tech-userlevel archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: dlopen(3) and symbol conflicts



On Fri 20 Mar 2015 at 08:13:58 +0000, Emmanuel Dreyfus wrote:
> On Thu, Mar 19, 2015 at 05:31:15PM +0000, Christos Zoulas wrote:
> > Did you use dlopen() with RTLD_GLOBAL?
> 
> Yes, but that does not help: when python dlopen() api.so and loads
> libglusterfs.so, libc.so has already been loaded, and uuid_compare()
> is already resolved to libc(). 

I suppose some reasoning for the way this effectively works could have
been that if some symbol has already been resolved, it would be too
surprising if it were to be changed when somebody calls dlopen.
And I suspect this reasoning would hold both for normal and weak
symbols.

> Emmanuel Dreyfus
-Olaf.
-- 
___ Olaf 'Rhialto' Seibert  -- The Doctor: No, 'eureka' is Greek for
\X/ rhialto/at/xs4all.nl    -- 'this bath is too hot.'

Attachment: pgpEi8CNVYbSp.pgp
Description: PGP signature



Home | Main Index | Thread Index | Old Index