120756 - function use before <script> declaration considered harmful

Status: RESOLVED FIXED

(Bugzilla :: Bugzilla-General, defect, P1)

Description

[timeless]

IE5.5 and 6 at the very least result in runtime errors if you do it.  The
attachments manager triggers this a lot. a non harmful example is to click edit
attachment as comment. you get a script error and no results.

Comment 1

[Bradley Baetz (:bbaetz)]

Much as we all love the lizard, I think that this is a 2.16 blocker. It would
possibly be easier for someone with access to IE to do this one.

Comment 2

[Christian Reis]

This doesn't happen with the query.cgi JS-foo, but I agree it's a good thing.
Should this be blocked by the "make JS a separate file" bug to make it easier to
implement?

Comment 3

[Gervase Markham [:gerv]]

Attachment: Patch v.1

This seems sensible, but I have no ability to test it. I just moved the entire
<script> block to the top of the file.

Gerv

Comment 4

[Dave Miller [:justdave]]

Need reviews from Javascript folks on this one.

Comment 5

[Christopher Aillon (sabbatical, not receiving bugmail)]

Comment on attachment 67406 [details] [diff] [review]
Patch v.1

r=caillon

Comment 6

[Andreas Franke (gone)]

Comment on attachment 67406 [details] [diff] [review]
Patch v.1

r=afranke

Comment 7

[Dave Miller [:justdave]]

-> patch author

Comment 8

[Dave Miller [:justdave]]

Comment on attachment 67406 [details] [diff] [review]
Patch v.1

I went to check this in, and I can't get the patch to apply.  diff claims it's
"reversed or already applied", and the patch fails whether I apply it in
reverse mode or not.

Comment 9

[Dave Miller [:justdave]]

That patch was diffed against r1.7 of edit.atml, current cvs tip is r1.8.  I did
try applying the patch to r1.7 then updating, but cvs failed to merge it also.

Comment 10

[Andreas Franke (gone)]

Attachment: Updated patch to rev. 1.8

This is the patch against rev. 1.8 of template/default/attachment/edit.atml .
I have recreated it manually, without using the old patch. (The patch just
swaps the order of occurrence of the main <form> and <script> elements.)
I have tried to test the patch in Netscape4 and Mozilla, and I could not find
any oddities. Seems to work fine.

Comment 11

[Andreas Franke (gone)]

Comment on attachment 69083 [details] [diff] [review]
Updated patch to rev. 1.8

Re-marking r=afranke.

Comment 12

[Gervase Markham [:gerv]]

Checked in. 

Gerv