hi all, according to the website, the 1.56 beta was scheduled on dec 31st and the release date is feb 3rd ... maybe the release managers could update the release schedule? shifting the dates by a month could be realistic, or what do other people think? cheers, tim
On Wed, Jan 15, 2014 at 9:01 AM, Tim Blechmann <tim@klingt.org> wrote:
hi all,
according to the website, the 1.56 beta was scheduled on dec 31st and the release date is feb 3rd ... maybe the release managers could update the release schedule? shifting the dates by a month could be realistic, or what do other people think?
It does not seem realistic to me to set the schedule until regression tests are cycling regularly and we have had a chance to assess problems the tests may turn up. We also have to decide exactly what goes in a release and someone has to rewrite and test the scripts that generate the release packaging. --Beman
On Jan 15, 2014, at 6:13 AM, Beman Dawes <bdawes@acm.org> wrote:
On Wed, Jan 15, 2014 at 9:01 AM, Tim Blechmann <tim@klingt.org> wrote:
hi all,
according to the website, the 1.56 beta was scheduled on dec 31st and the release date is feb 3rd ... maybe the release managers could update the release schedule? shifting the dates by a month could be realistic, or what do other people think?
It does not seem realistic to me to set the schedule until regression tests are cycling regularly and we have had a chance to assess problems the tests may turn up.
I agree 100%. — Marshall
according to the website, the 1.56 beta was scheduled on dec 31st and the release date is feb 3rd ... maybe the release managers could update the release schedule? shifting the dates by a month could be realistic, or what do other people think?
I think more than a month: regression tests still aren't fully running which means some of us have been holding off on merging changes to release until they are. Even once the tests are cycling again, developers will likely need some time to correct any issues that arise before they can release anything, so the chances of an early release actually containing any changes compared to 1.55 look pretty slim? So I'd suggest the beta starts at least a month after the tests start cycling again. Just my 2c, John.
according to the website, the 1.56 beta was scheduled on dec 31st and the release date is feb 3rd ... maybe the release managers could update the release schedule? shifting the dates by a month could be realistic, or what do other people think?
I think more than a month: regression tests still aren't fully running which means some of us have been holding off on merging changes to release until they are. Even once the tests are cycling again, developers will likely need some time to correct any issues that arise before they can release anything, so the chances of an early release actually containing any changes compared to 1.55 look pretty slim? So I'd suggest the beta starts at least a month after the tests start cycling again.
boost-1.55 does not really support msvc-2013. patches are trivial. imho, this could have justified a 1.55.1 release, but unfortunately boost does not have a notion of bugfix releases. one could argue that compiler vendors should test boost before releasing their compilers, but that would make life way too easy ...
boost-1.55 does not really support msvc-2013. patches are trivial. imho, this could have justified a 1.55.1 release, but unfortunately boost does not have a notion of bugfix releases.
We have done them very occasionally, and I guess if there was a pressing need to get msvc-2013 patches out then that would be a reason to do an early patches-only release. I'm not sure how much that would speed things up though, there's still a lot of figuring out to do before any release can be produced. If anything the big reorganization for Git requires a slower than usual release procedure. John.
On 15 January 2014 18:10, Tim Blechmann <tim@klingt.org> wrote:
boost-1.55 does not really support msvc-2013. patches are trivial. imho, this could have justified a 1.55.1 release, but unfortunately boost does not have a notion of bugfix releases.
I asked for patches before the release. If the will was there, we probably could have got such a release out. But doing it now is awkward as everyone is concentrating on git.
On Wed, Jan 15, 2014 at 1:51 PM, Daniel James <dnljms@gmail.com> wrote:
On 15 January 2014 18:10, Tim Blechmann <tim@klingt.org> wrote:
boost-1.55 does not really support msvc-2013. patches are trivial. imho, this could have justified a 1.55.1 release, but unfortunately boost does not have a notion of bugfix releases.
I asked for patches before the release. If the will was there, we probably could have got such a release out. But doing it now is awkward as everyone is concentrating on git.
+1 Once we are over this initial transition hurdle, it looks to me as if point releases will be much easier to do. But we have to get over the transition hurdle first. --Beman
participants (5)
-
Beman Dawes -
Daniel James -
John Maddock -
Marshall Clow -
Tim Blechmann