On Wed, 02 Mar 2005 20:34:58 -0500, Aaron Bentley <[EMAIL PROTECTED]> wrote: > Tom Lord wrote: > > Will you please elaborate? Without referring to old conversations > > we had long ago: what is wrong with the `--forward' option? > > Its granularity is too large. It either applies the entire patch, or > ignores the entire patch. It doesn't just ignore already-applied hunks. > > I'm going by what ddaa said. I haven't tested it, but I thought there > was consensus.
That's my memory too. I think it really, really, sucks that patch works this way, but it does. I can only imagine that when this option was added, many, many years, ago, it was assumed that people would examine the results --forward to make sure things were OK afterwards -- for instance, note that it saves the omitted patches in .rej files. I will note that in my non-arch usage, I've used patch --forward quite often, and never been bitten by this behavior, but it's entirely possible that I've simply been lucky, and I agree that it's just too dangerous to use in a system where people want to complex and merges. I suppose if tla is going to continue to use the standard patch command, it could do some sort of horrible kludgey thing with --forward by examining the results of patch, splitting up --forward-generated .rej files and trying to reapply them..... gack. -Miles -- Do not taunt Happy Fun Ball. _______________________________________________ Gnu-arch-users mailing list [email protected] http://lists.gnu.org/mailman/listinfo/gnu-arch-users GNU arch home page: http://savannah.gnu.org/projects/gnu-arch/
