Firefox dies on simple script

**kaeli** · Jul 23 '05, 06:23 PM

Re: Firefox dies on simple script

In article <d0qede$on8$1@c hessie.cirr.com >, ataru@nospam.cy berspace.org
enlightened us with...[color=blue]
> window.onerror= function( msg, url, line ) {
> switch( msg ) {
> default: break;
> case foo: break;[/color]

Why is your default at the beginning?

Just curious.

--
--
~kaeli~
Shotgun wedding: A case of wife or death.

http://www.ipwebdesign.net/wildAtHeart

http://www.ipwebdesign.net/kaelisSpace

**Christopher Benson-Manica** · Jul 23 '05, 06:23 PM

Re: Firefox dies on simple script

kaeli <tiny_one@nospa m.comcast.net> spoke thus:
[color=blue]
> Why is your default at the beginning?[/color]

In the actual code, the default case falls through to another case
after alerting the user appropriately.

--
Christopher Benson-Manica | I *should* know what I'm talking about - if I
ataru(at)cybers pace.org | don't, I need to know. Flames welcome.

**RobG** · Jul 23 '05, 06:24 PM

Re: Firefox dies on simple script

Christopher Benson-Manica wrote:[color=blue]
> Is Firefox 1.0.1 within its rights to crash when presented with the
> following script?
>
> <html>
> <head>
> <script>
> var foo="foo";
> window.onerror= function( msg, url, line ) {
> switch( msg ) {
> default: break;
> case foo: break;
> }
> }
> </script></head></html>
>
> I'm betting not; I'm planning to file a bug report.
>[/color]

No script should cause the browser itself to crash, regardless
of syntactical correctness, so yes it's a bug in Firefox.

The ECMA spec does not state explicitly that the default
statement must be in any particular position:

<URL:http://www.mozilla.org/js/language/E262-3.pdf>

Section 12.11

However, putting it first with a single 'break' statement means
that none of the following clauses will be evaluated and
therefore it's something that is only likely to happen whilst
debugging.

There are a number of work-arounds, the simplest being to put a
trivial clause ahead of the default so that it is no longer
first:

window.onerror= function( msg, url, line ) {
switch( msg ) {
case foo2: continue;
default: break;
case foo: break;
}
}

--
Rob

**Martin Honnen** · Jul 23 '05, 06:24 PM

Re: Firefox dies on simple script

Christopher Benson-Manica wrote:
[color=blue]
> Is Firefox 1.0.1 within its rights to crash when presented with the
> following script?
>
> <html>
> <head>
> <script>
> var foo="foo";
> window.onerror= function( msg, url, line ) {
> switch( msg ) {
> default: break;
> case foo: break;
> }
> }
> </script></head></html>
>
> I'm betting not; I'm planning to file a bug report.[/color]

Crash is always a bug, I seached whether you had filed one but couldn't
find it so I filed this as
<https://bugzilla.mozill a.org/show_bug.cgi?id =285755>

--

Martin Honnen

Attention Required! | Cloudflare

http://JavaScript.FAQTs.com/

rh · Jul 23 '05, 06:24 PM

Re: Firefox dies on simple script

RobG wrote:

<..>
[color=blue]
> The ECMA spec does not state explicitly that the default
> statement must be in any particular position:
>
> <URL:http://www.mozilla.org/js/language/E262-3.pdf>
>
> Section 12.11
>
> However, putting it first with a single 'break' statement means
> that none of the following clauses will be evaluated and
> therefore it's something that is only likely to happen whilst
> debugging.
>[/color]

I don't think 12.11 reads quite that way. All "case" clauses should be
tested prior to forcing the default, regardless of the source-text
position of the "default" clause.

So the only time the source-text position of the default should be of
consequence is when there is a fall-through from a preceding case
statement execution sequence.
[color=blue]
> There are a number of work-arounds, the simplest being to put a
> trivial clause ahead of the default so that it is no longer
> first:
>
> window.onerror= function( msg, url, line ) {
> switch( msg ) {
> case foo2: continue;
> default: break;
> case foo: break;
> }
> }
>[/color]

And while that's the right idea for a work-around, recall that
"continue" can only be used within an iterative construct. Something
like

case msg+'#':
default: break;

would seem to be a better choice as a (never-successful case) buffer.

../rh

**Christopher Benson-Manica** · Jul 23 '05, 06:37 PM

Re: Firefox dies on simple script

Martin Honnen <mahotrash@yaho o.de> spoke thus:
[color=blue]
> Crash is always a bug, I seached whether you had filed one but couldn't
> find it so I filed this as
> <https://bugzilla.mozill a.org/show_bug.cgi?id =285755>[/color]

I didn't file a bug report since the latest nightly build seemed to
have fixed it :)

--
Christopher Benson-Manica | I *should* know what I'm talking about - if I
ataru(at)cybers pace.org | don't, I need to know. Flames welcome.

Firefox dies on simple script

Firefox dies on simple script

Comment

Comment

Comment

Comment

Comment

Comment