Re: [Boost-testing] Re: Whitespace (again)

From: Victor A. Wagner Jr. (vawjr_at_[hidden])
Date: 2005-03-09 01:47:20

At Tuesday 2005-03-08 21:03, you wrote:
>Misha Bergal wrote:
>>del results\*.log
>>...
>>del results\*.xml /s >garbage
>>rem del results\*.obj /s >garbage
>>rem del results\*.exe /s >garbage
> From this I see that you do not completely clean up the results
>>directory (for sure *.exe files and *.test files are not being
>>deleted). This means that actually you are doing an incremental run.
>
>I can see what Victor is trying to do.. It's an attempt to do a clean run 
>by deleting the byproducts of the testing. But without deleting the build 
>byproducts so that things don't get rebuilt. And hence save on the compile 
>time.
>
>Unfortunately, even if you did delete all the byproducts, this would 
>delete the *.xml results for "compile" type tests. And those results would 
>get regenerated unless something in the sources changes. Which would cause 
>the white cells.
>
>If this is a wanted way of running tests it would require some changes of 
>the Boost.Build testing to provide a "testclean" target. The target would 
>do the job of deleting the minimal set of files to have a mostly clean re-run.

Actually, I probably don't need to split up all the regression.py 
stuff.  I've moved to a new computer and my system is set up a little more 
inline with the boost build things and I think I've got all my makefiles, 
etc pointing to the install directory  (\boost).  So I could probably 
invoke the regression test with a single line now.

Ideally, of course, it will get the latest from CVS, do an "install 
(multiple toolsets)" of boost itself (we still, apparently, have to build 
bjam every time since it cannot replace itself) and run the tests and 
upload results.  And it wouldn't do anything that it didn't need to, and 
would manage to get rid of results for tests that no longer exist and also 
those that _should_ have been regenerated but some early error in the test 
process prevents the tests from running.
A tall order, I know

>--
>-- Grafik - Don't Assume Anything
>-- Redshift Software, Inc. - http://redshift-software.com
>-- rrivera/acm.org - grafik/redshift-software.com - 102708583/icq
>_______________________________________________
>Boost-Testing mailing list
>Boost-Testing_at_[hidden]
>http://lists.boost.org/mailman/listinfo.cgi/boost-testing
>

Victor A. Wagner Jr.      http://rudbek.com
The five most dangerous words in the English language:
               "There oughta be a law"