From: Alexandre Oliva (oliva@dcc.unicamp.br)
Date: Wed Jan 13 1999 - 15:28:26 EST
On Jan 13, 1999, "Parmelan, Edouard" <EP510777@exchange.FRANCE.NCR.com> wrote:
> [ as you know if libtool fails to load the dynamic object
> it use the static one ]
Yep. I'm willing to move to the safe side and build only a static
module if there are any dynamic dependencies that can't be satisfied
for a module. This will at least get us working while we port
inter-library dependencies to other platforms.
> For libtool, how can I test that it's legal for all cases ?
Hard to tell. I still don't fully understand all the issues
involved. Gordon Metzigkeit used to have a web page explaining what
was the hard point about inter-library dependencies, and how different
systems were from one another WRT to this.
One reasonable approach is to give a try on pass_all and, if something
ever breaks, find out why and move to a more conservative mechanism.
But we shouldn't make it the default for all systems, otherwise we
risk breaking everything at once. Better have someone listed as the
ChangeLog responsible for each particular change, so that we can
contact that person and try to get him helping us finding a way to
improve ILD on each platform.
--
Alexandre Oliva http://www.dcc.unicamp.br/~oliva aoliva@{acm.org}
oliva@{dcc.unicamp.br,gnu.org,egcs.cygnus.com,samba.org}
Universidade Estadual de Campinas, SP, Brasil
This archive was generated by hypermail 2b29 : Sat Sep 23 2000 - 19:57:41 EDT