On Tue, Jun 7, 2011 at 6:38 PM, Allan McRae <al...@archlinux.org> wrote: > On 08/06/11 08:29, Dan McGee wrote: >> >> On Tue, Jun 7, 2011 at 5:19 PM, Rémy Oudompheng >> <remyoudomph...@gmail.com> wrote: >>> >>> On 2011/6/7 Dan McGee<d...@archlinux.org> wrote: >>>> >>>> -alpm_list_t SYMEXPORT *alpm_find_grp_pkgs(alpm_list_t *dbs, >>>> +alpm_list_t SYMEXPORT *alpm_find_grp_pkgs(pmhandle_t *handle, >>>> alpm_list_t *dbs, >>>> const char *name) >>>> { >>>> alpm_list_t *i, *j, *pkgs = NULL, *ignorelist = NULL; >>> >>> I don't see where the handle is needed in this function. >> >> It's needed in the subsequent patches that move pm_errno to the >> handle; I tried anything to make that one smaller and this was one bit >> that could be split out. Barely worth it, I know, but trying every >> little bit... >> > > It that for subsequent patches in this patchset or future patches? At the > end of the current patchset, I find these two functions that do not use the > handle argument:
Bagh, looks like I'm eating my words here. * alpm_find_grp_pkgs() needed the handle for the QUESTION() call, but as it is in a loop, we can get that off of the db (or the pkg) for that matter. So no need for the handle. * alpm_pkg_should_ignore() looks like one I missed from my original rough draft; as above the handle can be grabbed from the pkg object. I'll fix both of these up to their old signatures. -Dan