Thu Apr 26 13:29:00 GMT 2012
On Apr 26, 2012, at 4:10 AM, Fred Cooke wrote:
> Is there any reason why such significant things didn't go into 2.22.1
> and 2.22.2 and so on as they were found and resolved? (with matching
> ports to trunk, of course)
Backporting to the branch is done on request of the maintainers or the patch submitter.
I think this is a very sane and safe approach, and was the tradition.
> Is the release process quite cumbersome?
It takes time, and I have to fit it in my calendar.
> is there a concern for flooding downstreams with too many releases? or
In anycase, I don't think we want too many releases. I think that one yearly major release and
one minor release within the next 6 months makes sense. At binutils is free software, anyone can
build the release branch or the trunk.
> In any case, +1 for a patch release if there are known good tested
> fixes in the branch ready and waiting.
In any cases, I do a full check before each release.
> Excuse my curiosity!
> On Thu, Apr 26, 2012 at 3:28 AM, Mike Frysinger <firstname.lastname@example.org> wrote:
>> On Wednesday 25 April 2012 09:18:15 Joel Sherrill wrote:
>>> Tristan you mentioned that you are not sure there will
>>> be a 2.22.1. Does that mean 2.23 is close or that
>>> you don't see enough issues to justify 2.22.1?
>> fwiw, ppc32 is pretty badly hosed in the 2.22 release. just about every
>> produced includes TEXTRELs. the fix for it is already in the 2.22 branch.
>> hppa32 has incorrect TLS relocation behavior, and plt/ifunc behavior --
>> kind of bad. the fixes are already in the 2.22 branch.
>> s390x fails to build, but that isn't technically a regression from 2.21,
>> maybe not a big deal.
>> so might be nice to have a 2.22.1 release i think :).
More information about the Binutils