Strange commit

Ulrich Spörlein uqs at
Fri Apr 10 20:35:36 UTC 2020

Am Do., 9. Apr. 2020 um 21:54 Uhr schrieb Ed Maste <emaste at>:
> On Thu, 9 Apr 2020 at 15:39, Ryan Stone <rysto32 at> wrote:
> > On reflection, I believe that what as happening was that the converter
> > was using newly created files as a really bad heuristic for merges so
> > that it would not have to try and understand mergeinfo.
> Ah, ok - so that's consistent with what I've inferred from looking at
> a number of vendor updates.
> Yang (one of my co-op students) is currently working on a tool to
> parse svn log / svn diff / svn propget svn:mergeinfo and emit a list
> of merge-to/merge-from revisions.

Cool stuff, if he can do it in Qt/C++, then we can pull that directly into the

I've frankensteined a variant, where the places with a mergeinfo change
will at least dump an svn diff output (as I'm too stupid to understand the
SVN/APR API nonsense).

and followup commits.


More information about the freebsd-git mailing list