From: "Björn Persson" <bjorn@xn--rombobjrn-67a.se>
Subject: Re: architecture-independent GNAT project files?
Date: Tue, 28 Apr 2009 23:41:50 +0200
Date: 2009-04-28T23:41:50+02:00 [thread overview]
Message-ID: <75pbgvF19smr0U1@mid.individual.net> (raw)
In-Reply-To: 7b4268c9-3af5-49ec-ac6a-b5d901d5248c@x1g2000prh.googlegroups.com
Ludovic Brenta wrote:
> Yes. "gcc -v | grep ^Target:" yields the correct information. I
> normally use a minimal Makefile to set up such variables from the
> environment and then invoke gnatmake, passing it the variables.
Neither -m32 nor the setarch command seem to affect the output of "gcc -v".
Anyway, I was hoping to avoid putting a wrapper around gnatmake.
> It is my understanding that "gcc -m32" invokes the linker in such a
> way (i.e. "ld -melf_i386", "ld -mi386linux" or some such) that it
> automatically selects /usr/lib64 instead of /usr/lib. Is that not the
> case?
>
> If I'm right, your problem disappars as far as binary libraries go but
> still exists for target-dependent source files.
I don't know how ld is invoked, but it's not only a matter of finding the
right .so file. The ALI files appear to be architecture-specific, and their
location is specified in the project file with the Library_ALI_Dir
attribute.
--
Bj�rn Persson
PGP key A88682FD
next prev parent reply other threads:[~2009-04-28 21:41 UTC|newest]
Thread overview: 11+ messages / expand[flat|nested] mbox.gz Atom feed top
2009-04-27 7:32 architecture-independent GNAT project files? Björn Persson
2009-04-27 8:02 ` Dmitry A. Kazakov
2009-04-27 21:16 ` Björn Persson
2009-04-27 21:55 ` Ludovic Brenta
2009-04-28 21:41 ` Björn Persson [this message]
2009-04-28 7:24 ` Dmitry A. Kazakov
2009-04-28 20:56 ` sjw
2009-04-30 12:17 ` Björn Persson
2009-04-30 12:31 ` Ludovic Brenta
2009-04-27 12:06 ` anon
2009-04-27 19:10 ` sjw
replies disabled
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox