This bug was filed from the Socorro interface and is report bp-2b117400-de08-44ae-8b08-10fb62111015 . ============================================================= This is a new crash showing up in 10.7.2 only and affects versions up to trunk. https://crash-stats.mozilla.com/report/list?signature=FinderKit@0x78d87 Comments mention downloading and saving pictures. Perhaps a different manifestation of Bug 670842? https://crash-stats.mozilla.com/report/index/2b117400-de08-44ae-8b08-10fb62111015 Comments Frame Module Signature [Expand] Source 0 FinderKit FinderKit@0x78d87 1 FinderKit FinderKit@0x78eef 2 FinderKit FinderKit@0x119f6d 3 FinderKit FinderKit@0x162d72 4 FinderKit FinderKit@0x162bee 5 FinderKit FinderKit@0x1630ef 6 FinderKit FinderKit@0x16c560 7 FinderKit FinderKit@0x164541 8 FinderKit FinderKit@0x19fc94 9 FinderKit FinderKit@0x19d890 10 FinderKit FinderKit@0x19d96a 11 Foundation Foundation@0xc259 12 AppKit AppKit@0x9af897 13 CoreFoundation CoreFoundation@0x58f83 14 CoreFoundation CoreFoundation@0x58ad5 15 CoreFoundation CoreFoundation@0x39470

chofmann's mac only report shows this as the #5 top crash on 10.7 -> https://crash-analysis.mozilla.com/chofmann/20111017/top-mac-crashes-all.txt

> Comments mention downloading and saving pictures. Perhaps a > different manifestation of Bug 670842? Groan. It does sound like another file picker bug :-( Probably different from bug 670842, though -- that seems genuinely to have been fixed.

Nominating for tracking 9 based on the fact that this is a new crash, although not high volume yet, that could be potentially when more people are running 10.7.2.

There were 597 of this bug's crashes in the last week (all three signatures taken together). That makes this bug the #12 Mac topcrasher. It's true bug 670842 was higher-volume -- before OS X 10.7.2 was released, it was the #1 Mac topcrasher. But this bug is still a topcrasher.

With combined signatures, it's #1 top crasher on Mac OS X in 8.0.1. According to some comments, it occurs while downloading a PDF or an image.

> According to some comments, it occurs while downloading a PDF or an image. That just confirms this is a file picker bug. It doesn't get us anywhere near decent steps to reproduce. I've been periodically checking the comments on this stack, looking for detailed enough information that it might help us find STR. So far I've tried a few, but haven't found any STR. It'd be really helpful for others besides myself to be doing this.

I have tried reproducing with my two 10.7.2 boxes to no avail yet. When I get back to Mt. View I will try again. This may be a lot like some other bugs where it is related to timing, making it a little difficult to reproduce. (In reply to Steven Michaud from comment #7) > > According to some comments, it occurs while downloading a PDF or an image. > > That just confirms this is a file picker bug. It doesn't get us anywhere > near decent steps to reproduce. > > I've been periodically checking the comments on this stack, looking for > detailed enough information that it might help us find STR. So far I've > tried a few, but haven't found any STR. It'd be really helpful for others > besides myself to be doing this.

There is more than a 50% correlation to VDH with FinderKit@0x78d87 signature: 55% (6/11) vs. 10% (207/2073) {b9db16a4-6edc-47ec-a1f4-b86292ed211d} (Video DownloadHelper, https://addons.mozilla.org/addon/3006) The other two signatures do not seem to have enough volume for correlations. Will keep trying to get a set of STR.

I have been trying all variations of downloading and saving PDFs, canceling, invoking the file picker over and over with by mouse and keyboard commands on the lab machine using Mozilla/5.0 (Macintosh; Intel Mac OS X 10.7; rv:9.0) Gecko/20100101 Firefox/9.0 - so far no crash. Am wondering if this has something to do with a particular kind of Mac hardware? My lab machine is a Mac mini and the other machine I was testing on is a MBP laptop that is one generation behind the latest one.

Were your tests with the Video DownloadHelper, Marcia? If not, I might try testing with it for an hour or two, to see what happens. As for hardware, all I have is one few-months-old MacBook Pro to test with.

Yes, I tried a bit with that addon as well as Down them all which showed up in another crash report. According to the uptime in the signature summary, I think you are right about spending more time than I did. I will keep plugging away. (In reply to Steven Michaud from comment #11) > Were your tests with the Video DownloadHelper, Marcia? > > If not, I might try testing with it for an hour or two, to see what happens. > > As for hardware, all I have is one few-months-old MacBook Pro to test with.

Other things I tried today, with still no luck: *Starting in 32 bit mode *Tweaking some of the Keyboard settings (FKA) One of the crash reports mentions it happens to them 1/3 of the time. So apparently this isn't one of those crashes that can be triggered that easily (at least from my perspective so far).

