Patch-review changes

Andreas Pakulat apaku at
Sun Feb 21 18:12:52 UTC 2010

On 21.02.10 17:44:23, David Nolden wrote:
> Am Sonntag 21 Februar 2010 17:07:34 schrieb Andreas Pakulat:
> > Well, I've forked the svncpp lib from rapidsvn, and the updating I did
> > recently cost me 30 minutes. But I never changed their original code, I
> > only added things, so that made it a lot easier. Spending an hour just
> > to update to a newer version of libxyz is something that nobody will
> > really do unless there's a really really important thing. If instead it
> > takes only 5 minutes, somebody might do it on each patch-level release.
> Which in this case is not even wanted. The "diff" file format hasn't changed 
> for many many years, and as long as we don't see issues in libdiff2 that are 
> fixed in a newer version, we simply don't need to update it. Any update just 
> brings the danger of new issues.

The same may be true for libdiff2, even if the diff2 format doesn't
change a lot. The obvious example is support for prefixes as git
produces them (I can even imagine throwing a diff at patch-review which
has two different-length paths as prefixes). I don't know how often
libdiff2 is changed, but apparently its more often than once in three
years, else there would've been no difference between our clone (without
your changes) and the original one.

Anyway, as you said, case closed.


You have an ability to sense and know higher truth.

More information about the KDevelop-devel mailing list