[pacman-dev] [patch] hide pmdepend_t behind alpm_depcmp

[Dan McGee]

On 10/24/07, Xavier <shiningxc at gmail.com> wrote:
> On Wed, Oct 24, 2007 at 03:13:33AM +0200, Nagy Gabor wrote:
> > The following may be surprising: I don't really like this patch.
> > I started to work on that, and after I had done ~50% of the job, I
> > realized that it is not so good as I expected. But when you did ~%50%
> > of the job you won't stop and drop away the result... (see Concorde
> > -- but sometimes this is the best decision) ;-)
> > My main goal was to simplify the code. This goal was (imho) reached.
> > BUT there are other things I don't like in pacman code, too:
> > For example, functions share too few results. And my patch joined
> > alpm_splitep + alpm_depcmp -- which is not good. Before my patch the
> > depcmp-caller split the dependency, then it could call alpm_depcmp 5000
> > times on the splited dep (upgrade-requiredby fro example). After my
> > patch alpm_splitdep is called 5000 times too, which is totally needless.
> > But at least I found some small bugs while I worked on that; so I
> > got some benefits too ;-)
>
> I like the simplification it gives, but we indeed lose flexibility in the
> same time. That may not be a big deal however.
>
> But if your patch is rejected, then you should probably also provide one
> separate patch with only the small bug fixes.

Either Nagy or Xavier- it would be awesome if you could split this up
as best as possible if it is good for inclusion. Thanks!

-Dan