Boost logo

Boost-Build :

From: Rene Rivera (grafikrobot_at_[hidden])
Date: 2007-08-15 13:42:27

Vladimir Prus wrote:
> On Wednesday 18 July 2007 13:31:43 Vladimir Prus wrote:
>> Rene,
>> have you had the chance to look into
>> Essentially, it seems that your changes to introduce regex-based
>> matching in Boost.Build testsuite had undesirable consequences.
>> I'd like to roll Boost.Build release soon, and therefore would like
>> to get testing back to life. What do you think is the most
>> straight-forward way to that?
> Ping?

OK, I fixed the missing exceptions from re_remove and glob_remove. But
this does nothing to fix the incorrectly reported test failures for
tests that use those two in my Windows msvc tests. And specifying the
toolset doesn't help either. The problem is that what the tests specify
for removing (i.e. expect_touch) does not match because the variant path
is different for msvc-8.0. For example here's the failure for

File bin/msvc-8.0*/debug/c.exe not added as expected
-------- all changes caused by last build command ----------
Added files : ['bin/msvc-8.0/debug/threading-multi/a.exe',
Removed files : []
Modified files: []
Touched files : []
STDOUT ============

STDERR ============

END ===============
FAILED test of bjam
-sBOOST_BUILD_PATH=c:\DevRoots\Boost\boost\tools\build\v2\test -d0 --quiet
        at line 428 of
c:\DevRoots\Boost\boost\tools\build\v2\test\ (fail_test)
        from line 442 of
c:\DevRoots\Boost\boost\tools\build\v2\test\ (expect_addition)
        from line 11 of

For GCC on Linux the "threading-multi" is missing. This is going to be a
problem for just about any toolset has different possible variants than
the default.

Any ideas on how to address this?

-- Grafik - Don't Assume Anything
-- Redshift Software, Inc. -
-- rrivera/ - grafik/
-- 102708583/icq - grafikrobot/aim - grafikrobot/yahoo

Boost-Build list run by bdawes at, david.abrahams at, gregod at, cpdaniel at, john at