Corinna Vinschen wrote:
> If I'm looking over a package for inclusion I'm currently accepting > two styles: > > package-ver-subver/ > ... Both "style 1" and "style 2" in my original email obey this. The difference is that "style 2" packages -- gcc, binutils, make, etc -- don't have package-ver-subver/CYGWIN-PATCHES/a-patch in fact, they don't have 'a-patch' at all. They are, in effect, forks of the antecedent project. There is no way, given just gcc-2.95.3-5-src.tar.bz2, to "revert to the 'original' source" -- short of also downloading the 2.95.3 source from www.gcc.org, unpacking both, and doing 'diff -r cygwin-version-of-gcc gnu-version-of-gcc'. Granted, new packages should never be style 2. But style 2 is in use. > or > > package-ver-subver.patch > package-ver-subver.sh > package-ver.tar.[bg]z[2*] <-- The pristine source > > Can we agree to use and document only these styles? The question is, should I document all styles in use, or only those styles which are acceptable for new packages? I could argue either way. --Chuck