From: Vladimir Prus (ghost_at_[hidden])
Date: 2003-04-01 05:00:18
> Hi. I got the cygwin version of bjam to pass most of the regression tests
> (fails railsys and print). Two sets of changes were necessary.
Hmm... I though tests were mostly passing on cygwin... but that was only my
idea ;-) Thanks for taking care of them!
The 'railsys' fails for a simple reason: it uses Qt.
I'm not sure why 'print.py' fails. Could you provide the exact error messages?
I think this issue is for Doug.
> The least important ones are related to the tests themselves, mainly
> related to missing 'import gcc ;' in project-root. My guess is that
> something unrelated to cygwin is wrong elsewhere.
Eeh... I don't think it's missing. You should have "using gcc ;" either in
user-config.jam or in "test-config.jam" file. Please see the top of
"test-system.html" for docs.
> The important changes are attached and involve replacing strings with
> lists so that the call to bjam (through popen2) doesn't go through the
> shell (and uses execvp instead). It would be great if someone more
> experienced could take a look at them. I know of a few issues already
> (commented out lines about 'interpreter', for instance).
I have two quick points:
1. Why going through shell is bad?
2. Could you please resend the patch as unified diff ("cvs diff -u"). Somehow,
emacs fails to convert your context diff into unified one, which is much more
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