|
From: | Sander Niemeijer |
Subject: | Re: Using libtool 2.0 in autoconf tests |
Date: | Mon, 22 Nov 2004 14:11:41 +0100 |
On maandag, nov 22, 2004, at 12:05 Europe/Amsterdam, Peter O'Gorman wrote:
Ralf Wildenhues wrote:C'mon Gary, two questions: is it *possible* to provide the old behavior without too much pain? Would that destroy some cool abstraction or somereally fundamental thing? Or are you just waiting for a patch to do this? (ok, that was three questions now).I would approve a patch which added LT_TRY_LINK or some such macro, and created a temporary configure time libtool script to do so. I think it would be a fairly hefty patch though. There is certainly enough information available to allow this as soon as the LT_INIT macro is complete, though there will be issues with quoting, as things are quoted one extra time because of the intervening config.status.
If such a macro ever gets created it is essential that it is possible to chose the kind of target one wants to link (i.e. program, library (static/shared) or module). The whole reason why I had to step away from the autoconf AC_CHECK_LIB macro is because I needed to create a shared library instead of a program (otherwise, my libtool 1.5.x solution would just have been to replace CC and LD temporarily with the appropriate libtool commands and call AC_CHECK_LIB).
The issue is a real one, and should be addressed. I am not convinced that it need be addressed for libtool-2.0 though, nor that it requires reverting to libtool-1.5 behavior.
Don't tell me you are saying that I won't be able to use libtool 2.0 unless I create such a LT_TRY_LINK myself (or follow the lt_ variables/ltmain.sh approach). This rather shatters the hope that Ralf was trying to give me.
Best regards, Sander
[Prev in Thread] | Current Thread | [Next in Thread] |