299792 - krum.rz.uni-mannheim.de/cabench/bgk86.txt crashed browser when clicking "File->Print Preview"

Status: RESOLVED INCOMPLETE

(Core :: Print Preview, defect)

Description

[Yuri]

User-Agent:       Mozilla/5.0 (X11; U; FreeBSD i386; en-US; rv:1.7.7) Gecko/20050420 Firefox/1.0.3
Build Identifier: Mozilla/5.0 (X11; U; FreeBSD i386; en-US; rv:1.7.7) Gecko/20050420 Firefox/1.0.3

see summary

Reproducible: Always

Comment 1

[Martijn Wargers (dead)]

Works for me with current trunk build.

Comment 2

[nwdbc]

I tried this with 3 different computers and the latest stable release (version 1.5.0.4) on each one.

for this web page

http://www.bell.ca/shopping/PrsShpWls_RtpFamilyShare.page

Using print preview or print it crashes (calls \windows\system32\dwwin.exe invoking the "Please tell Microsoft..." dialogue and closes out of firefox).

Comment 3

[timeless]

[e-mail]: please don't use random bugs for new reports of similar problems. especially bugs like this. file your own bug.

Comment 4

[nwdbc]

"random bug" would imply that I had picked one at random.  I did not.  I chose one that seemed to be the same problem I was describing (crash on print preview).

If my unfamiliarity with the bug reporting system led me to file incorrectly then 'whoops'.  However, perhaps a better use of your efforts would be to attempt to improve the bug submission instruction page, which directs people to check that the bug has not yet been submitted.

Thanks for the comment though.  Nice to know someone is listening if nothing else. 

Comment 5

[timeless]

I'm listening. the problem in this case is actually that we have an entire component for print preview, so "crash on print preview" should describe *all* crash bugs in the entire component (quickly moves bug to proper component).

There are 35 atm using this search: https://bugzilla.mozilla.org/buglist.cgi?component=Print+Preview&keywords_type=allwords&keywords=crash&resolution=DUPLICATE&resolution=---

This search yields 18 bugs, only one open:
https://bugzilla.mozilla.org/buglist.cgi?&short_desc_type=allwordssubstr&short_desc=preview&component=Printing&keywords_type=allwords&keywords=crash&resolution=DUPLICATE&resolution=---

I'm not sure how to address the fact that it seems like a good match. I definitely care about the Bugzilla experience and you have my sympathies. If you have any ideas about how we can fix this, please let me know (we could move this into a new bug instead of spamming this bug, but it was opened nearly a year ago and is unlikely to go anywhere, so I'd sooner use it for some good -- yes that's counter to my point about you recycling bugs, but...).

A few things to consider (and which we should make clear in the entry page):
* if the bug is actually characterized and your problem has the same detailed characterization, then it doesn't matter if the platforms don't match.
* if the bug is uncharacterized, then it should focus on the specific instance (a certain web page, on a certain platform, following a quirky step list -- if the step list is generic, then recycling the bug isn't a good idea, especially if the platforms or only interesting details [web page, platform, version] don't match)
* bugs about crashing really need talkback for windows/linux/macos(x) or a stack trace (for e.g. freebsd), filing one without this is not going to get it very far unless it actually happens to be something that someone else decides to reproduce and they get one
* bugs should be filed with a reference, if you *always* crash on print preview, then say that. if not everything you do triggers a crash then mention "print preview works on this other page <>". i.e. it helps to explain what's *special*. If it really happens no matter where you go, then it really helps to specify this.
* for printing (and the guided form should suggest this), it's a good thing to specify the names of all your printers, and their drivers and their version numbers.

In short, the distinguishing features about a problem are what (any) you should use to decide whether to share a bug. Print Preview being its own component and crashing being something that happens a lot are not very good distinguishing characteristics. Web pages tend to be distinguishing, and that a bug is filed against a strange platform (FreeBSD) can be distinguishing. Note that which characteristics are distinguishing will change over the life of the bug. currently the only interesting bits we have in this bug are the url and the platform. If a stack trace (from a core or talkback) appears, the summary would be changed to indicate a signature which would replace the url as a distinguishing characteristic.

Anyway, whatever suggestions you have (possibly just rewording what I've written, possibly writing something else) should hopefully be used to improve things.

And of course, please try to install talkback and file a bug about your specific crash (once you get an incident id).

Note that the reason I'm using Bugzilla for this comment is because it means other people can (hopefully) gain something from it later, private emails get lost.

Comment 6

[nwdbc]

Thanks for taking the time to respond.  Some good points that have expanded my knowledge a bit.

When initially looking for an appropriate bug that matched my own I was reading descriptions more than anything else.  Even now without your mention of FreeBSD I would be unsure if the dropdownlists at the top of the page are preselected with the info that was used when initially reporting this bug.  If this was text rather than dropdowns it would be more obvious that I posted incorrectly (since I'm running Windows XP Pro SP2 on these three machines).

I guess when I got to the part of the bug reporting guidelines that said, "before you enter your bug confirm that has not been previously reported" I followed the link away to the 'searching' page.  Searched.  Entered my comment in support of what I (mistakenly) presumed to be the same bug.  Ideally, I should have read the rest of the initial submission guide down to the point where it makes reference to things like platform used.

In any event, my print preview works flawlessly on many many pages.  I'm presuming there's something funky in this particular page that I mentioned.

As you suggest, I'll install the talkback component.  Try the page again and refile as a new bug ... if I find the time.  Though it's a small quirk that obviously doesn't happen on many pages.  A minor bug and so perhaps not likely to be looked at even if I do report it.

Thanks again.

Comment 7

[Wayne Mery (:wsmwk)]

still see this problem?
if not please close this bug accordingly (probably worksforme)

Comment 8

[Tyler Downer [He/Him]]

Please reopen this bug if it appears with a recent version of firefox in a new profile.