Re: LDAP extended and revamped

Oscar Figueiredo writes: > >>>>> "Robert" == Robert Pluim <rpluim(a)bigfoot.com&gt; writes: > > Robert> Hrvoje Niksic writes: > >> Oscar Figueiredo <Oscar.Figueiredo(a)di.epfl.ch&gt; writes: > >> > >> > *** IMPORTANT NOTE *** > >> > Due to the function renaming, this patch introduces an incompatible > >> > change that breaks the packages using the current LDAP API, most > >> > notably EUDC and W3. > >> > >> Renaming doesn't necessarily entail breaking backward compatibility. > >> If you want to be kind to your users, define functions with old names > >> that do nothing except call the new functions. > > Robert> Wouldn't define-{obsolete,compatible}-function-alias be better here? > Robert> Or are they intended only for 'core' emacs functions? > > Of course I've been concerned by this problem. But I don't > know what to do for the particular case of ldap-search. > Currently ldap-search is the name of the high-level search > function (vs. ldap-search-internal which is the low-level > function). But according to the recent discussion on the > subject, what we want is the low-level function to be called > ldap-search and the high-level one called something else (my > patch proposes ldap-search-entries). I'm in favor of keeping compatibility with the old interface and, if necessary, accepting some awkwardness in the new names. But given that low level and high level search functions take different parameters, ldap-search could be smart and call the high level function if it looks like that's what the application wants. So maybe the problem can be avoided.