Re: svn commit: r1935862 - in subversion/branches/svnsync-atomic-author/subversion: include include/private libsvn_ra_local libsvn_ra_serf libsvn_ra_svn libsvn_repos mod_dav_svn svnserve svnsync tests/cmdline tests/libsvn_repos

Jordan Peck <[email protected]>
Newsgroups gmane.comp.version-control.subversion.devel
Message-ID <pony-f79afb3b-16af-4093-90d6-a46e1d30f565-dev@subversion.apache.org>
Sorry about that, I did review the commit message and made changes to the llm output (which is why there are 2 versions of the message in the log). I edited/reviewed the message in my IDE then copied it incorrectly into the SVN commit window. I've amended the log message.

On 2026/07/04 11:17:20 Branko Čibej wrote:
> On 4. 7. 2026 12:59, Evgeny Kotkov via dev wrote:
> > Jordan Peck<[email protected]> writes:
> >
> >>    Read 1 file
> >>
> >> ● Update(subversion\libsvn_repos\fs-wrap.c)
> >> Added 1 line, removed 1 line
> >>                                     svn_repos_t *repos,
> >>                                     svn_revnum_t rev,
> >> _t *pool)
> >> {
> >>    apr_array_header_t *revprops;
> >>
> >> ● Update(subversion\libsvn_repos\fs-wrap.c)
> >> Added 1 line, removed 1 line
> > …
> >> Fix 5c — state commit_callback nullability on editor6 directly
> >> (replacing the stale pre-1.6 parenthetical):
> >>
> >> ● Update(subversion\include\svn_repos.h)
> >> Added 1 line, removed 2 lines
> > …
> >> You've hit your session limit · resets 9:50pm (Europe/London)
> >> /upgrade or /usage-credits to finish what you’re working on.
> >>
> >> ✻ Baked for 5m 21s
> >>
> >> ❯ Can you put together a brief commit message for this change
> >>
> >> Here's a commit message in the Apache Subversion log-message convention
> >> (summary paragraph plus per-file annotations), ready for svn commit -F:
> > Looks like the commit message accidentally picked up part of your Claude
> > session, not just the generated message itself.
> 
> 
> I'm going to be less diplomatic. We know how LLMs are prone to 
> "hallucination". And yet you didn't bother to review the commit message 
> it prepared for you. How do we know that the code wasn't generated with 
> a similar attention to detail?
> 
> This is open source. You're never in too much of a hurry to get things 
> committed. We prefer ... or rather, I prefer quality over quantity. 
> Please do better next time.
> 
> -- Brane
>
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.