Subject: Re: [Boost-build] [python-port] how to port boost.jam?
From: Juraj Ivanèiæ (juraj.ivancic_at_[hidden])
Date: 2011-11-29 18:36:04
On 29.11.2011. 13:07, Vladimir Prus wrote:
> I am sorry for the delay. I've checked in your previous patches -- it would be appreciated if you
> grab a fresh SVN checkout and see if everything still works. I suppose that setting up local VCS
> will help if you plan to work on some big changes. For quick fixes, I am sure we'll be able to
> promptly get them in one way or another.
I don't plan on making anything much bigger than I already made. After
boost.jam I'll definitely do python.jam. If time permits I might try
porting what is missing to build boost if only to have a nice big test
case for Python port. But this might be a bigger bite than I can chew.
Here are a few patches.
The first patch attempts to correct some minor issues I noticed while
reading code, but did cause me an actual failure. If you are wondering
why I removed 'module' variable - I believe it is unused.
The second patch fixes some issues with <tag>, which is a prerequisite
for boost.py module. A few comments are in order. In
LinkingGenerator::generated_targets() I changed s.real_name() to
s.name() to catch any changes made by tag feature. Tag makes changes in
the base class, in AbstractFileTarget.name_. Without this change <tag>
is ignored for searched libs. I have absolutely no idea why was
real_name() here in the first place, so I'm not 100% sure if this is a
proper fix. The rest of patch 2 is obvious.
The third patch contains ported boost.py.
Might as well stuff the fourth patch - couple of errors in toolset.flags
rules in msvc.py.
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