Hi, Am 12.07.2018 um 12:07 schrieb Jan Stary <h...@stare.cz>: > On Jul 11 10:27:50, ghar...@sonic.net wrote: >> I would *personally* prefer that we not have generated configure files in >> Git and require that autoconf be run (or, if it needs to be run with >> particular arguments, supply an autopen.sh file and require that it be run); >> if anybody has an argument against it, let us know. > > The ./configure script does need to be generated in the first place. > I would very much preffer we have a simple, hand-written ./configure, > and avoid the GNU auto* hell altogether, much like e.g. the extremely > portable http://mandoc.bsd.lv/ does.
As a non-Linux distro maintainer (OpenCSW for Solaris) I can tell you that „hand-written“ configure scripts are the worst you can get as every package then uses a different approach and it takes a lot of time to figure out the standard things like compiler selection, flags for preprocessor, compiler and linker, install relocation, etc. - if this is taken cared of at all. I don’t say autotools is perfect, by far it is not, but it is a standard and you need to figure out stuff only once and then reuse it for all packages that use this tool. I personally maintain ~1600 packages, with manually written configure scripts this would just not be doable. Just my 0,02€, there may be more important reasons to change the configure mechanism, but for me autotools works pretty good for most of the upstream software I package. Maybe adjusting the configure.ac to be more flexible is a netter approach than reinvfenting the wheel once again. Best regards — Dago -- "You don't become great by trying to be great, you become great by wanting to do something, and then doing it so hard that you become great in the process." - xkcd #896 _______________________________________________ tcpdump-workers mailing list tcpdump-workers@lists.tcpdump.org https://lists.sandelman.ca/mailman/listinfo/tcpdump-workers
