Re: [Boost-build] V2/1.34 -- msvc + issues

From: Reece Dunn (msclrhd_at_[hidden])
Date: 2006-02-14 03:43:39

Vladimir Prus wrote:
>On Tuesday 14 February 2006 01:06, Reece Dunn wrote:
>
> > 1.  I get an error about zlib/bzip2 for the iostreams library (the patch 
>I
> > provided previously fixes this issue);
> > 2.  If you don't configure python, bjam/bbv2 fails saying it can't find
> > boost_python after warning that python support is disabled;
>
>I'll take a look on those.

Great.

> > 3.  I get the same error that Alexey got - applying his patch fixed this
> > issue.
>
>The patch is in CVS already.

Picked it up last night :). Also, running the tests last night, the 
serialization tests were unable to find boost_wserialization.dll, so I will 
try again with your patch on the main V2/1.34 thread and rerun the tests.

>Can you send me the .diff file?

I uploaded it to the TestingOnBoost page like Alexey. NOTE: I didn't post 
the actual V2 results.

>I've also fixed some issues recently, so maybe
>you could rerun the tests? BTW, if you have problems with CVS access, you 
>can
>get:
>
>   http://engineering.meta-comm.com/boost/snapshot/boost-CVS-HEAD.tar.bz2
>
>I don't know how often it's updated, but seems like changes appear there
>within hour from the time when they are comitted.

CVS access is usually good, but I'm using anonymous access, so that is 
usually 3 hours behind IIRC.

> > I will now try and run daily V1/V2 regressions :).
> > I can also try this for CodeWarrior, intel and borland compilers and 
>even
> > msvc-6.5 (spacetime permitting :)).
>
>I think that would be great, and probably, due to feature freeze, we don't
>need to run them all at the same time. Let's get mscv into shape and then
>we'll handle other toolsets.

In that case, I could rotate once msvc is stable :).

BTW: Is it possible to generate the split results like in the metacomm 
regressions, that way I could put the results on my website, and is it also 
possible to combine the V1 and V2 results like metacomm do so we could have 
side-by-side results as well as the diff.

- Reece