From: David Abrahams (dave_at_[hidden])
Date: 2004-08-10 10:08:03
"Aleksey Gurtovoy" <alexy_at_[hidden]> writes:
> David Abrahams writes:
>> Aleksey Gurtovoy <alexy_at_[hidden]> writes:
>> > The "BUGS, LIMITATIONS" section in the Jam man page
>> > (http://public.perforce.com/public/jam/src/Jam.html) says:
>> > "With the -j flag, errors from failed commands can get staggeringly
>> > mixed up."
>> > This is a real showstopper for running regression tests with -jn, where
>> > n > 1 -- due to this mixing of the tools' output the resulting log
>> > just doesn't make much sense. Is there any chance for this to get fixed?
>> What does "fixed" mean in this context?
> Made to produce a meaningful output that is suitable for parsing.
>> What kind of output would be acceptable?
> Just not mixing up the output of different actions (see a fragment of the
> current output below) would be enough.
I think it's nontrivial to do with a portable (to Win95, e.g.) bjam.
We'd need to port os.popen from Python or something like that. While
we need something like that in bjam eventually I don't think there's
time before the release.
Probably your best bet is to run all the tests with -jn to bring
things up-to-date quickly, and then run them again without -jn to get
the output. Sorry :(
-- Dave Abrahams Boost Consulting http://www.boost-consulting.com
Boost-Build list run by bdawes at acm.org, david.abrahams at rcn.com, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk