Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

js related bug

2 views
Skip to first unread message

albor

unread,
Nov 22, 2003, 5:10:09 PM11/22/03
to
Hello,


Test case:

<http://albor.ruweb.net/opera/bugs/bug4.html> (30KB)

This bug is mysterious for me...
The result depends upon very strange things (e.g. size of the file)...
I failed to catch the real cause...

Is it a bug in Opera (I'm sure it is) or in the code?


--
albor{at}pisem.net

Opera 7.22 build 3221 JRE 1.4.1 Win2000 Proxomitron Naoko 4.5

Da Twink Daddy

unread,
Nov 22, 2003, 5:46:00 PM11/22/03
to
On Sun, 23 Nov 2003 04:10:09 +0600, albor <al...@pisem.NOSPAM.net> wrote:

> <http://albor.ruweb.net/opera/bugs/bug4.html> (30KB)

<URL:
http://validator.w3.org/check?uri=http%3A%2F%2Falbor.ruweb.net%2Fopera%2Fbugs%2Fbug4.html&charset=%28detect+automatically%29&doctype=XHTML+1.0+Transitional
>

> This bug is mysterious for me...
> The result depends upon very strange things (e.g. size of the file)...
> I failed to catch the real cause...

The real cause is probably related to '<'s and '&'s in your js.

--
Da Twink Daddy
datwin...@icqmail.com
ICQ: Da Twink Daddy (514984)
YM: DaTwinkDaddy
AIM: DaTwinkDaddy

Lee Harvey

unread,
Nov 22, 2003, 9:39:24 PM11/22/03
to
albor wrote:
> Test case:
>
> <http://albor.ruweb.net/opera/bugs/bug4.html> (30KB)
>
> This bug is mysterious for me...
> The result depends upon very strange things (e.g. size of the file)...
> I failed to catch the real cause...
>
> Is it a bug in Opera (I'm sure it is) or in the code?

Odd. Definitely something going on here.

When I changed your <script> block from this...

<script type="text/javascript">
..
</script>

...to this...

<script type="text/javascript">
<!--
..
// -->
</script>

NOTICE: Added comment block <!-- // -->

I got a lot more (all?) results printed!

Hmm.

--
Feel free to submit a report: <URL: https://bugs.opera.com/wizard/ >


albor

unread,
Nov 24, 2003, 3:51:53 PM11/24/03
to
Hello,

>> <http://albor.ruweb.net/opera/bugs/bug4.html> (30KB)

> <URL:
>
http://validator.w3.org/check?uri=http%3A%2F%2Falbor.ruweb.net%2Fopera%2Fbug
s%
>
2Fbug4.html&charset=%28detect+automatically%29&doctype=XHTML+1.0+Transitiona
l>

It is the rare case when validator is wrong.

>> I failed to catch the real cause...

> The real cause is probably related to '<'s and '&'s in your js.

Content of <script> element is CDATA and SGML rules do not allow character
references in CDATA element content.

I'm almost sure it is a bug in Opera (and in validator).

Nevertheless try another test case with js in external file, it is valid:

<http://albor.ruweb.net/opera/bugs/bug4.1.html>

albor

unread,
Nov 24, 2003, 4:32:03 PM11/24/03
to
Hello,

>>> <http://albor.ruweb.net/opera/bugs/bug4.html> (30KB)

>> <URL:
>>
http://validator.w3.org/check?uri=http%3A%2F%2Falbor.ruweb.net%2Fopera%2Fbug
>> s%
>>
2Fbug4.html&charset=%28detect+automatically%29&doctype=XHTML+1.0+Transitiona
>> l>

> It is the rare case when validator is wrong.

Pardon me, validator was right... Valid test case:

<http://albor.ruweb.net/opera/bugs/bug4.valid.html>

Da Twink Daddy

unread,
Nov 25, 2003, 12:20:51 AM11/25/03
to
On Tue, 25 Nov 2003 03:32:03 +0600, albor <al...@pisem.NOSPAM.net> wrote:

> Pardon me, validator was right... Valid test case:
>
> <http://albor.ruweb.net/opera/bugs/bug4.valid.html>

Yeah, It looks good to me... mostly.
I still think you should put a <![CDATA[]]> or <!-- --> around the script,
but it should work as is.

OmegaJunior

unread,
Nov 25, 2003, 3:51:16 AM11/25/03
to
"Da Twink Daddy" wrote ...

> albor wrote:
>
> > Pardon me, validator was right... Valid test case:
> >
> > <http://albor.ruweb.net/opera/bugs/bug4.valid.html>
>
> Yeah, It looks good to me... mostly.
> I still think you should put a <![CDATA[]]> or <!-- --> around the script,
> but it should work as is.

Last time I checked, Opera ignored script in cdata block tags. But that was
a week ago with xhtml1.1.

