Wrong use of -export-symbols in src/Makefile.am

LRN lrn1986 at gmail.com
Thu Mar 21 16:10:30 CET 2013

Hash: SHA1

On 21.03.2013 18:39, Werner Koch wrote:
> On Thu, 21 Mar 2013 12:28, lrn1986 at gmail.com said:
>> The reason is, i believe, that EXPORTS is not on the first line.
>> If it's moved to the first line (above the comments), then
>> libtool groks it correctly.
> Are we talking about master or 1.5?
Both. I've had this problem with 1.5, and i still have it with git
master HEAD.

> I tested 1.5 and it worked without problems.
Which version of libtool did you use? Did you cross-compile (not sure
how that would affect libtool behaviour though...)?

> Is your plan to port master to Windows?
libgcrypt does not require porting (which is good!). It only has some
small hiccups in the buildsystem: .def file EXPORTS placement, libdir
creation, and fig2dev detection - that's all.
It even passes all the tests (i have the output of `make check', if
you need it).

> I would like to postpone this until a release.
As with everything else i've been pushing your way, i'm perfectly
happy to apply my patches locally, until you fix the problems on your

> It is just to much work to fix minor Windows incompatibilities
> during development.
Of the problems listed above:
* the fig2dev issue does not affect tarballs, since they come with
pre-generated image files,
* libdir creation only affects package maintainers (i.e. when
installing into empty staging directory and with -jX).,
* EXPORTS placement issue is the only one that will screw people
compiling from git AND from tarballs. Fix that, and, hopefully, 1.6.0
tarballs will come out fine.

> I don't test it anyway.
That's bad.

Version: GnuPG v1.4.11 (MingW32)
Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/


More information about the Gcrypt-devel mailing list