Arch Linux

commit 4a72c0964abc4b4836c2aafb9c59716299215802
Author: Xavier Chantry <chantry.xavier@gmail.com>
Date: Sun Oct 17 11:16:27 2010 +0200

pacman/upgrade: switch to new interface

Note that there is a behavior change here : if the same package name
appeared several times in the target list, the alpm_add_target interface
chooses the new package, while alpm_add_pkg returns PKG_DUP.

I don't see why we cannot unify the behavior of -S and -U, and just
choose one behavior that applies to both.

Otherwise, it's always possible to handle these different behaviors in
the frontend, it just requires more work.



19:59 toofishes >> the backend should allow for flexible behavior
19:59 shining^ >> so a flag or two add_pkg functions ?
19:59 toofishes >> if we want to be smart, we should do that frontend- stinks becaue ew have to look at hte current target list every time and be smart
19:59 toofishes >> and remove then add
19:59 toofishes >> you don't even need flags or options, just do it the old fashioned way i would think
19:59 toofishes >> keep the operations simple
20:01 shining^ >> so allow the feature to be implemented in the frontend ?
20:03 toofishes >> oh damn
20:03 toofishes >> there is no way to remove a package from the *add* list once you put it there?
20:03 toofishes >> i was thinking add and remove operated ont he same list, silly me
20:04 shining^ >> yeah we cant
20:04 toofishes >> maybe one more function- alpm_replace_pkg(pmpkg_t *pkg) ?
20:05 toofishes >> so we would call alpm_pkg_add, get DUP back, and then do a vercmp to see if we should call replace?
20:05 toofishes >> replace should probably add too if not found, rather than error
20:05 shining^ >> vercmp with what ? :p
20:07 toofishes >> what do you mean?
20:07 toofishes >> we have two package objects at that point (if you find the one from the current target list)
20:08 shining^ >> yeah but .. looks weird to search target for dup in pkg_add , and then do it again in frontend when it returns PKG_DUP
20:09 toofishes >> then do the search first, i don't care :)
20:09 toofishes >> i figure 95% of the time you wno't have dups so who cares
20:09 toofishes >> what happens in the backend stays in the backend
20:09 toofishes >> and we're going to make pkg_find fast anyway with allan and i's hash changes
20:10 shining^ >> I wonder if the frontend can even use that
20:11 toofishes >> of course it can, it does
20:11 toofishes >> that would be stupid if it couldn't
20:13 shining^ >> where ?
20:18 toofishes >> well i suppose we need to expose a find_pkg method, but it should be there :)
20:20 toofishes >> apparantly we did not