A Good Resource for Learning XUL & Javascript?

RJabelman asks: "I'm trying to write a Mozilla extension, but I can't find a decent resource to learn from. Tutorials abound for packaging up an extension, and the web's littered with Javascript snippets to pretty up a web page, but there seems to be very little authoritative information for doing serious work with Javascript, XUL and Mozilla (and more specifically, manipulating XML). I can find my one true resource for every other language or API I've learned: but not this. Can anyone point me to theirs?"

Comments

LAME

[drfrog]

propietary web browser language

need i say more?

Re:LAME

[dn15]

so if its not proprietary how does one using IE or opera or whatever view these web based apps?

That's the thing, they're not supposed to because we're not talking about making web-based apps at all. We're talking about creating software within a framework that also happens to be host to a web browser. You could write something that extends the browser's functionality but you could also write something that has nothing to do with browsing at all and doesn't even need to access the Internet. If appropriate you could spin it off as a separate app that does not integrate browsing functionality, like the Sunbird project.

Re:XulPlanet

[Jerf]

On the down side, welcome to the land of magic and wonder, where arcane bugs haunt the long forgotten planes of DOM...

Yeah, I dug into Mozilla development gung ho over the past year or so, and I've completely abandoned it... well, I'm in the process of creating a replacement XBL, but after that's done, I'm out. In painfully slow succession, I've tried in a serious way Moz's RDF support, XUL, and XBL, and in every case I have rapidly exceeded their capabilities (to the tune of segfaults) and gone back to JS + DOM. (For RDF, I now dump out my data as a JS file and use XMLHttpRequest to retrieve it and "eval" to run it; this is what Google Suggest uses. I beat them by about a month, I know others beat them by more :-) but it works and you'll be stunned how fast it is, even for many kilobytes of data; by far the fastest solution I've found, even independent of the fact that loading any sort of XML in Moz causes a huge stall of some kind and tends to cause crashes and memory leaks in nothing flat.)

I'm preparing a series of half rants, half detailed indictments on why the Mozilla specific technologies are not just poorly implemented at the moment for any task other than being a web browser, but why Javascript + DOM is usually, on the balance, a superior solution. ("On the balance" means that while my JS implementation of XBL is neither a subset nor a superset of "true" XBL, on the whole the benefits level out in favor of my JS implementation... and if I were willing to go pure Moz instead of cross-platform it would be a total win.) I don't want to repeat them here in toto, but, well, that's actually the basic argument: Me, a single schmoe, can replicate most of XBL in a couple of weeks, in Javascript, and it is actually much more reliable too, for reasons that will only make sense if you used Moz's XBL support for anything serious, like widgets that can load remote data or include other widgets in interesting ways. How much time has been spent on XBL, which is still behind?

Unless you need a XUL widget like "popup" that has no good HTML replacement, you're just better off with JS and DOM. Most people don't understand how powerful JS really is, and I've found it to be surprisingly speedy, too.

(To show I'm not just spouting off randomly, here is my current XBL in JS implementation. Still in development, but it oughta show I'm serious about this, and even now I'm finding it more pleasing to work with overall than real XBL. What stops me from releasing the rants right now is, well, there's some writing to be done yet, but instead of just bitching I want to have some constructive solutions as well; xbl.js is a big part of that, and right now I'm working on the POPUP element because I need that for my app. Ranting is great but I find they are even more powerful when they are not purely negative.)

Details forthcoming at a later date, but next time you're reaching for XUL or XBL, if it isn't for a Mozilla extension, stop and make sure you don't really want to do it in cross-platform JS + DOM. See, the thing is, those libraries are well tested and optimized in a variety of situations; I'm not encountering bugs hourly like I felt like I was in Moz.

Finally, to preempt some of the obvious responses, I'm not saying XBL or something was a bad idea; in fact the idea is so good I'm re-implementing it. I'm saying the implementation right now is so dodgy it isn't worth playing with when there is another less cool, but more functional, alternative available today in the form of JS + DOM.... and what advantages XUL or XBL have over my JS implementation are only a few small hooks away from being exposed to the JS as well.

Re:OReilly book

[jalefkowit]

Yes, there is. And it sucks.

I bought that #*#$@ing book when it first came out. And proceeded to waste prodigious amounts of time trying to make even the simplest code samples from it work. It turned out that the code samples were riddled with typos that made them un-runnable.

The book is, quite honestly, a waste of your time and money. Maybe they've since gone back and fixed the errors, but considering that it's still in the 1st edition (which targeted Mozilla 1.0 -- god only knows how much has changed between that and Firefox 1.0) I seriously doubt it.