From kde-core-devel Tue Jun 14 13:04:37 2005 From: Josef Spillner Date: Tue, 14 Jun 2005 13:04:37 +0000 To: kde-core-devel Subject: Re: Build system for KDE4 Message-Id: <200506141504.38587.spillner () kde ! org> X-MARC-Message: https://marc.info/?l=kde-core-devel&m=111875431819188 Am Dienstag, 14. Juni 2005 14:38 schrieb Guillaume Laurent: > Honestly, I can understand people arguing on which solution is best to > replace the autotools, but if you feel you first need to establish that > they need replacement, you're in serious need of a reality check. Most of the discussion blurred the line between concepts and implementation. For instance the automake syntax is more or less as easy as it can get (including the am_edit extensions), and unsermake has shown that it is possible to implement a fast parser and makefile.in generator. Now that leaves the configure checks, but replacing those will have to be thought of carefully as well, including a restructuring of the kde autoconf macros to something like: - check for Qt (incl. version) - check for KDE (incl. version) (the 2 above are the only relevant macros for 3rd party projects) - check for python/ruby/perl/xxx-lang-dev - check for libraries used in kdelibs - check for module libraries (like libxine in kde-mm) There's no use in starting a replacement when afterwards we still have python checks in 5 modules, openssl checks in 2 and big-endian-little-endian in 25, all without documentation and without reusability. Josef