On Jan 15, 2014, at 6:13 AM, Beman Dawes wrote:

On Wed, Jan 15, 2014 at 9:01 AM, Tim Blechmann 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.

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 ...

On 15 January 2014 18:10, Tim Blechmann 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.