Re: FYI: duplicate removal patch [Was Re: ok, new libtool for cygwinupdates]

[Gary V . Vaughan]

On Tuesday 03 April 2001  7:04 pm, Robert Boehne wrote:
> Hello,

Hi!

>  I have been on vacation for several days, so I wasn't
> here to add to this discussion until now.
> The piecewise linking code collapses all the *.lo arguments
> to a link into a single object file, so there is some limit
> to how much it can do. In practice this solves the problem
> for the most part, unfortunately the current problem with
> this is that dependency libs are repeatedly added,
> seemingly in an infinte loop.

It is not an infinite loop.  It is just that complexities of the 
interdependencies of the libraries in a project as large as KDE conspire to 
produce combinatorially explosive chains of dependant library names.

> IMHO, any duplicate libraries should be removed on
> the platforms where it is possible.

Agreed.

Unfortunately it is a surprising difficult thing to do properly.  My current 
thinking is that we should add a new configure time variable which will turn 
back on duplicate stripping (for *shared* libs only) on hosts where it is 
known to be harmless.  I don't know how we will recognize which libraries in 
the list are shared (and subject to duplicate stripping), especially on hosts 
with pass_all deplibs_check_method =(O|

Considering it took 2 years for anyone to report the bug, but only 2 days for 
people to complain about the fix, I am beginning to think that we should 
deliberately ship 1.4 with the bug reintroduced.  We can then fix it properly 
in the next release.  Pragmaticallym, the half cocked fix in place right now 
seems to be more trouble than it is worth.  Comments?

Cheers,
        Gary.

> Robert
>
> Nick Hudson wrote:
> > Michael Matz wrote:
> > > Hi,
> > >
> > > On Sun, 1 Apr 2001, Gary V. Vaughan wrote:
> > > > > > I have applied the following to HEAD (and similar to MLB).
> > > > >
> > > > > Why also MLB?  Was it really broken there too?  I ask, because I
> > > > > _definitely_ got multiple libraries in link commands.
> > > >
> > > > Try out the new depdemo-dups.test on an old libtool script, and
> > > > you'll see what I mean.  Perhaps I have found and fixed a similar but
> > > > different bug?
> > >
> > > I see.  Argh, This then again means, that libtool will probably
> > > generate excessively large link commands for KDE.  We sometimes list
> > > also dependent libs in the makefiles (history and lazyness), and this
> > > then cumulates over many libraries.  Hmm, OK I need to check, if this
> > > really is so, but I suspect it.
> >
> > I can confirm this. Some .la files created by NetBSD KDE 2 pkgs have
> > dependency_libs entries that are bigger than 2048 characters long. ATM,
> > the buffer in ltdl.c is currently set to 1024.
> >
> > -lm and -lXmu appear far more than they need to...
> >
> > Nick
> > --
> > aka address@hidden, address@hidden
> >
> > _______________________________________________
> > Libtool-patches mailing list
> > address@hidden
> > http://mail.gnu.org/mailman/listinfo/libtool-patches

-- 
  ___              _   ___   __              _         mailto: address@hidden
 / __|__ _ _ ___ _| | / / | / /_ _ _  _ __ _| |_  __ _ ___       address@hidden
| (_ / _` | '_|// / |/ /| |/ / _` | || / _` | ' \/ _` | _ \
 \___\__,_|_|\_, /|___(_)___/\__,_|\_,_\__, |_||_\__,_|//_/
home page:  /___/                      /___/                  gpg public key:
http://www.oranda.demon.co.uk           http://www.oranda.demon.co.uk/key.asc