Mailing List Archive: linker error during compilation

linker error during compilation

Feb 2, 2003, 1:30 PM

Post #1 of 4 (734 views)

I grabbed the cvs version of libgcrypt (the PACKAGE_VERSION from config.h is
noted as being 1.1.13-cvs) today and received the following error from the
linker:

....
/usr/bin/ld:./libgcrypt.vers:23: parse error in VERSION script
collect2: ld returned 1 exit status
make[2]: *** [libgcrypt.la] Error 1
make[2]: Leaving directory
....

According to the documentation that I have for my linker, it says that the
syntax for the VERSION command is:

VERSION { commands }

where VERSION should not be blank.

I have attached a tiny patch that fixes this problem on my system. You may
wish to alter the version identifier. Thanks for the great software.

-Blake

--
Blake Matheny "... one of the main causes of the fall of the
bmatheny@purdue.edu Roman Empire was that, lacking zero, they had
http://www.dbaseiv.net no way to indicate successful termination of
http://ovmj.org/GNUnet/ their C programs." --Robert Firth

Re: linker error during compilation [ In reply to ]

wk at gnupg

Feb 3, 2003, 10:41 AM

Post #2 of 4 (733 views)

Permalink

On Sun, 2 Feb 2003 15:30:29 -0500, Blake Matheny said:

> VERSION { commands }

> where VERSION should not be blank.

According to Ulrich Drepper's document on DSOs the existing script
should be okay. It might be that your ld is too old. I know, we
should write an autoconf test.

Thanks,

Werner

Re: linker error during compilation [ In reply to ]

Marcus.Brinkmann at ruhr-uni-bochum

Feb 4, 2003, 7:57 AM

Post #3 of 4 (728 views)

Permalink

On Sun, Feb 02, 2003 at 03:30:29PM -0500, Blake Matheny wrote:
> According to the documentation that I have for my linker, it says that the
> syntax for the VERSION command is:
>
> VERSION { commands }
>
> where VERSION should not be blank.

Your report is very incomplete, and thus we don't know how to fix this
problem. More important than a patch is always an accurate description of
the problem, and that includes the operating system you are using, and the
version of the linker (and if it is the system linker or GNU ld).

Without a VERSION, it is an anonymous version map. This is supported by
binutils 2.13 or later, probably also by earlier versions. It should also
be supported by Solaris.

Please let us know which system and version you are using.

Thanks,
Marcus

--
`Rhubarb is no Egyptian god.' GNU http://www.gnu.org marcus@gnu.org
Marcus Brinkmann The Hurd http://www.gnu.org/software/hurd/
Marcus.Brinkmann@ruhr-uni-bochum.de
http://www.marcus-brinkmann.de/

Re: linker error during compilation [ In reply to ]

bmatheny at purdue

Feb 4, 2003, 8:17 AM

Post #4 of 4 (728 views)

Permalink

It would be helpful if libgcrypt had a script for collecting this type of
information. In any case, here is the relevant information:
OS : Linux
OS RELEASE : 2.4.18-4_72smp
HARDWARE : i686
gcc version : 2.96
Gnu gmake : 3.79.1
autoconf : 2.57
automake : 1.6.3
ld : 2.11.90.0.8

The version of ld I'm using is older. If there are restraints on the versions
of pieces of software that people need, perhaps adding some checks to
scripts/autogen.sh is in order? Thanks.

-Blake

Whatchu talkin' 'bout, Willis?
> On Sun, Feb 02, 2003 at 03:30:29PM -0500, Blake Matheny wrote:
> > According to the documentation that I have for my linker, it says that the
> > syntax for the VERSION command is:
> >
> > VERSION { commands }
> >
> > where VERSION should not be blank.
>
> Your report is very incomplete, and thus we don't know how to fix this
> problem. More important than a patch is always an accurate description of
> the problem, and that includes the operating system you are using, and the
> version of the linker (and if it is the system linker or GNU ld).
>
> Without a VERSION, it is an anonymous version map. This is supported by
> binutils 2.13 or later, probably also by earlier versions. It should also
> be supported by Solaris.
>
> Please let us know which system and version you are using.
>
> Thanks,
> Marcus
>
> --
> `Rhubarb is no Egyptian god.' GNU http://www.gnu.org marcus@gnu.org
> Marcus Brinkmann The Hurd http://www.gnu.org/software/hurd/
> Marcus.Brinkmann@ruhr-uni-bochum.de
> http://www.marcus-brinkmann.de/

--
Blake Matheny "... one of the main causes of the fall of the
bmatheny@purdue.edu Roman Empire was that, lacking zero, they had
http://www.dbaseiv.net no way to indicate successful termination of
http://ovmj.org/GNUnet/ their C programs." --Robert Firth

Mailing List Archive

Mailing List Archive

Attached Files: