Boost logo

Boost Testing :

From: David Abrahams (dave_at_[hidden])
Date: 2007-09-04 16:54:33


on Tue Sep 04 2007, Roland Schwarz <roland.schwarz-AT-chello.at> wrote:

> David Abrahams wrote:
>> That's not accurate at all, unless someone has changed how this works
>> since 1.34.0 was released.
>
> "changed how this works" relates possibly to two out of sync branches.
>
> Robert obviously tries to use trunk (which I conclude from presence of
> the message "Graph library does not contain optional GraphML reader."
> which was not present in RC_1_34_0.
>
> Since building on trunk was so broken, that I was not even able to start
> regression building for mingw/cygwin I tried to help Volodya and Rene
> with merging the good parts from RC_1_34_0 over to trunk.
>
> There still might be some holes to fill.
>
>> I don't know who generates this message:
>>
>> warning: Python location is not configured
>> warning: the Boost.Python library won't be built
>
> BOOST_ROOT/libs/python/build/Jamfile.v2
>
> (both in RC_1_34_0 and trunk)

Oh, that's just stupid. I should have deleted that else clause.

> Please let me take the opportunity to reiterate my question, since this
> problem seems to be related:
>
> Should trunk be stabilized? I.e. merge the good parts over from
> RC_1_34_0 to trunk?
>
> Or should we go the other way round? Merge the new parts from trunk to
> RC_1_34_0?

Neither.

1. The trunk needs to be tagged or renamed (e.g., to oldtrunk) so that
   peoples' work checked in there is still available.

2. The 1.34.1 release should be copied into trunk/

3. We should initiate nightly (at least) testing on a known stable set
   of release requirement compilers

4. People should start moving their stuff from oldtrunk into the trunk
   if they are sure it is stable. Anyone who breaks tests on the
   trunk should back out his changes until he thinks he has a fix.

-- 
Dave Abrahams
Boost Consulting
http://www.boost-consulting.com
The Astoria Seminar ==> http://www.astoriaseminar.com

Boost-testing list run by mbergal at meta-comm.com