due to the complete lack of specification and stabilisation on GCC's side, but many Linux distributions decided nevertheless to enable it by default. More importantly, some of them were already released with Clang 3.8 as the only release they'll ever have (example, Ubuntu LTS), so if we don't back-port the ABI patches to 3.8.x (whatever x), Ubuntu LTS will not have a working Clang until the next LTS comes along. AFAIK, many patches were accepted and committed to implement the ABI change, but all of them are for 3.9, which will hopefully be out in time for the next Ubuntu (October).

So far, to my knowledge, no ABI patch was proposed to 3.8 yet. Because of the LTS nature of some Linux distros, this would be relevant even if *after* 3.9.


> I carefully tracked 3.7 to 3.7.1 one changes, but 3.8 came so close on its heels that it turned out I shipped 3.8 to my customers and not the 3.7.1 patch.

All the patches I've seem were pretty minimal. Worth of a 3.8.1, if only early enough to not be made redundant by 3.9.

cheers,
--renato