Subject: Re: [Boost-build] [git] direct b2 support for modularised git
From: Jürgen Hunold (jhunold_at_[hidden])
Date: 2013-01-03 06:20:01
On Wednesday, 2. January 2013 12:31:24 Rene Rivera wrote:
> On Wed, Jan 2, 2013 at 12:05 PM, "Jürgen Hunold" <jhunold_at_gmx.eu> wrote:
> > > Ok to commit?
> > The same question again.
> I would like the change to Jamroot be made conditional (with a CLI option)
> to reduce the likelihood of breaking the current trunk. If you can do that
> change go ahead and check it in. After that..
Well, first it has to work. It seems either configure.jam is broken or the new
link.jam has some errors.
When I use (from line 127 in configure.jam, see diff attached)
for local t in $(targets)
rule object(property-set)@878.actualize unknown in module configure
(full log attached, too).
If I only use
for local t in $(targets[2-])
compilation works, but symlink checks are broken. I think I'm missing
something obvious here... The original code used $(targets[2-]) and was
factored out into try-build by Steven. Have to think about that and take a
> The behavior doesn't seem to match the cmake script which creates links to
> the header files directly instead of to the directories. Or am I reading
> the diffs incorrectly?
Well, the behaviour is to create symlinks for individual files, too.
Unfortunatly, the non-symlink fallback is to create copies.
I think that a separate solution providing either symlinks or forward-headers
using link.jam when symlinks are possible would be best. Especially if a
feature controling this is provided. So the end-user can choose between
symlinks and Qt-style forward headers.
I'm runinng short on time, so do not expect new patches from me.
-- * Dipl.-Math. Jürgen Hunold ! * voice: ++49 4257 300 ! Fährstraße 1 * fax : ++49 4257 300 ! 31609 Balge/Sebbenhausen * jhunold_at_gmx.eu ! Germany
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