On Jul 28, 3:06 pm, "Rick Faircloth" <[EMAIL PROTECTED]> wrote: > We need some kind of system where plug-ins are tested and deemed > compatible with this version or that version of the core and not with > some other version. > > An automated system that checks for compatibility and dependencies.
That's not technically plausible/possible because to automatically test for compatibility the automated system must understand the following: a) Which functions belong to jQuery and which not. b) For ANY GIVEN INPUTS, what are the expected outputs. That requires knowing a hell of a lot about semantics. It is up to the PLUGIN AUTHORS to verify whether or not their plugin works with any given jQ version. That is NOT to say that plugin authors are obligated to test against every jQ release, because that's just not plausible. Authors SHOULD, however, say, "this version of my plugin was tested with version XYZ of jQuery and appears to work." Aside from that, plugin authors have done us the favour of giving us something to work with and are in no way obligated to support us beyond that. Though of course most plugin authors do support their code, my point is simply that they're not obligated to, and we cannot enforce upon them that they test with every version of jQuery. There can never be a 100% guaranty that any given plugin works in 100% of cases with a given version of jQuery, because the whole environment which jQuery lives in (web browsers) is simply too fluid and full of incompatibilities between browsers (e.g. see how many subject lines in this forum have "IE" in them). > I'm just trying to figure out a way to manage a fast-growing library > and all its little plug-in children which are multiplying very rapidly! i sympathize completely with what you're after, but i've got over two decades of programming experience which tells me that an automated tool to do what you're looking for cannot work, at least it cannot work 100% correctly 100% of the time. And, in my experience, a tool which fails 5% of the time is worse than one which fails 90% of the time because people will come to trust the 95% tool and will let the 5% of failures simply slip through without a second glance.
