[prev in list] [next in list] [prev in thread] [next in thread] 

List:       ruby-talk
Subject:    [ruby-talk:00159] Re: inability to load extension modules in 1.2, c
From:       Clemens Hintze <c.hintze () gmx ! net>
Date:       1998-12-29 9:49:05
[Download RAW message or body]

On 27 Dec, Julian Fondren wrote:

Okay now the bad news! I don't find any fault. I think the warnings on
the end of the compilation stage can be ignored. But the missing symbol 
__GLOBAL_OFFSET_TABLE_, indicates for me that whether it was forgotten
to link a certain lib to ruby, or your system may be corrupt. I feel
strange, that the symbol begins with "__" and ends only with "_".

I have a Linux system not a BSD system like you. But I nevertheless
think, that there is something wrong with your system. There is no
symbol beginning with "__" and ending with "_" in all shared libs of my
system. So you should ask your system, who requests that symbol.

I have used following command line to search for such symbols:

nm /lib/*.so /usr/lib/*.so /usr/local/lib/*.so \
   /usr/X11R6/lib/*.so                       | \
ruby -ane 'print "#{$F[2]}\n"'|grep '^__.*[^_]_$'

[...]

A further strange think I have found is, that in the linking line there
are several libc linked in. But sometimes there is a -L/usr/local/lib
prepended. Do you have a libc in /usr/local/lib, so that perhaps the
linker will link in two different libc?

Furthermore you link in a -ltk80 (strange too, BTW; that should be the
name in Windows95, under UNIX it should be -ltk8.0?!?) but you only
link in a -ltcl which should not be called -ltcl8.0 or -ltcl80????

As I have no BSD Unix, I cannot help you further, sorry. In my system
all works well regardless if I use dynamic loadable extensions or static
ones.

I hope others can help you.

> cya,

Ok,
Cle.

[prev in list] [next in list] [prev in thread] [next in thread]