caml-list - the Caml user's mailing list
 help / color / mirror / Atom feed
From: "Richard W.M. Jones" <rich@annexia.org>
To: Philippe Veber <philippe.veber@gmail.com>
Cc: Gerd Stolpmann <info@gerd-stolpmann.de>, caml users <caml-list@inria.fr>
Subject: Re: [Caml-list] Compiling the ocaml distribution under multiarched Debian/Ubuntu
Date: Fri, 13 Jan 2012 14:21:08 +0000	[thread overview]
Message-ID: <20120113142108.GA1299@annexia.org> (raw)
In-Reply-To: <CAOOOohS-R4NTz5wdVzxPzjwWTEimNj-+71sLyxo9v46mRhSRvQ@mail.gmail.com>

On Fri, Jan 13, 2012 at 01:44:06PM +0100, Philippe Veber wrote:
> 2012/1/13 Gerd Stolpmann <info@gerd-stolpmann.de>
> 
> > Am Freitag, den 13.01.2012, 12:59 +0100 schrieb Philippe Veber:
> > > Dear camlers,
> > >
> > > Debian and Ubuntu have not so recently switched to multiarch binaries
> > > (including libs, see http://wiki.debian.org/Multiarch/Implementation).
> > > This is an important change for ocaml C bindings since the libraries
> > > are now to be found in /usr/lib/<arch description> instead
> > > of /usr/lib. I was just bitten by this, when realizing that the ocaml
> > > configure script couldn't find libX11.so and wouldn't install
> > > graphics. A similar problem was handled by the people in charge of
> > > debian ocaml packages
> > > (http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=619344). Now my
> > > question is the following: will this evolution be a problem for
> > > GODI/odb packages (and more generally source distribution), and what
> > > is the advised fix for it?
> >
> > Good question. For GODI it means that the automatic library search is
> > broken. You can set the library path in most cases manually, though. (So
> > far I overlook it, this is only a problem for the libs
> > in /usr/lib/<arch>/<pkg> but not for the ones in /usr/lib/<arch> because
> > the latter can be linked without -L switches.)
> >
> That's what I thought too, but I'm hurt even for libX11 which is in
> /usr/lib/<arch>:
> 
> pveber@gesundheit:/usr/lib$ locate libX11.so
> /usr/lib/x86_64-linux-gnu/libX11.so
> /usr/lib/x86_64-linux-gnu/libX11.so.6
> /usr/lib/x86_64-linux-gnu/libX11.so.6.3.0
> 
> This is related to ocaml configure script, which wants to see the lib
> before trying to compile
> 
> for dir in \
>     $x11_lib_dir          \
>     $x11_try_lib_dir      \
>                           \
>     /usr/X11R6/lib64      \
>     /usr/X11R6/lib        \
>     /usr/lib/X11R6        \
>     /usr/local/X11R6/lib  \
>     /usr/local/lib/X11R6  \
>     /opt/X11R6/lib        \
>                           \
>     /usr/X11/lib          \
>     /usr/lib/X11          \
>     /usr/local/X11/lib    \
>     /usr/local/lib/X11    \
>     /opt/X11/lib          \
> <SNIP>
>     ; \
> do
>   if test -f $dir/libX11.a || \
>      test -f $dir/libX11.so || \
>      test -f $dir/libX11.dll.a || \
>      test -f $dir/libX11.dylib || \
>      test -f $dir/libX11.sa; then
> 
> So basically it may be an issue in all cases ...

I think this is just an argument for using autoconf instead of
hand-coded shell scripts.

> > In the future we will have many errors because not
> > all package developers will test on multiarch-enabled systems.
> >
> Just curious: are there many others than debian/ubuntu?

It seems unlike that Fedora will implement multiarch like
this, since Fedora has done its own thing for a long time
(and it's broken too, but in a different way).

Rich.

-- 
Richard Jones
Red Hat

  reply	other threads:[~2012-01-13 14:21 UTC|newest]

Thread overview: 12+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2012-01-13 11:59 Philippe Veber
2012-01-13 12:25 ` Gerd Stolpmann
2012-01-13 12:44   ` Philippe Veber
2012-01-13 14:21     ` Richard W.M. Jones [this message]
2012-01-13 13:18 ` Stéphane Glondu
2012-01-13 13:56   ` Gerd Stolpmann
2012-01-13 14:17     ` Philippe Veber
2012-01-13 15:22       ` Gerd Stolpmann
2012-01-13 16:13         ` Adrien
2012-01-13 16:47           ` Gerd Stolpmann
2012-01-13 16:53             ` Philippe Veber
2012-01-14 15:06               ` Philippe Veber

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20120113142108.GA1299@annexia.org \
    --to=rich@annexia.org \
    --cc=caml-list@inria.fr \
    --cc=info@gerd-stolpmann.de \
    --cc=philippe.veber@gmail.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).