--- "Alfred M. Szmidt" <[EMAIL PROTECTED]> wrote: > Fixing > broken Makefile.am's > that don't use DESTDIR is also a good task for > someone who is is > looking for something to do.
This is what I really feel as important, and what I am trying to do since some time (with a dramatic slowdown recently for my new fulltime, time-sucking, but life-sustaining proprietary software development work). There are really a lot of possibilities that would open up if, at least for GNU projects, build systems really converge around the GNU standards. I offered help on -discuss to any GNU project to make the move towards the standards, which lead to many very positive responses, fixes, and new build systems, and also some negative responses [and probably even more ignores]. Summarized, current state is that most GNU projects support, at least, ./configure --prefix=PREFIX make make install and most also support DESTDIR. When all above is supported, more or less small shortcomings in the build system cause sometimes the package to not really be --prefix-able, or to be installable only as root, or to support DESTDIR only partially, [...] Read more here if interested: http://www.gnu.org/software/sourceinstall/gnutest.html [warning: not so recent] This test could be of course automated and connected to the release process, as it requires otherwise a daunting amount of time to perform/maintain manually, as I experienced first hand. Many other 'management' tasks could be made much simpler if we really prepare all our packages in a standard way (something similar to sourceinstall, or even simpler, could be used as a validation tool). > Please help improve the Hurd, GNU Mach, and other > missing and broken > bits, this is far far more important. > A system that > crashes while > compiling it self is not useful! I have looked with new interest at the Hurd, and I think there are great ideas there, and I find some of the proposed ideas here on -system-discuss intriguing. For example, getting rid of the PATH, LD_LIBRARY_PATH and similar variables through removal of the need for different prefixes is something I dubbed as crazy at first, but now recognize as possible (with effort) and desiderable. I am currently not knowledgeable enough to be able to really help in Hurd development, and the time at my disposal for learning aggressively is shrinking day by day. I would love to, but at the moment I have to restrict my contributions to less daunting tasks. Back to usual lurking I guess. Claudio Chiacchiera con i tuoi amici in tempo reale! http://it.yahoo.com/mail_it/foot/*http://it.messenger.yahoo.com _______________________________________________ gnu-system-discuss mailing list [email protected] http://lists.gnu.org/mailman/listinfo/gnu-system-discuss