Looking again at comments on recent crash stacks, I discovered one that looks interesting: "Was viewing a website's directory page (lots of filenames with links to jpg's). Selected half the file names (299). right clicked and selected 'save link as… Save As dialogue came up. Typed in a file name and clicked 'Save'. Firefox window disappeared." Don't know where I'm going to find hundreds of jpg's to test with, though. I also noticed that none of the reports are for OS X 10.7.3 (all for 10.7.2). 10.7.3 hasn't been released yet. But there are surely a few hundred people testing 10.7.3, and the fact that none of them have showed up might indicate that Apple has fixed this bug in 10.7.3.

I just checked *all* FinderKit signatures from the last week to be sure: Most of them are on 10.7.2, and they're a few on 10.7.1 or 10.7, but none of them are on 10.7.3.

Hi Steven. Can you please provide a link to such a web page in order to test this? I'm also running Mac OS X 10.7.2. Thaks. (In reply to Steven Michaud from comment #14) > Looking again at comments on recent crash stacks, I discovered one that > looks interesting: > > "Was viewing a website's directory page (lots of filenames with links to > jpg's). Selected half the file names (299). right clicked and selected 'save > link as… Save As dialogue came up. Typed in a file name and clicked 'Save'. > Firefox window disappeared." > > Don't know where I'm going to find hundreds of jpg's to test with, though. > > I also noticed that none of the reports are for OS X 10.7.3 (all for > 10.7.2). 10.7.3 hasn't been released yet. But there are surely a few > hundred people testing 10.7.3, and the fact that none of them have showed up > might indicate that Apple has fixed this bug in 10.7.3.

> Can you please provide a link to such a web page in order to test this? I don't have one. If I did I'd have tested with it myself, and posted it :-)

We need to continue to track this so we can verify if it's fixed in 10.7.3. Nothing we can immediately fix.

(Following up comment #14 and comment #15) I've now found one FinderKit crash on OS X 10.7.3 in the last two weeks: bp-adbfa020-c68b-420f-ba36-66aff2111226 So it's likely this bug *isn't* fixed in 10.7.3. But we still have reason to hope there will be fewer of these crashes.

Assigning this tracked bug to Steven in an effort to ensure that all tracked bugs have assignees with next actions.

This is still sitting at #8 for mac crashes on 9.0.1. Is it just a matter of getting a fix from Apple. Is there anything else actionable on our side.

> Is it just a matter of getting a fix from Apple. Hopefully. It's definitely Apple's bug. But Apple has a very poor record with fixing file picker crashes -- new ones pop up as soon as older ones get fixed. > Is there anything else actionable on our side. It's hard to say. It'd take many hours of work just to find out, and I'm probably the only one who can do it. And I'm currently very busy with other bugs. For now I think it's best to wait for OS X 10.7.3, and see what effect it has on the situation.

For what it's worth, I can (as in comment #19) only find one FinderKit crash (at any address) on OS X 10.7.3 (a pre-release version) over the last four weeks: bp-adbfa020-c68b-420f-ba36-66aff2111226 So we still can hope that 10.7.3 will bring a reduction in the number of file picker crashes.

Alex, I don't think it makes sense to track this anymore since that isn't going to make anything happen faster. We probably just need to see how it plays out with 10.7.3.

(In reply to Sheila Mooney from comment #24) > Alex, I don't think it makes sense to track this anymore since that isn't > going to make anything happen faster. We probably just need to see how it > plays out with 10.7.3. Agreed.

The 10.7.3 crash signatures are FinderKit@0x7884a and FinderKit@0x7bdbc. Comments talk about Gmail or downloading.

> The 10.7.3 crash signatures are FinderKit@0x7884a and FinderKit@0x7bdbc. These are quite numerous. Sigh. Doesn't look like these crashes will happen less often on 10.7.3 :-(

https://crash-stats.mozilla.com/report/list?signature=libsystem_c.dylib@0x4cf91 seems to be this crash happening on 10.7.3. All the comments seem to mention downloading.

I just got this crash twice in a row: https://crash-stats.mozilla.com/report/index/bp-dff07427-9ac0-4ef7-a279-6048c2120227 and https://crash-stats.mozilla.com/report/index/bp-dca5bf4d-e033-4d30-b008-41c152120227 I was downloading files from my server, which is the file list generated by lighttpd when no index files are present. I right-clicked on a file link, clicked "Save link as...", but I can't remember whether the file picker dialogue appeared or not before crashing. Of course, now that I'm paying attention to what I'm doing I can't recreate the crash.

I'm getting this crash regularly when using "Save Link As" on linked pictures. bp-dc235115-2de1-4f49-a8a0-a07832120412 bp-33ba5006-632d-4e03-8303-ca9f02120419 bp-ae891169-ac2e-4f80-aaf4-f58072120401 bp-ca13161b-03ff-4c76-b744-b87982120407

Reuben, please provide specific examples, and detailed STR. The specific nature of the examples, and the exact nature of the steps you follow in reproducing these crashes, is very likely *critically important* in reproducing them. I can't tell you how many times I've used "Save Link As" and *not* reproduced these crashes. It may also matter exactly which extensions and plugins you have, exactly which version(s) of FF you're using, and on exactly what version of OS X. But those details should be recoverable from your crash logs.

