From: Mat Marcus (mat-lists_at_[hidden])
Date: 2005-07-29 12:27:40
-----BEGIN PGP SIGNED MESSAGE-----
Some darwin issues:
Passing -s to the linker generates bad code with the current gcc from
Apple. The <debug-symbols>off could be changed to "-Wl,-x" for the moment,
following bbv1. We might also want to change -g to -gfull and turn on
- -dead_strip for release variant.
No support for varying the architecture between i386, ppc and
universal/fat. the ar tool is not properly supported for universal binaries
and bbv2 builds will fail.
No support for varying the SDK (with -isysroot) or deployment target (via
I can offer you the current (hacked) set of patch files to address all of
these matters. I have (much of) the domain knowledge to specify the
necessary settings, but my boost build skills are still weak. It is
possible that by collaborating we could arrange for much better darwin
support with relatively little additional effort. Interested? I'd love to
see M11 work out of the box for current darwin users.
- --On Friday, July 29, 2005 5:13 PM +0400 Vladimir Prus <ghost_at_[hidden]>
> As I've said before, I'd like to roll Milestone 11 of Boost.Build V2
> in early August.
> There's a couple of pending things:
> - msvc patches from Andrey (waiting for him to send a diff against CVS)
> - Qt4 support from Alo (will examine on Monday)
> If anybody knows of some other thing that must be addressed for M11 --
> like toolsets not working, or regressions with respect to M10 -- please
> let me know.
> - Volodya
> Vladimir Prus
> Boost.Build V2: http://boost.org/boost-build2
-----BEGIN PGP SIGNATURE-----
Version: PGP Desktop 9.0.1 (Build 2185)
-----END PGP SIGNATURE-----
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