Thanks for the report

Doing a quick look against trunk/SVN head, a lot of these have been moved, and possibly fixed.

A lot of them come up to be comment lines, or even line numbers past the end of the file.

It's also quite likely a lot of them have been fixed in SVN, and then not backported, especially if they've come from third party libraries (ie if they're directly from jquery)

We do have access to everywhere, so fixing them isn't much of an issue :)

Even the phase3 ones, seemingly to be comments etc.

If you were to run it against a SVN copy, and report the errors, that'd be appreciated :)

I'll leave it open for Roan/Trevor to have a look at the latter issues for the moment :)

I use v1.16 so I really don't know where some files were moved but I tried my best and here is the result from the SVN trunk:

WikiEditor/modules…
contentCollector.js

• 434

ext.wikiEditor…
.addMediaWizard:

• 15: }); (the whole line)

.tests.toolbar:

• 245: setTimeout( function() { button.slideDown( 'fast' ); }, 2000 ); (the whole line)

jquery.wikieditor…
.dialogs:

• 38

.iframe:

• 1143
• 1155
• 1167
• 1313: }; } )( jQuery ); (the whole line)

.templateEditor:

• 256
• 381
• 626
• 650

.toc:

• 474: setTimeout( function() { $.wikiEditor.modules.toc.fn.unhighlight( context ); }, 1000 ); (the whole line)

.toolbar:

• 288
• 455
• 468
• 472

/skins/common/ajax.js:

• 36
• 65

/skins/common/ajaxwatch.js doesn't contain this error :)

If you can tell me now where the following files are located, I could inspect those also:

• UsabilityInitiative/js/usability.js
• /plugins/jquery.async.js
• /plugins/jquery.textSelection.js

I'm not sure, but as far as I know, only the jquery.async.js file is from a 3rd party which contains error.

/skins/common/ajax.js:

• 36 - Is a } after "return true;"
• 65 - Is a } after "return A;"

If it's wanting a ; after those (which makes some sense), line 167 and 177 should also have them.

I've just added in r77872 to fix common/ajax.js

Looking at ajaxwatch.js, that already has those trailing ; at SVN head.

In SVN, most of the phase3 ones won't have been moved, I don't think.

For the Usability Extensions, they've been split down to separate extensions:
UsabilityInitative/WikiEditor -> WikiEditor

I'm interested to see how (or even, IF), the JS minifier Roan and Trevor are using do anything to correct this automatically (it's possible it might).

Though, it wouldn't be bad practise to do it manually.

Most of the other javascript stuff that's been folded into phase3, are in the resources directory (phase3/resources)

I'm interested to see how (or even, IF), the JS minifier Roan and Trevor are
using do anything to correct this automatically (it's possible it might).

Auto-semicolon-ing stuff sounds like something that would cause mysterious hard to find bugs. If you did have a smart algorithm for adding the semicolons, it'd probably be the same as the auto-semicolon insert rules in js, in which case there really wouldn't be much of a point.

No, I don't have one; I did the inspection semi-automatic.

I used Zend Studio which indicated me where syntax errors are. Maybe it worth a try to investigate Aptana, which is an open source software probably with a similar capability.

My IDE found 192 in phase3...

Fixed those in r77922, r77923, r77924

Another in r77926

And WikiEditor in r77928

Does this need any more work now that ResourceLoader is going in?

(In reply to comment #8)

Does this need any more work now that ResourceLoader is going in?

It's almost more needed now. The lack of semi colons when you have minified the files, cause more problems to the browsers (see original post)

I've probably got *most* of them, but likely not, so I'll poke them all and see how many more there seems to be...

ResourceLoader uses a minifier that's magic semi-colon safe. If there are cases where magic semi-colons are indeed breaking because of JavaScriptDistiller, that's a bug.

Closing this. New problems == new bugs. Re Trevor in IRC:

<TrevorParscal> if someone finds a bug to do with magic semi-colons breaking,

then it can be re-opened, it would be a valid regression