Bug 827407 is a top-crasher in Fx21 and per https://bugzilla.mozilla.org/show_bug.cgi?id=827407#c23 , it will really help the mobile team speed up on investigation if we can help with Flash versions here. :Kairo mentioned this could be possible to gather from the raw data given we have the annotation in Fx21 per Bug 839456 .

I think we need someone to do a pig query here, unless we have that field in the DB. Benjamin has done some querying with pig, so CCing him as well.

I have been trying to work on this. My access to hbase://crash_reports seems to no longer work for reading. I'm accessing from sp-admin01 and this worked a few weeks ago. https://gist.github.com/7ae89f2216e6e39ade1a I pinged someone in #metrics.

(In reply to bhavana bajaj [:bajaj] from comment #0) > Bug 827407 is a top-crasher in Fx21 and per > https://bugzilla.mozilla.org/show_bug.cgi?id=827407#c23 , it will really > help the mobile team speed up on investigation if we can help with Flash > versions here. > > :Kairo mentioned this could be possible to gather from the raw data given we > have the annotation in Fx21 per Bug 839456 . Here's the count of instances containing the FlashVersion and the signature from bug 827407. [sdeckelmann@sp-admin01.phx1 flash_versions]$ cat * | sort | uniq -c 440 3 10.3 r186 2 11.0 r1 6 11.1 r102 7 11.1 r111 74 11.1 r112 1633 11.1 r115

Dates are: 4/12/13 to 4/25/13 pig -param start_date=130412 -param end_date=130425 findFlashVersion.pig

Thanks, Selena! It looks like this problem worsened in Flash 11.1r115 or, more likely, most of our users are good about updating to the latest version of Flash.

Is there any way to get the release number? r115 is about 20 different versions over the last year plus.

I went back to March 1-30, and got this: [sdeckelmann@sp-admin01.phx1 pig]$ cat flash-march1-30/* | sort | uniq -c cat: flash-march1-30/_logs: Is a directory 3593 2 11.1 r102 1 11.1 r111 20 11.1 r112 536 11.1 r115 [sdeckelmann@sp-admin01.phx1 pig]$ cat flash_versions/* flash-april1-12/flash-versions--130401-130411-1/* | sort | uniq -c 1109 3 10.3 r186 3 11.0 r1 14 11.1 r102 16 11.1 r111 107 11.1 r112 2675 11.1 r115 So, that was a similar number of days, but we se fewer NULLs for FlashVersions, and quite a few more crashes for r115 in April.

(In reply to Kevin Brosnan [:kbrosnan] from comment #6) > Is there any way to get the release number? r115 is about 20 different > versions over the last year plus. I don't have any additional release number information from a raw crash. Acquiring it would need to be a change to breakpad, if it can be done.

Breakpad updates won't change things ;-) This is the version number that Flash reports to Firefox from it's description field: http://mxr.mozilla.org/mozilla-central/source/dom/plugins/base/nsPluginHost.cpp#2734 So we actually just don't have any better information at the moment.

(In reply to Kevin Brosnan [:kbrosnan] from comment #6) > Is there any way to get the release number? r115 is about 20 different > versions over the last year plus. From what I understand, this is 1:1 what our annotation feeds into the crash reports - and as Benjamin says, it looks like that's all we have for now - unless you know a better way to get the full version number inside our process (and if so, please file a bug on Firefox for Android to update the annotation with that).

(In reply to Benjamin Smedberg [:bsmedberg] from comment #9) > Breakpad updates won't change things ;-) This is the version number that > Flash reports to Firefox from it's description field: > http://mxr.mozilla.org/mozilla-central/source/dom/plugins/base/nsPluginHost. > cpp#2734 > > So we actually just don't have any better information at the moment. Heh. I was afraid of that. :/ Thanks for the pointer.

Our only hope is to use the debug IDs and match against the list in bug 794265.

I don't think I need a more detailed breakdown, since the bug affects all major releases (in varying degrees).