2009/7/30 Alexey Zakhlestin <indey...@gmail.com> > 2009/7/30 Johannes Schlüter <johan...@schlueters.de>: > > On Thu, 2009-07-30 at 00:08 -0300, Guilherme Blanco wrote: > >> What do you think about the possibility to support ArrayObject > >> instances in array_* functions? > >> If you all agreed on this, I can definately help to complete the > >> patch, but I need some initial guidance to finish at least the first > >> function. > > > > In general I'm for bringing Iterators/ArrayObject/.. on a line with > > arrays but please don't simply patch some functions but let's try to > > find a as consistent as possible global solution. > > > > Questions included there are of these kinds (just examples) > > - is that specific to ArrayObject or do we "need" interfaces like > > "Sortable" or "Shuffable" > > - What should stuff like aray_merge(ArrayObject, array) do > > - Should we do this globally? (should ldap_set_option() > > allow an ArrayObject as 3rd param?) > > > > Especially the latter shows that a proper solution might be not to patch > > individual functions but to introduce some form of abstraction from > > HashTables for "array operations" and then use that interface. > > > > Simply extending a few functions will end in a mess but looking around > > people seem to love these structures so improving them is good. > > agreed. > the proper solution would be to define several interfaces, make > functions accept objects of these interfaces and then add some > "hacks", which would pretend, that array() implements these. >
I would vote to extend this concept to user-land functions as well if at all possible. It would make my life easier to not care about the implementation of a set of data so long as it conformed to an interface. -- -Nathan Gordon If the database server goes down and there is no code to hear it, does it really go down? <esc>:wq<CR>
