win32 binary build, committed to VC6 for life of 2.2.x?

[Eric Covener <co...@gmail.com>]

Will the transition to a new MS compiler necessarily be on a major
release boundary? Or is this really on the table at any given time
based on the binaries being more or less a convenience?

wrt third-party modules, it seems like we will have to eventually pull
the trigger and lead (force) the upgrade.


-- 
Eric Covener
covener@gmail.com

Re: win32 binary build, committed to VC6 for life of 2.2.x?

[Jorge Schrauwen <jo...@gmail.com>]

I'd like to give some input on this aswel.
Has the compatibility been tested between MSVC80 and MSVC80Free binary <->
module compilations?

I'd like to do some testing on that but my hardware doesn't allow for a vm
at my dorm at the moment.

On 10/16/07, William A. Rowe, Jr. <wr...@rowe-clan.net> wrote:
>
> Eric Covener wrote:
> > Will the transition to a new MS compiler necessarily be on a major
> > release boundary? Or is this really on the table at any given time
> > based on the binaries being more or less a convenience?
> >
> > wrt third-party modules, it seems like we will have to eventually pull
> > the trigger and lead (force) the upgrade.
>
> I envision this being httpd 2.4 - following our interest in maintaining
> binary compatibility.  A plethora of MSVCRT's out there adds to the
> complexity of following the principal for Windows.
>
> Of course there is the issue of loading third party binaries, and I hope
> that ActiveState and others finally catch up, too.
>
> We won't /break/ VC 6/7 builds, but our defacto binary would likely be VC8
> along the timetable of the next httpd minor or major release.
>
> FYI the next httpd 2.2.7 should build just fine on VC8 for either x86 or
> x64.
> With luck it will even compile cleaner on x64 than it does now.  Only APR
> is really thoroughly 'tested' as being ready for x64, however.
>



-- 
~Jorge

Re: win32 binary build, committed to VC6 for life of 2.2.x?

["William A. Rowe, Jr." <wr...@rowe-clan.net>]

Eric Covener wrote:
> Will the transition to a new MS compiler necessarily be on a major
> release boundary? Or is this really on the table at any given time
> based on the binaries being more or less a convenience?
> 
> wrt third-party modules, it seems like we will have to eventually pull
> the trigger and lead (force) the upgrade.

I envision this being httpd 2.4 - following our interest in maintaining
binary compatibility.  A plethora of MSVCRT's out there adds to the
complexity of following the principal for Windows.

Of course there is the issue of loading third party binaries, and I hope
that ActiveState and others finally catch up, too.

We won't /break/ VC 6/7 builds, but our defacto binary would likely be VC8
along the timetable of the next httpd minor or major release.

FYI the next httpd 2.2.7 should build just fine on VC8 for either x86 or x64.
With luck it will even compile cleaner on x64 than it does now.  Only APR
is really thoroughly 'tested' as being ready for x64, however.