Conversation

Notices

1. Embed this notice
   iced depresso (icedquinn@blob.cat)'s status on Friday, 15-Dec-2023 02:11:42 JST iced depresso
   > the vcs is still missing features like 'blame' after >5 years development
   
   💀
   • Sexy Moon and Haelwenn /элвэн/ :triskell: like this.
   • Embed this notice
     iced depresso (icedquinn@blob.cat)'s status on Friday, 15-Dec-2023 02:16:13 JST iced depresso

     in reply to
     although i found amusement that sapling uses an sccs weave to implement the absorb command. :blobcatcoy:
     
     weaves are ... neat. i looked in to them a little when reading the bitkeeper source code. they eventually did get big enough to warrant changing formats. bitkeeper came up with a new bytecode version and they were running lz4 or something on top of it. McVoy was apparently consulting at faceboogle and I'm guessing this is what they were talking to him about.
     
     the big dumb with weaves though is you have to rewrite an entire weave file on every content update. mercurial operates with append-only logs/filesystems (hello venti) and git *can* but often doesn't because of repacking. whereas a weave commit *has* to replace an entire file, because the file *is* the entire log of all changes interwoven with control codes for the versions.
     
     the :blobcatthumbsup: with weave merges is ... they're very smart. having the histories of both files being merged together allows it to get out of failing. which is apparently why BK rarely knows what merge conflicts are while git merge/rebases are always shitting themselves.
   • Embed this notice
     iced depresso (icedquinn@blob.cat)'s status on Friday, 15-Dec-2023 02:42:58 JST iced depresso

     in reply to

     • THOT POLICE
     • Sacha Ligthert 🇳🇱🇬🇧
     @ligthert @s8n my guess is because its a google engineer's side project, and he's just playing around with some prototypes for things google might want.
     
     https://martinvonz.github.io/jj/v0.12.0/
     
     some parts of it seem neat.
   • Embed this notice
     Sacha Ligthert 🇳🇱🇬🇧 (ligthert@mastodon.nl)'s status on Friday, 15-Dec-2023 02:43:00 JST Sacha Ligthert 🇳🇱🇬🇧

     in reply to

     @icedquinn Blame is like one of these high-priority features you implement as an ass-covering method. Just why isn't it in there yet?

   • Embed this notice
     iced depresso (icedquinn@blob.cat)'s status on Friday, 15-Dec-2023 03:11:33 JST iced depresso

     in reply to

     • THOT POLICE
     • Sacha Ligthert 🇳🇱🇬🇧
     @ligthert @s8n i doubt there will be anything 'disruptive' for a long time. git has pretty much suffocated the entire environment (even though its horribly designed.)
     
     i tinkered around with 'topgit' which basically does patch queues but each patch retains its own history in a shadow branch. it looks like the big corpos have blessed patch queue lifestyle (but now its "stacked commits")
     
     monorepo stuff i'm kind of eh. bitkeeper was right, i think. monorepos are a non-goal and just handling repository nesting in a sane way gets you what you want there. bk nested isn't significantly different from git subrepos, but the tooling actually bothers to check if you are about to shoot yourself in the foot prior to allowing commits and pushes.
     
     i think, really, we just want bitkeeper with some modern polishing and a couple of extensions for the way workflows have turned out.
   • Embed this notice
     Sacha Ligthert 🇳🇱🇬🇧 (ligthert@mastodon.nl)'s status on Friday, 15-Dec-2023 03:11:34 JST Sacha Ligthert 🇳🇱🇬🇧

     in reply to

     • THOT POLICE

     @icedquinn @s8n
     Nothing disruptive, but nice to see. :-)