OJ


Rijk van Geijtenbeek

unread,
Nov 25, 2003, 5:46:55 AM11/25/03
to
On Tue, 25 Nov 2003 09:51:16 +0100, OmegaJunior <gu...@what.not> wrote:
> "Da Twink Daddy" wrote ...
>> albor wrote:

>>> Pardon me, validator was right... Valid test case:
>>> <http://albor.ruweb.net/opera/bugs/bug4.valid.html>

>> Yeah, It looks good to me... mostly.

Except for the problems when rendering this :(
In 7.23, the list doesn't show at all when ID-ing as Opera, and with lots
of 'undefined' items from 77 onward when ID-ing as MSIE.

Please report a bug.


>> I still think you should put a <![CDATA[]]> or <!-- --> around the
>> script, but it should work as is.
>
> Last time I checked, Opera ignored script in cdata block tags. But that
> was a week ago with xhtml1.1.

Opera 7.0-7.2 ignores any SCRIPT in XML documents (those really served and
treated as XML).

--
The Web is a procrastination apparatus: | Rijk van Geijtenbeek
It can absorb as much time as | Opera Software ASA
is required to ensure that you | Documentation & QA
won't get any real work done. - J.Nielsen | mailto:ri...@opera.com

Robert Blaut

unread,
Nov 25, 2003, 6:54:07 AM11/25/03
to
On Tue, 25 Nov 2003 11:46:55 +0100, Rijk van Geijtenbeek <ri...@opera.com>
wrote:

> Opera 7.0-7.2 ignores any SCRIPT in XML documents (those really served
> and treated as XML).

So I guess Opera 7.30 has no problem with it?

--
Robert Błaut
Using M2, Opera's revolutionary e-mail client: http://www.opera.com/m2/

Rijk van Geijtenbeek

unread,
Nov 25, 2003, 7:18:59 AM11/25/03
to
On Tue, 25 Nov 2003 12:54:07 +0100, Robert Blaut <qui...@despammed.com>
wrote:

> On Tue, 25 Nov 2003 11:46:55 +0100, Rijk van Geijtenbeek
> <ri...@opera.com> wrote:
>
>> Opera 7.0-7.2 ignores any SCRIPT in XML documents (those really served
>> and treated as XML).
>
> So I guess Opera 7.30 has no problem with it?

Some parts of JavaScript, like document.write, are not really compatible
with XML. It would make the document invalid in most cases. So that will
not be supported.

But if no big problems crop up during testing, DOM manipulation of XML
documents might work in the next non-security upgrade.

Robert Blaut

unread,
Nov 25, 2003, 7:33:45 AM11/25/03
to
On Tue, 25 Nov 2003 13:18:59 +0100, Rijk van Geijtenbeek <ri...@opera.com>
wrote:

> But if no big problems crop up during testing, DOM manipulation of XML

> documents might work in the next non-security upgrade.

Sounds good. We're waiting very impatient for first public appearance of
the new 7.30 line :)

Tim Altman

unread,
Nov 25, 2003, 10:53:46 AM11/25/03
to
On Tue, 25 Nov 2003 11:46:55 +0100, "Rijk van Geijtenbeek"
<ri...@opera.com> wrote:

>On Tue, 25 Nov 2003 09:51:16 +0100, OmegaJunior <gu...@what.not> wrote:
>> "Da Twink Daddy" wrote ...
>>> albor wrote:
>
>>>> Pardon me, validator was right... Valid test case:
>>>> <http://albor.ruweb.net/opera/bugs/bug4.valid.html>
>
>>> Yeah, It looks good to me... mostly.
>
>Except for the problems when rendering this :(
>In 7.23, the list doesn't show at all when ID-ing as Opera, and with lots
>of 'undefined' items from 77 onward when ID-ing as MSIE.
>
>Please report a bug.

I'm getting an error in the Javascript console due to one of the
characters on line 79. Anyone else?

--
Tim Altman
Customer Service
Opera Software

roy...@myrealSP-AMbox.com
No SP-AM is good spam.

Andrew Gregory

unread,
Nov 25, 2003, 9:59:02 PM11/25/03
to
On Tue, 25 Nov 2003 10:53:46 -0500, Tim Altman <add...@in.sig> wrote:

> On Tue, 25 Nov 2003 11:46:55 +0100, "Rijk van Geijtenbeek"
> <ri...@opera.com> wrote:
>
>> On Tue, 25 Nov 2003 09:51:16 +0100, OmegaJunior <gu...@what.not> wrote:
>>> "Da Twink Daddy" wrote ...
>>>> albor wrote:
>>

>>>>> <http://albor.ruweb.net/opera/bugs/bug4.valid.html>


>>
> I'm getting an error in the Javascript console due to one of the
> characters on line 79. Anyone else?

I don't get any JS errors, 78 through 423 show as undefined on the page.
The line shown for 77 (a[76] in the code) is truncated.

--
Andrew Gregory
<URL: http://www.scss.com.au/family/andrew/ >

Tim Altman

unread,
Nov 26, 2003, 11:11:25 AM11/26/03
to

Alright, please file a bug report.

0 new messages