Makes sense. I unfortunately don't have cheap access to AI. So may be best done by people who get "free" access via either work or academia etc.
G.W. Haywood wrote at about 15:45:59 +0100 on Saturday, October 3, 2026: > Hi there, > > On Fri, 2 Oct 2026, backuppc@kosowsky... wrote: > > > Have you let AI loose on the BackupPC codebase > > No. Others have, at least on parts of it, for example > > https://github.com/backuppc/rsync-bpc/issues/50 > > which at the time I found astonishing. Now I'm getting used to it. > > > to: > > - Identify bugs & vulnerabilities > > - Improve robustness, logging, etc (wherever you think there is a lag) > > - Identify speed/efficiency improvements - including possibly moving > > other code to C to make it faster (if it would give a significant speedup) > > All the areas you mention could fruitfully be investigated, and for my > own purposes in particular I would like to improve some of the logging > and log configuration as thesedays I seem to be spending time that I'd > rather use elsewhere just tweaking log output. Hopefully now that I'm > a lot more familiar with things that won't be so much of an issue for > me, but for others who come along after me I've already felt the pain. > > BackupPC version 4 uses Perl::XS to do most of the heavy lifting, so I > think we're well into diminishing returns on speed improvement effort > just by converting Perl code to C. Having said that, there's a Github > issue here > > https://github.com/backuppc/backuppc/issues/390 > > which could turn into a nicely self-contained project for someone who > wants to try out their own or their AI's skills. Be warned, however, > that it's likely to be a very intrusive patchset and it would need a > lot of testing before being released into the wild. The ultimate hope > would be to provide significant improvements both to speed and storage > capacity virtually for free. Obviously whatever code comes out of the > exercise would need for the foreseeable future to be able to recognize > and handle files that are compressed either with Zlib or with Zstd. > > The two most important parts that use C extensively are rsync-bpc and > BackupPC-XS. Most of rsync-bpc has already been scrutinized fairly > thoroughly by AI as part of the recent rsync efforts but there's still > a bunch of code in there that's specific to BackupPC and in some areas > of which I have niggling doubts. If anyone is willing to put in some > effort prompting their pet AI to look at this stuff I'd be more than > happy to look at the results. > > -- > > 73, > Ged. > > > _______________________________________________ > BackupPC-users mailing list > [email protected] > List: https://lists.sourceforge.net/lists/listinfo/backuppc-users > Wiki: https://github.com/backuppc/backuppc/wiki > Project: https://backuppc.github.io/backuppc/ _______________________________________________ BackupPC-users mailing list [email protected] List: https://lists.sourceforge.net/lists/listinfo/backuppc-users Wiki: https://github.com/backuppc/backuppc/wiki Project: https://backuppc.github.io/backuppc/
