https://bugzilla.wikimedia.org/show_bug.cgi?id=69924
--- Comment #21 from Bartosz Dziewoński ---
I quite like the document.write one, seems it will have the least impact on
users with non-broken user JS and should be easy to implement?
--
You are receiving this mail because:
You are the assigne
https://bugzilla.wikimedia.org/show_bug.cgi?id=69924
--- Comment #20 from Krinkle ---
(In reply to Krinkle from comment #19)
> Another solution would be to return early instead of wrap the whole thing.
>
> While a return statement is not allowed in the global scope, a throw works
> fine.
>
>
>
https://bugzilla.wikimedia.org/show_bug.cgi?id=69924
--- Comment #19 from Krinkle ---
Another solution would be to return early instead of wrap the whole thing.
While a return statement is not allowed in the global scope, a throw works
fine.
if (!cond)
throw '';
.. rest of script ..
.. mw
https://bugzilla.wikimedia.org/show_bug.cgi?id=69924
--- Comment #18 from Krinkle ---
Another solution would be to return early instead of wrap the whole thing.
While a return statement is not allowed in the global scope, a throw works
fine.
if (!cond)
throw '';
.. rest of script ..
.. mw
https://bugzilla.wikimedia.org/show_bug.cgi?id=69924
--- Comment #17 from Krinkle ---
Alternatively, we might be able to do the condition outside the script. Perhaps
something like replace this:
with:
if (cond) document.write("\u003Cscript
src=\"./load.php?modules=user\u0026only=scripts\u0026
https://bugzilla.wikimedia.org/show_bug.cgi?id=69924
--- Comment #16 from Michael M. ---
What about only wrapping the
mw.loader.state({"user":"ready"});
at the bottom in an if-clause (or prefixing it with window.mw && ) and leaving
the rest of the script alone? I hadn't seen that call to mw.load
https://bugzilla.wikimedia.org/show_bug.cgi?id=69924
--- Comment #15 from Krinkle ---
Open to other ideas. Of neither #1 or #2 is acceptable, we may have to remove
the if-wrap for the time being since modern Firefox triumps priority over grade
C browsers.
--
You are receiving this mail because:
https://bugzilla.wikimedia.org/show_bug.cgi?id=69924
--- Comment #14 from Krinkle ---
Hm.. so yeah, that's annoying. The condition-wrap is supposed to hide an issue
in old browsers, not introduce issues in modern browsers.
So far I've only seen two possible solutions:
1) Wrap in $.globalEval, p
https://bugzilla.wikimedia.org/show_bug.cgi?id=69924
--- Comment #13 from Krinkle ---
> quux(); if (true) { function quux() { return 1; } } else { function quux() {
> return 2; } }
V8> 2
SpiderMonkey> quux is undefined
--
You are receiving this mail because:
You are the assignee for the bug.
Y
https://bugzilla.wikimedia.org/show_bug.cgi?id=69924
--- Comment #12 from Krinkle ---
> > function bar() { return 1; }
> > bar();
> >> 1
>
> > if (true) {
> > function bar() { return 1; }
> > } else {
> > function bar() { return 2; }
> > }
> > bar();
> >> 2
btw, in Firefox this yields 1.
-
https://bugzilla.wikimedia.org/show_bug.cgi?id=69924
--- Comment #11 from Krinkle ---
This is not a new problem. Users using older browsers have always had fatal
errors because jquery/mediawiki are not loaded in blacklisted browsers not
supporting our required version for the javascript engine. T
https://bugzilla.wikimedia.org/show_bug.cgi?id=69924
--- Comment #10 from Michael M. ---
As actually nobody explicitly quoted the relevant standards yet:
* ECMAScript defines FunctionDecleration and FunctionExpression.
(http://ecma-international.org/ecma-262/5.1/#sec-13)
* Firefox (and other bro
https://bugzilla.wikimedia.org/show_bug.cgi?id=69924
--- Comment #9 from Michael M. ---
(In reply to Erik Moeller from comment #7)
> Michael, if common.js is non-blank, this will lead to site-wide errors in
> IE6 (and other disabled browsers).
My comment was limited to user scripts, and did not
https://bugzilla.wikimedia.org/show_bug.cgi?id=69924
--- Comment #8 from Andre Klapper ---
Another example:
https://de.wikipedia.org/w/index.php?title=Wikipedia:Technik/Werkstatt&oldid=133432119#.C3.84nderung_im_Javascript.3F
--
You are receiving this mail because:
You are the assignee for the
https://bugzilla.wikimedia.org/show_bug.cgi?id=69924
--- Comment #7 from Erik Moeller ---
Michael, if common.js is non-blank, this will lead to site-wide errors in IE6
(and other disabled browsers). That was one of the motivations for Timo's
change. But I'll let him comment on strategies to resol
https://bugzilla.wikimedia.org/show_bug.cgi?id=69924
--- Comment #6 from Michael M. ---
Why do user scripts have to be enveloped in that if-clause anyway?
1. In most browsers the global mw will be defined before the user script is
executed, so there is no need to test for it.
2. If for some reas
https://bugzilla.wikimedia.org/show_bug.cgi?id=69924
--- Comment #5 from Fomafix ---
$.globalEval() [1] as suggested in bug 63587 would be a possible solution.
Maybe this can added as a general optional parameter for all modules.
[1] https://api.jquery.com/jQuery.globalEval/
--
You are receiv
https://bugzilla.wikimedia.org/show_bug.cgi?id=69924
Andre Klapper changed:
What|Removed |Added
Priority|Unprioritized |High
--
You are receiving this mail b
https://bugzilla.wikimedia.org/show_bug.cgi?id=69924
Helder changed:
What|Removed |Added
See Also||https://bugzilla.wikimedia.
19 matches
Mail list logo
