From mboxrd@z Thu Jan  1 00:00:00 1970
X-Spam-Checker-Version: SpamAssassin 3.4.4 (2020-01-24) on polar.synack.me
X-Spam-Level: 
X-Spam-Status: No, score=-1.3 required=5.0 tests=BAYES_00,FREEMAIL_FROM,
	INVALID_MSGID autolearn=no autolearn_force=no version=3.4.4
X-Google-Language: ENGLISH,ASCII-7-bit
X-Google-Thread: 103376,7ab2d703299100e6
X-Google-Attributes: gid103376,public
From: Juergen.Pfeifer@t-online.de (Juergen Pfeifer)
Subject: Re: GNAT shared libraries under Linux
Date: 1998/05/31
Message-ID: <6ksc6a$9ap$1@news00.btx.dtag.de>#1/1
X-Deja-AN: 358451263
Content-Transfer-Encoding: 7bit
References: <3571A970.41C6@cl.cam.ac.uk>
X-Sender: 06162912249-0001@t-online.de
Content-Type: text/plain; charset=us-ascii
Organization: Pfeifer Family
Mime-Version: 1.0
Newsgroups: comp.lang.ada
Date: 1998-05-31T00:00:00+00:00
List-Id: <comp.lang.ada>


Markus Kuhn wrote:
> 
> Is it possible to compile libada.a into a shared library with
> gnat-3.10p under Linux?
> 
> At the moment, the way I use gnat, every executable I produce
> is statically linked with the Ada run-time system, which causes
> even 20 lines long programs to produce easily over 100 kb long
> executables (even with -O2 -gnatp and "strip"ed).
> 
I built it from the sources on my redhat 5.0 system (requires a tiny
patch). The important thing is to use "make gnatlib-shared" if you
build the library. Next step is to make sure, that the linker knows
about the shared library. I simply put a symlink /usr/lib/libgnat.so
to the shared library located in the gcc library subdirectory. After
the usual ldconfig it works fine with the shared runtime.

I'm actually experimenting now with glibc-2.1 und kernel 2.1.103.
This again requires some small patches that I'm currently testing,
because the glibc developers have changed the LinuxThreads a little
bit. I've it up und running now, but there are problems with tasking.
      - Every of the test programs of florist get's a CONSTRAINT_ERROR
        and I've problems to debug this because gdb gets always
        interrupted by the realtime signals the glibc now uses for
        threading.
      - Programs using tasking are not properly terminated when killed
        by a SIGINT. There are always remaining threads that have to be
        killed from the commandline.

Cheers
Juergen