ftrain
Parse error: http://techcrunch.com/2012/12/30/im-bored-whats-next/
TechCrunch doesn't seem to parse at all. None of its sentences can be shared on Twitter using the bookmarklet. This might be a feature.
I've tried running this code fragment on several sites:
jQuery(document).ready(function(){console.log('foo');})
GitHub, The Atlantic, and others log "foo" to the console. TechCrunch does not. For some reason, jQuery is not executing the ready callback on that site, and SavePublishing depends on this feature of jQuery.
I'd do more digging, but there's no time.
This is an awesome step to have taken, thank you. The issue is most likely going to be in the order of execution of on those pages, and there's something blocking my jQuery (which I reassign to the global JQ) from executing. Also, something may be loading after the bookmarklet.
The script should likely fail with an error if it can't find the JQ variable.
There's an uncaught TypeError throw in the TechCrunch source, on this line of code:
new techcrunch.module.PostDetail("#module-post-detail");
If you set a breakpoint before this TypeError and inject the undefined object, you can prevent the exception and allow SavePublishing to load.
Obviously this is not really a solution, but it demonstrates that an uncaught in exception in an earlier ready callback will prevent SavePublishing from running.
I was fiddling with this today just because I was curious, and I wrote a little scriptish user script to catch and handle all the TechCrunch page load errors. And that works -- it gets the test url here loaded without any JS errors and document.ready functions afterward and everything, but SavePublishing still doesn't work. All the console.log statements fire up to "Evaluating: run.coffee" but nothing happens on the page itself. Essentially it looks the same as if I hadn't done anything. My sense is that the errors during the page load aren't the problem, or if they are, it's not in a very straightforward way.
I don't know if any of this is useful but writing it down here at least means I didn't totally waste my time.
So I forked this morning and fiddled more aggressively, and I got this page to work, but how I did it might be crazytime and unacceptable. A bunch of debugging showed that indeed for whatever reason, document.ready wasn't firing on this page. So I removed JQ(document).ready -> in run.coffee, so all the stuff that would have been run by it would just go without waiting for a ready to fire. And it appears to work just fine.
I'm posting this here rather than just submitting a pull request because I don't know if this is a really bad idea for other reasons or not. Also I haven't tested it on anything but this one url.