On Fri, Jul 20, 2007 at 04:13:34PM +0300, Bozhidar Batsov wrote:
2. By reimplementing everything there is the slight possibility that you'd do something better than it was done in the original - if it happens, it can be retrofitted there. 3. It is much more fun do it all...
I've searched a bit about other implementations, I found at least two of them : 1) libpypac + apport client, used in Enlisy distribution http://bbs.archlinux.org/viewtopic.php?id=34051 http://bbs.archlinux.org/viewtopic.php?id=14911 http://home.gna.org/libpypac/ http://enlisy.org/ - http://enlisy.com/en/ This one looks up to date. Does anyone know about technical details, how it compare to pacman ? features implemented, performance, code readability, etc. 2) Pry http://bbs.archlinux.org/viewtopic.php?id=20447 that one looks pretty much dead. Maybe there are others also, but this is apparently not the first time it has been done. About your project, I'm not sure that a dependency on mono will be appreciated for the core software of Arch. Not to mention this particular language is Microsoft's baby, so I would find that rather funny, but it may be just me :) I personnaly wouldn't mind if pacman used a higher language though, unless someone with impressive C skills rewrote it from scratch. Because I'm not a huge fan of the current codebase..
P.S. Everything would have been much easier had the format of the database been thoroughly explained somewhere. I had to guess and ask a lot before I could achieve a working state of things.
Help is welcome ?