292837 - crash on viewing this particular SVG

Status: RESOLVED WORKSFORME

(Core :: SVG, defect)

Description

[edA-qa mort-ora-y]

User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.8b2) Gecko/20050503 Firefox/1.0+
Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.8b2) Gecko/20050503 Firefox/1.0+

When attempting to view the attached SVG file, Firefox crashes.  Other SVG files
do work (with GDI+) -- even those referred in other Bugzilla issues.


Reproducible: Always

Steps to Reproduce:
1. Load attached file.
2. Crash
3.

Actual Results:  
QA Feedback tool

Expected Results:  
Rendered the SVG

Comment 1

[edA-qa mort-ora-y]

Attachment: Sample file that crashses

Comment 2

[edA-qa mort-ora-y]

Sometimes it doesn't crash, rather only renders a single box in the SVG (one of
several hundred).

Comment 3

[tor]

No crash here.  Lots of rectangles on a build tree - might be a dup of bug 293004.

Comment 4

[Hermann Schwab]

Attachment: screenshot showing buggy file in inkscape

http://inkscape.org/
Open Source Scalable Vector Graphics Editor

There is a black rectangle drawn with a thick line leaving the sheet,
proceeding to the border of the window. Don´t know if this is a bug in Inkscape
or in Firefox, maybe it is a bug in the file.
The comma at the end of the vector like targets[0] = [1,1,1,1,... 1,0,]; seems
to be at least needless.

       <script type="text/ecmascript"><![CDATA[

	   targets = new Array();
targets[0] =
[1,1,1,1,1,1,1,0,0,0,1,1,1,1,1,1,1,1,0,0,1,0,0,1,1,1,1,1,1,1,1,1,1,0,1,0,0,1,1,1,1,1,0,1,1,1,0,1,1,1,1,1,1,0,0,1,1,1,1,0,0,1,1,1,1,1,0,1,1,0,1,1,1,1,1,1,1,1,1,0,0,1,1,1,0,1,1,1,1,1,1,1,0,0,0,];

Comment 5

[Hermann Schwab]

Attachment: screenshot firefox

Comment 6

[Hermann Schwab]

Mozilla/5.0 (Windows; U; Win98; en-US; rv:1.8b2) Gecko/20050504 Firefox/1.0+
I didn´t crash, but jsconsole is showing a lot of errors, like 

Error: Unknown property ''.  Declaration dropped.
Source File: https://bugzilla.mozilla.org/attachment.cgi?id=182574
Line: 0

Error: hoverOn is not defined
Source File: https://bugzilla.mozilla.org/attachment.cgi?id=182574
Line: 1

Error: hoverOff is not defined
Source File: https://bugzilla.mozilla.org/attachment.cgi?id=182574
Line: 1

Comment 7

[edA-qa mort-ora-y]

(In reply to comment #4)
> Created an attachment (id=182757) [edit]
> screenshot showing buggy file in inkscape
> http://inkscape.org/
> Open Source Scalable Vector Graphics Editor
> There is a black rectangle drawn with a thick line leaving the sheet,

This looks about as expected (as viewed in Adobe SVG viewer).
Though the overlapping black text shouldn't be shown (only one line of text at a
time is expected).

Comment 8

[edA-qa mort-ora-y]

(In reply to comment #5)
> Created an attachment (id=182758) [edit]
> screenshot firefox
> 

This appears to be missing the numbers/text items.  (Note, I used Adobe as a
reference viewer, that may also explain missing properities -- which Adobe SVG
wasn't complaining about).

Comment 9

[José Jeria]

WFM: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8b2) Gecko/20050506
Firefox/1.0+

Comment 10

[edA-qa mort-ora-y]

To the missing items and whatnot, I had got this SVG working in Adobe SVG okay.
 I decided to run this particular file through the SVG validator
<http://jiggles.w3.org/svgvalidator/validator> and I get this report:

Validation report for the chart.svg file:

    * 1 error was found.
         1.

            Line: 7
            The 'version' attribute of the <svg> element has a fixed value: it
can not be changed.
    * no warning was reported.

Note: I'm not sure if that is checking the script aspects however...

Comment 11

[Hermann Schwab]

(In reply to comment #7)

> > screenshot showing buggy file in inkscape

> > There is a black rectangle drawn with a thick line leaving the sheet,
> 
> This looks about as expected (as viewed in Adobe SVG viewer).
> Though the overlapping black text shouldn't be shown (only one line of text
>at a time is expected).

Did you compare my screenshots from inkscape and firefox? In Inkscape something
missing in Firefox is shown, but that missing element is leaving the window and
overflowing to somewhere. I got a crash with inkscape triggering DocWatson, but
I couldn´t crash Firefox. But DocWatson data is pretty useless compared to
Talkback data. 
As you got the QA Feedback tool comming up, could you please reproduce the
crash, fill in as comment 'Bug 292837' and submit?
Info about talkback: http://www.mozilla.org/quality/qfa.html

After the report is sent, a TalkbackID is sent back from the server. Please open
talkback.exe in the firefox/components directory and copy&paste this talkback
number in a comment over here.
You also can have a look yourself at the sourcecode where it crashed, if you
search for your report at http://talkback-public.mozilla.org/talkback/fastfind.jsp
Insert the talkback ID top-left in FastFind and press go.
If the server is offline try later, if the server is busy you can bookmark and
retry later if the report is ready. The links at the right side of your report
are links into the source code.

If you compare Adobe, Firefox and Inkscape, they all show part of the file
sameway, and another part missing or different. Can you make a smaller testcase,
showing this part only?

Comment 12

[edA-qa mort-ora-y]

With the newest nightly I cannot cause the crash to occur anymore.

However, it isn't rendering correctly (being simply completely wrong when
compared with Adobe or Inkscape).  I now get a single box and a large circle
covering half of the screen.

I can see if I can get a smaller reproduction case.

Comment 13

[edA-qa mort-ora-y]

Attachment: Showing text rendering error.

This shows that text rendering is failing (possibly with the window size
chosen).  And this may be the root cause of why the larger SVG is failing to
render correctly.

Comment 14

[edA-qa mort-ora-y]

Bug 290044 (relating to minimum font-size) is causing part of the rendering
problem.  If I turn off the minimum font size I at least see all the boxes on
the display correctly, however, I don't see any text whatsoever.

(Is ecmascript supported yet in the SVG viewer?)

Comment 15

[edA-qa mort-ora-y]

Bug 293852 has been created to deal with an incorrect font rendering issue.
Bug 267953 already exists indicating incorrect script naming handling (even
allowing for this the script isn't working in the native SVG)


Given that the crash no longer occurs, and this file exhibits a culmination of
other SVG viewer defects, I'm okay if we close this issue and I will revisit
this sometime in the future when the other issues are resolved.

Comment 16

[Chris Blore]

Ok, I'll close this bug and mark it WORKSFORME as I cannot find any crash and
there are currently other bugs open relating to the other issues you mention.
Feel free to get this bug reopened if the problem should resurface...