Boost logo

Boost-Build :

From: Vladimir Prus (ghost_at_[hidden])
Date: 2006-10-07 14:48:16


On Monday 18 September 2006 10:53, Vladimir Prus wrote:

Rene, did you seen the following message?

Sorry for bothering, but I noticed that your sequence.unique optimization
causes a couple of tests to fail -- which is a bit annoying when I try to
figure out of some change of mine caused real regressions.

Thanks,
Volodya

> On Monday 18 September 2006 04:22, Rene Rivera wrote:
> > Sorry for the long post to follow, not sure how much is relevant :-\
> >
> > Vladimir Prus wrote:
> > > I recall you telling you had some problems running V2 regression tests.
> > > They work just fine for me -- can you tell me what you did and what
> > > errors you got, so that we can set everything right?
> >
> > OK, trying it again as I don't immediately remember what the problems
> > where last time I tried it :-)
> >
> > 1. I launch the MinGW/MSYS shell.
> > 2. export DO_DIFF=diff
> > 3. cd /c/DevRoots/Boost/boost/tools/build/v2/test
> > 4. clear ; python test_all.py
> >
> > The first failure is:
> >
> > --- c:\docume~1\grafik\locals~1\temp\tmp_kizivexpected Sun Sep 17
> > 18:55:59 2006
> > +++ c:\docume~1\grafik\locals~1\temp\tmpzp8sd5actual Sun Sep 17
> > 18:55:59 2006
> > @@ -1,9 +1,9 @@
> > -A.act t1: X1-t1
> > -B.act t1: X1-t1 X2-B
> > -act t1: X1-t1 X2-global X3-global
> > -A.act t2: X1-A X2-t2
> > -B.act t2: X2-t2
> > -act t2: X1-global X2-t2 X3-global
> > -A.act t3: X1-A X3-t3
> > -B.act t3: X2-B X3-t3
> > -act t3: X1-global X2-global X3-t3
> > +A.act t1: X1-t1
> > +B.act t1: X1-t1 X2-B
> > +act t1: X1-t1 X2-global X3-global
> > +A.act t2: X1-A X2-t2
> > +B.act t2: X2-t2
> > +act t2: X1-global X2-t2 X3-global
> > +A.act t3: X1-A X3-t3
> > +B.act t3: X2-B X3-t3
> > +act t3: X1-global X2-global X3-t3
> > module_actions : Expected STDOUT ==========
> > A.act t1: X1-t1
> > B.act t1: X1-t1 X2-B
> > act t1: X1-t1 X2-global X3-global
> > A.act t2: X1-A X2-t2
> > B.act t2: X2-t2
> > act t2: X1-global X2-t2 X3-global
> > A.act t3: X1-A X3-t3
> > B.act t3: X2-B X3-t3
> > act t3: X1-global X2-global X3-t3
> >
> > Actual STDOUT ============
> > A.act t1: X1-t1
> > B.act t1: X1-t1 X2-B
> > act t1: X1-t1 X2-global X3-global
> > A.act t2: X1-A X2-t2
> > B.act t2: X2-t2
> > act t2: X1-global X2-t2 X3-global
> > A.act t3: X1-A X3-t3
> > B.act t3: X2-B X3-t3
> > act t3: X1-global X2-global X3-t3
> >
> > DIFFERENCE
> > Unable to compute difference
> > FAILED
> >
> > Which is weird because there's no difference from the expected and
> > actual output AFAICT. And it is indeed able to do the diff as it
> > displays it.
>
> I recall I seen this failure before, but I though it's gone. It's related
> to some differences in print module behaviour on Linux and Windows --
> either Windows adds some space somewhere or something like that.
>
> > But that's not the interesting problem... Second failure is:
> >
> > generators_test :
> > "['c:\\DevRoots\\Boost\\boost\\tools\\build\\v2\\test\\../../../jam/src\\
> >bi n.ntx86\\bjam',
> > '-sBOOST_BUILD_PATH=c:\\DevRoots\\Boost\\boost\\tools\\build\\v2\\test',
> > '-d0', '--quiet', 'gcc']" returned 1
> > STDOUT ============
> > warning: no toolsets are configured.
> > warning: you won't be able to build C++ programs.
> > warning: please consult the documentation at
> > warning:
> > http://boost.org/boost-build2/doc/html/bbv2/advanced/configuration.html
>
> The testsuite will check for file test-config.jam in the 'test' directory.
> If present, it will be loaded, and user-config.jam will be ignored. It
> looks like you have such file, but it does not define anything. Can you
> check if that's the case?
>
> > So my suggestion would be to put in
> > the test dir a test.bat and test.sh scripts for other people to readily
> > find with the default/best way to run tests with some brief instructions.
>
> Sounds like a good idea.
>
> - Volodya


Boost-Build list run by bdawes at acm.org, david.abrahams at rcn.com, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk