[EMAIL PROTECTED] (Karl Berry) wrote: > The only reason maintainer-clean typically has not removed files like > configure and Makefile.am is that they have been (until relatively > recently) version controlled. > > That doesn't seem quite right to me. Makefile.am is typically
Whoops. Thanks. Of course, I don't mean Makefile.am, in spite of the number of times I wrote it. Only files created as part of ./bootstrap, configure, make, etc. operations should be remove by "make maintainer-clean". > version-controlled and always has been; furthermore, configure is nearly > always a generated file and may or may not be version-controlled by design. > As far as I can see, the issue isn't version control of a given file but > whether a provided "maintainer tool" can reinstate it. Exactly. > In general, I agree it would seem desirable for maintainer-clean to > delete everything which the package's bootstrap script (if any) can > recreate, now that bootstrap scripts are in common use. But we must > recognize that not all packages (I venture to say the vast majority, in > fact) use your particular fancy bootstrap script. > > So it seems to me the rule shouldn't be "always delete configure and > Makefile.am", but rather "delete everything which is recreated by a > maintainer bootstrapping the package". If there is no bootstrap script > at all, the current definition of maintainer-clean would apply. > > This could be implemented by options to automake/autoconf/gnulib-tool, I > guess. But I know I don't want maintainer-clean in Texinfo to go > deleting my Makefile.am's. :-) > P.S. Ok, "bootstrap" isn't technically the right term for all this, but > you know what I mean :). Ideas for a better name?