(In reply to Steven Michaud from comment #31) > Reuben, please provide specific examples, and detailed STR. > > The specific nature of the examples, and the exact nature of the steps you > follow in reproducing these crashes, is very likely *critically important* > in reproducing them. I've only reproduced this problem while browsing imageboards (as those are the only places where I use "Save Link As"). I don't have STR to reproduce the crash reliably, it just seems to crash every now and then. Try saving a few images using "Save Link As" in an imageboard (like http://kusabax.cultnet.net/simg/) > It may also matter exactly which extensions and plugins you have, exactly > which version(s) of FF you're using, and on exactly what version of OS X. > But those details should be recoverable from your crash logs. I'll try to reproduce it in safe mode. I've been seeing those in all Nightlies for the past month or so. OS X 10.7.3 (11D50).

Adding Mac OS X 10.7.4 crash signature.

Haven't been able to reproduce this in a while now.

I've been seeing this fairly regularly on 10.7.4 with Firefox 13.0. The crashes are pointing to FinderKit- is this likely an Apple problem? Has someone reported it to them? If it's not an Apple issue, is someone looking into it? Perhaps someone with good knowledge might be able to provide some pointers where to look...?

> is this likely an Apple problem? Yes, almost certainly. > Has someone reported it to them? No, because there's not yet enough information to make a good bug report. If you want to help, please find a way to reproduce this bug on demand. At a minimum this will involve finding a detailed set of steps that always (or almost always) makes the bug happen. You may also find that your steps don't work unless a particular plugin and/or extension is loaded. This is *not* easy work. But it would be extremely valuable if you could accomplish it. Theses crashes are deep in Apple system code, so this bug will (most likely) stay dead in the water until someone finds good steps-to-reproduce or Apple fixes it. Since variants of this bug have effected all versions of OS X 10.7, I think it's very unlikely that Apple will fix it. But if we can reproduce it, we may be able to find a workaround. At the very least, once we're able to reproduce this bug we can open a bug report with Apple.

(In reply to Steven Michaud from comment #37) > > is this likely an Apple problem? > > Yes, almost certainly. > > > Has someone reported it to them? > > No, because there's not yet enough information to make a good bug report. > Ah, I see. > If you want to help, please find a way to reproduce this bug on demand. At > a minimum this will involve finding a detailed set of steps that always (or > almost always) makes the bug happen. You may also find that your steps > don't work unless a particular plugin and/or extension is loaded. > Fair enough. I've made some efforts to reproduce it, but haven't been able to. It was happening a lot, then it stopped for a while, now it's happened a few times again. There's no obvious correlation I can see which would be associated with the crashes. > it's very unlikely that Apple will fix it. But if we can reproduce it, we > may be able to find a workaround. At the very least, once we're able to > reproduce this bug we can open a bug report with Apple. I'll take another look at reproducing it then.

Some correlations show that many of the crashes in [@ libsystem_c.dylib@0x4cf91 ] are correlated to Firebug: libsystem_c.dylib@0x4cf91|EXC_BAD_ACCESS / KERN_INVALID_ADDRESS (22 crashes) 41% (9/22) vs. 6% (81/1279) firebug@software.joehewitt.com (Firebug, https://addons.mozilla.org/addon/1843) 27% (6/22) vs. 9% (116/1279) {b9db16a4-6edc-47ec-a1f4-b86292ed211d} (Video DownloadHelper, https://addons.mozilla.org/addon/3006) 27% (6/22) vs. 11% (141/1279) {d10d0bf8-f5b5-c8b4-a8b2-2b9879e08c5d} (Adblock Plus, https://addons.mozilla.org/addon/1865) 18% (4/22) vs. 4% (49/1279) {e4a8a97b-f2ed-450b-b12d-ee082ba24781} (Greasemonkey, https://addons.mozilla.org/addon/748) 14% (3/22) vs. 2% (31/1279) {DDC359D1-844A-42a7-9AA1-88A850A938A8} (DownThemAll!, https://addons.mozilla.org/addon/201) 9% (2/22) vs. 0% (2/1279) pig1717@gmail.com 9% (2/22) vs. 1% (13/1279) artur.dubovoy@gmail.com (Flash Video Downloader, https://addons.mozilla.org/addon/6584) 9% (2/22) vs. 1% (15/1279) {19503e42-ca3c-4c27-b1e2-9cdb2170ee34} (FlashGot, https://addons.mozilla.org/addon/220) 9% (2/22) vs. 1% (17/1279) {e968fc70-8f95-4ab9-9e79-304de2a71ee1} (User Agent Switcher, https://addons.mozilla.org/addon/59) 9% (2/22) vs. 3% (41/1279) {D4DD63FA-01E4-46a7-B6B1-EDAB7D6AD389} (Download Statusbar, https://addons.mozilla.org/addon/26)

For what it's worth, Apple seems to have fixed this bug (or bugs) in MountainLion. All of this bug's signatures are pretty high in the list of Mac topcrashers (at least for FF releases). But there don't appear to any signatures (reasonably high in the topcrasher list) corresponding to this bug on OS X 10.8.X.

I'm marking this bug as WORKSFORME as bug crashlog signature didn't appear from a long time (over half year) in Firefox (except some obsolete Fx <49, no crashes starting in Fx49).