Re: [boost] VC 7.1/RudbekAssociates test anomaly (was: [1.32 Release] Branch-for-release date)

From: Victor A. Wagner Jr. (vawjr_at_[hidden])
Date: 2004-09-13 16:22:33

All I can say is that the .bat file which runs the regression is available 
in the RudbekOld or RudbekAssociates links
btw the reason for two sets of results is I'm bringing up a new machine and 
the results appeared to be different between my last run on the old system 
and the ones on the new system.  Until we get this resolved it might be a 
good idea to publish both.

As for WHY it hasn't run since Sep 7, I have no clue.  I get the stuff from 
CVS, I run the tests (--incremental)...
if bjam doesn't run all the tests it should, then we need to look there.
I DO note that for some reason we're still running the teststreams test on 
VC8.0 (I don't know how to turn tests on/off, and I see my role as tester 
as someone who will run _exactly_ what's in the CVS.

well, I did have to make some changes so that it would find my installation 
of vc8.0 and NOT try to build dot...
Changes of the three files which show up as "M" follow

C:\Projects\boost>cvs diff -w Jamfile.v2
Index: Jamfile.v2
===================================================================
RCS file: /cvsroot/boost/boost/Jamfile.v2,v
retrieving revision 1.10
diff -w -r1.10 Jamfile.v2
26c26
< build-project libs/graph/build ;
---
 > # build-project libs/graph/build ;
C:\Projects\boost>cvs diff -w tools/build/v1/vc8.0-tools.jam
Index: tools/build/v1/vc8.0-tools.jam
===================================================================
RCS file: /cvsroot/boost/boost/tools/build/v1/vc8.0-tools.jam,v
retrieving revision 1.1
diff -w -r1.1 vc8.0-tools.jam
9c9
<     VC80_ROOT ?= $(ProgramFiles:J=" ")"\\Microsoft Visual Studio .NET 
Whidbey\\VC7" ;
---
 >     VC80_ROOT ?= "g:\\program files\\Microsoft Visual Studio 8\\VC" ;
C:\Projects\boost>cvs diff -w  tools/build/v2/user-config.jam
Index: tools/build/v2/user-config.jam
===================================================================
RCS file: /cvsroot/boost/boost/tools/build/v2/user-config.jam,v
retrieving revision 1.13
diff -w -r1.13 user-config.jam
20c20
< #  using gcc ;
---
 >   using gcc ;
29c29,30
< #  using msvc ;
---
 >   using msvc : 7.1 ;
 >   using msvc : 8.0 ;
At Monday 2004-09-13 02:52, you wrote:
>Stefan Slapeta writes:
> > The assert test in MPL is failing on many compilers. Can you imagine
> > what is the reason for that?
>
>Fixed in the CVS already.
>
> > (BTW ... another test with different
> > results for VC 7.1 on different machines: one failing, one passing!).
>
>Hmm, this one _is_ a little wierd, since the regression was fixed only a
>few hours ago. [Digging into "RudbekAssociates.xml"...] Oh, it looks like
>the test hasn't been re-run since 2004-09-07, which explain why it's
>passing (everything was well back then; the breaking change to
>"boost/mpl/assert.hpp" was committed on "Wed Sep 8 17:47:27 2004 UTC"):
>
><test-log library="mpl" test-name="assert" test-type="compile" 
>test-program="libs/mpl/test/assert.cpp"
>target-directory="bin/boost/libs/mpl/test/assert.test/vc7.1/debug/threading-multi" 
>toolset="vc7.1" show-run-output="false">
><compile result="succeed" timestamp="2004-09-07 04:56:26 UTC">
>
>I don't have an explanation _why_ it hasn't been re-run, though. May be
>Victor can shed some light on this.
>
>--
>Aleksey Gurtovoy
>MetaCommunications Engineering
>
>_______________________________________________
>Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
Victor A. Wagner Jr.      http://rudbek.com
The five most dangerous words in the English language:
               "There oughta be a law"