On Tue, Aug 07, 2001 at 11:23:04AM +0200, Ralf Corsepius wrote: [...] > You might want to try --trace with autoconf > 2.13. > > BTW: This type of error message in most cases indicates m4 quoting > problems. Hmm, ok I got it finally: it was hard to detect since I didn't consider what autoconf actually does ... The problem itself was a missing ']' bracket. But it was "hidden" by the fact that it was in acinclude.m4 which in turn is translated in aclocal.m4, but only if you run the appropriate tool. If you do not run the proper sequence while debugging the results are out of sync with the source files ... "autoconf >2.13" is another thing to discuss. I try to stay "cool" and don't bash the auto* authors more than necessary this time ... We have reports of systems which do not seem to be supported by libtool 1.3.5 which we decided to use for 0.93.0 (well, actually I pushed hardly to do so). libtool 1.4 already requires autoconf >2.13 (some people didn't see this; it's a fact on one system here; strange that even the authors don't have an overview all the time). autoconf 2.50 in turn breaks the vast majority of all configure.in setups of various packages on my harddisk. And everytime I ask on another project's list I here: "no, our script is not 2.50 ready" or "2.50 sucks" or similar. Also it's clear that I can neither debug all configure.in's nor autoconf itself - that's not an option. So what's left? libtool people insist on going ahead, I think 1.4 even could be built using autoconf 2.13, CVS version no longer does so. I don't like the idea of having two sets of auto* installed on my systems just to work on LessTif. Any ideas?? -- Alexander Mai [EMAIL PROTECTED]
