Does anyone object to the third option here?
Thanx,
juli.
----- Forwarded message from Peter Pentchev <[EMAIL PROTECTED]> -----
Delivered-To: [EMAIL PROTECTED]
Date: Wed, 11 Dec 2002 12:48:28 +0200
From: Peter Pentchev <[EMAIL PROTECTED]>
To: Juli Mallett <[EMAIL PROTECTED]>
Cc: [EMAIL PROTECTED]
Subject: Re: lang/mono port
In-Reply-To: <[EMAIL PROTECTED]>
User-Agent: Mutt/1.5.1i
On Wed, Dec 11, 2002 at 02:36:05AM -0800, Juli Mallett wrote:
> * De: Peter Pentchev <[EMAIL PROTECTED]> [ Data: 2002-12-11 ]
> [ Subjecte: lang/mono port ]
> > Hi,
> >
> > In browsing the gnome@ archives, I found that you were interested in
> > updating the lang/mono port to 0.17. You mention some issues with
> > boehm-gc; IMHO, there are no issues really, except for the fact that the
> > Mono build has changed a bit, and the files that need to be tweaked in
> > the pre-patch phase are different: the #include <gc.h> is now in a
> > header file, and is not caught by the find/xargs/sed-i in the port's
> > pre-patch target.
>
> Hrm, the version of libgc I installed from ports didn't seem to install
> the _THREADS related defines that are required, and I had problems with
> GC_MALLOC, and I don't understand why the <gc.h> thing needs changed...
> I'm the one who added the gc wrapper, and I added <gc.h> so it would work
> on FreeBSD. If the CPPFLAGS are wrong, or not being obeyed, then that
> is the problem. Thoughts? If you can clear that up for me, then I'll
> gladly give it a go and push happily for this to be included.
The problem is that there is a gc.h in the mono/metadata/ directory.
Since mono/metadata/Makefile.in has a "DEFS= ... -I.", gc_wrapper.h's
inclusion of <gc.h> actually picks up mono/metadata/gc.h instead of
Boehm's ${LOCALBASE}/include/gc.h :( This is *exactly* the reason for
the problems with GC_MALLOC: Mono's gc.h does not define it, so the
parts of Mono which try to use Boehm's GC cannot see it.
The problem could be solved in three ways:
1. Make the FreeBSD port of boehm-gc install gc.h into <gc/gc.h>
instead of <gc.h>, as it does now - IMHO, this might break lots of
other ports.
2. Remove the -I. from mono/metadata's DEFS - I have no idea what
other changes will be necessitated by this.
3. Rename mono/metadata/gc.h to, say, mono-gc.h, so that there are no
chances of its ever conflicting with any other header file named gc.h
or similarly - this will require propagating the change all through
the Mono sources, but IMHO this is the cleanest solution.
G'luck,
Peter
--
Peter Pentchev [EMAIL PROTECTED] [EMAIL PROTECTED]
PGP key: http://people.FreeBSD.org/~roam/roam.key.asc
Key fingerprint FDBA FD79 C26F 3C51 C95E DF9E ED18 B68D 1619 4553
I am the meaning of this sentence.
----- End forwarded message -----
--
Juli Mallett <[EMAIL PROTECTED]>
OpenDarwin, Mono, FreeBSD Developer.
ircd-hybrid Developer, EFnet addict.
FreeBSD on MIPS-Anything on FreeBSD.
_______________________________________________
Mono-list maillist - [EMAIL PROTECTED]
http://lists.ximian.com/mailman/listinfo/mono-list
