[pacman-dev] [GIT] The official pacman repository branch, master, updated. v3.1.4-163-gfb09d35

[Xavier]

Nagy Gabor wrote:
>> commit f43805d875ad5c672afbbfff48bded2087204773
>> Author: Chantry Xavier<shiningxc at gmail.com>
>> Date:   Sat May 10 18:47:42 2008 +0200
>>
>>      Cleanup usages of alpm_list_find and alpm_list_remove.
>>
>>      * remove obsolete and unused *_cmp helper functions like deppkg_cmp and
>>      _alpm_grp_cmp
>>
>>      * new alpm_list_remove_str function, used 6 times in handle.c
>>
>>      * remove _alpm_prov_cmp / _alpm_db_whatprovides and replace them by
>>      a more general alpm_find_pkg_satisfiers with a cleaner implementation.
>>      before: alpm_db_whatprovides(db, targ)
>>      after: alpm_find_pkg_satisfiers(alpm_db_getpkgcache(db), targ)
>
> Warning: pkg literal also satisfies pkg. But in most cases we called
> what_provides if we didn't find a literal.
>

I know, it's not exactly equivalent but I think it makes more sense that 
way and as you noticed, it works the same for our use case.

>>      * remove satisfycmp and replace alpm_list_find + satisfycmp usage by
>>      _alpm_find_dep_satisfiers.
>>      before : alpm_list_find(_alpm_db_get_pkgcache(db), dep, satisfycmp)
>>      after : _alpm_find_dep_satisfiers(_alpm_db_get_pkgcache(db), dep)
>
> Warning: possible slowdown, the old way just stopped after a satisfier (which
> is ideal in checkdeps), now we scan the whole db.
>

Right, I knew about that too, I just wanted to keep the code as clean as 
possible and didn't find another way. Though it might be worth to do 
some benchmarking / profiling. If it's really too bad, it will have to 
change.

>>
>>      * remove _alpm_pkgname_pkg_cmp, which was used with alpm_list_remove,
>> and
>>      use _alpm_pkg_find + alpm_list_remove with _alpm_pkg_cmp instead.
>>
>
> Imho this is ugly. First we find it, then we again find it via list_remove.
>
>

Yeah, it's not ideal either. But neither are dummy pkg or fake 
asymmetric cmp functions. I just preferred it like that.
Imo the real problem here is that our data structures suck and are 
inefficient. Linear search ftw.


Anyway, other better suggestions for these 3 points are always welcome.