On Fri, Nov 30, 2007 at 08:39:22PM +0100, Nagy Gabor wrote:
OK, but I leave the benchmark for you;-) Note: Long "dependency paths" like in smoke001.py may cause notable slowdown. I made a quick "hack", so probably I made some mistakes...
Bye
Well, resolvedeps is also quite hard to read (but seems OK). Imho we should share some code between alpm_sync_addtarget and resolvedeps: I mean they could use common search_for_literal(), search_for_satisfier() functions. By doing this alpm_resolvedeps would be much clearer and '-S provision' would be more general, user could do -S 'depend>=2.0' if he wants to do so (he got an unsatisfied dependency error after -A) and we can slowly kill whatprovides... (foo provides bar is a special case of foo satisfies bar)
On this subject, see http://bugs.archlinux.org/task/8763 .
Well, I wanted to ask this earlier: is it possible (and acceptable) to "forward" pacman FS activity to here automatically? (/me never reads FS) Back to 8763: Imho this is a bit off here. But my opinion hasn't changed since -Sg: user should know and define what he wants to do. So my vote: "-S literal" only, if this wasn't successful, pacman can ask the user whether he wants to search for providers etc... And a special option to install provider (or satisfier). Bye