GNU Arch – first impressions

Something really strange happened today. I never thought that this is possible, even less that it could happen to me, but it did. CVS did something really strange to my Intranet/Webcentral repository. All the code is there, but there is no more diffirentiation between versions. Whatever you try to see, you get all the code at once. Yuck!

That happened probably because I was waiting for something like this to happen. Not like this, but, well, whatever. I’ve been constantly irritated with CVS recently. Maybe I am using it wrong, maybe it’s the wrong tool for the job, maybe something else. Anyway, I switched to using arch. It seems to be better then CVS in certain areas, so I decided to give it a shot.

Here are the first impressions:

  • No RPM found in standard places like FreshRPMs and RPMfind.net. Maybe they are there, but I was not looking hard enough.
  • Building from source is simple, but two things I hate: 1. Files like INSTALL and README start with a “=” sign. “=README” and “=INSTALL” is something that disturbes me. :) 2. configure didn’t want to start in the same directory I got after extracting the .tar.gz . I had to make another one and start configure from there. Messy.
  • Organization of archives, projects, branches and versions is much cleaner then in CVS.
  • Resource like CVS Home is either missing or not easy/fast to find. Again, I was not looking hard enough just yet.
  • Generation of GNU Changelogs is sweet.
  • …so are the patchset, although until now I just like the idea, but have no practice. I can see that it’s possible, just haven’t got myself that far yet.
  • Distributed repositories and ease of setting one up using http/ftp/sft/etc blows my imagination. Simpliest so far. Mine is over sftp. :)

I’ll be writing more code in all the different place for the next few days, I guess.