Re: Regression test results with subincludes

From: Gennadiy E. Rozental <rogeeff_at_[hidden] (rogeeff_at_[hidden])
Date: 2003-01-06 21:01:00

--- In jamboost_at_[hidden], Beman Dawes <bdawes_at_a...> wrote:
> Okay, regression tests with subincludes have now been tested. See
> http://boost.sourceforge.net/regression-logs/cs-win32-index.html

Wow! This is that rare case when mount came to Mohammed. I was for a 
while in a state "about to add my tests to restression testing 
system". Now reression system came to the Boost.Test to pick up te 
tests ;-)

> I would appreciate it if some of the Boost.Build experts would
> look at libs/test/test/Jamfile and decide if what Gennadiy has
> done needs to be coped with, or is too off-the-wall for regression
> tests.

>From Dave reply (thanks for the compliment) it seems that my Jamfile 
is correct (other that minimal_test should be run_fail - fixed). But 
there is one issue that we will need to cope with.
Two of the Boost.Test unit tests are using external pattern files to 
match against. They are located in libs/test/test/ directory. During 
test run it's assumed that they are in the same location as the 
program. Lack of the patterns files is causing both of them to fail.
Another small issue is that some of the tests create temporary 
files as a side effect. Do we want to delete them after test is 
finished?

> Being able to subinclude regression tests is really great! Thanks 
to Dave 
> for the help providing enough information for the status reporting 
to be 
> able to cope with subincludes.

Second this. Thank you, Dave.

Gennadiy.