[boost] Re: unit_test_development merge progress

From: Gennadiy Rozental (rogeeff_at_[hidden])
Date: 2002-08-10 21:16:35

"Beman Dawes" <bdawes_at_[hidden]> wrote in message
news:4.3.2.7.2.20020810185319.01b1f2f0_at_mailhost.esva.net...
> Gennadiy's unit_test_development branch has been merged into the main CVS
> trunk.
>
> As expected, this major revision to Boost.Test broke a lot of regression
> tests.  Usually just minor stuff like status/Jamfile needing
> "<lib>../libs/test/build/test_exec_monitor" added to a test, or the test
> program failing to #include <iostream> or <boost/cstdlib.hpp>, which
> apparently were #included in the prior version of test_tools.hpp.

Looks like You are managed to wipe out all my changes I committed yesterday.
Particularly *all* issues above were addressed there.
I am fixing it right now by recommitting the changes.

> Originally there were a lot of changes that had to be made to
> status/Jamfile, but Gennadiy made many of them, and then I made some more,
> but it may still have some residual problems.
>
> GCC is causing a certain amount of grief; I had to comment out a function
> (compare_and_continue_impl) in test_tools.hpp to get it to work at
> all.  That's something Gennadiy knows about, and presumably will fix soon.
Was fixed yesterday ;-)

>
> I've cleared over 100 fails already.  (That isn't hard to do; when five
> compilers all fail a test, one fix clears five fails.)  See
> http://boost.sourceforge.net/regression-logs/cs-win32-index.html
boost/cstdlib.hpp I added to test_tools myself. So no need to add it
everywhere.

>
> You can see the remaining problems most clearly on
> http://boost.sourceforge.net/regression-logs/cs-win32-diff.html, which
> shows test results that have changed in a larger font.
>
> I'm exhausted; perhaps someone else could work on clearing errors for
> awhile.  Even if you just concentrate on one compiler, often fixing fails
> for it will fix the fails for other compilers too.

Once I finish this work I will post an update with detailed description for
what was changed.
Could you run the regression test again then?

Gennadiy.