Adds patch-rpm support to rpm. rpm4 port not really well tested...
Created attachment 93983 [details] patch-rpm patch
The patch looks OK, but there is more to do to integrate with other features in rpm. I am still philosophically opposed to "patch packages", as I think they are overly complicated and obscure for insufficient gain. I.e. I don't believe that most packagers can figure out how to release a patch package that upgrades robustly everywhere. NEEDINFO so I don't have to see the bug, while I continue muddling what to do. At minimum, I'm willing to add patch somewhen underneath compile time option, default off, somewhen so that there is at least a reference rpm source code base.
*** Bug 131768 has been marked as a duplicate of this bug. ***
I once had to download a 300 MB openoffice RPM to fix a one-line problem in the /usr/bin/ooffice shell script. That alone should be sufficient motivation for patch RPMs.
Nowadays I prefer delta rpms, as applying the delta will get you a bytewise identical rpm, they are thus less error prone. The disadvantage of delta rpms is, of course, that a delta rpm is based on exactly one version, whereas patch rpms can be based on multiple versions.
for comment #4: Speeding up downloading is no argument for patch/delta rpms. For starters, servers everywhere would need to generate patch/delta rpm's, not going to happen imho. Furthermore, using rsync would have speeded up your download without the complexity of patch/delta rpm's. A local template of OO.o rpm's (if not already available) could have been generated by erasing and repackaging OO.o. rpm payload's have been prepared with the equivalent of gzip --rsyncable switch for quite some years now.
Comments on whether to apply the patch or not requested at https://lists.dulug.duke.edu/pipermail/rpm-devel/2007-April/002240.html
Current feedback says the patch packages are unreliable and unnedded. WONTFIX
Yeah, agreed also on rpm-maint: https://lists.dulug.duke.edu/pipermail/rpm-maint/2007-May/000317.html