552088 - crash reports with email addresses but no comment don't all bubble up to the top of the "comments" tab

Status: RESOLVED WONTFIX

(Socorro :: Database, task)

Description

[Wayne Mery (:wsmwk)]

crash reports with email addresses but no comment don't all bubble up to the top of the "comments" tab

bp-8fc32203-633a-4183-bd75-405c62100308 has an email address, but no comment, and is not listed in the "comments" tab of the query from which I obtained the crash link https://crash-stats.mozilla.com/report/list?product=Thunderbird&build_id=&query_search=signature&query_type=exact&query=memcpy%20|%20CWaveHeader%3A%3ARead%28CFrameFifo%26%2C%20unsigned%20char*%26%2C%20unsigned%20int%26%2C%20int%26%29&date=&range_value=4&range_unit=weeks&process_type=all&plugin_field=&plugin_query_type=&plugin_query=&do_query=1&signature=memcpy%20|%20CWaveHeader%3A%3ARead%28CFrameFifo%26%2C%20unsigned%20char*%26%2C%20unsigned%20int%26%2C%20int%26%29&missing_sig=&page=1

bp-f8b8457f-2941-4101-983e-8e6cd2100311 is another example

Comment 1

[Laura Thomson :laura]

Is this now a requirement?

Comment 2

[Wayne Mery (:wsmwk)]

(In reply to comment #1)
> Is this now a requirement?

ludo, standard8, bienvenu, 

I filed this bug because (I think) we want testcases for most crashes - as I understand it we often need a testcase to fix a crash, or it makes it much easier. (Is it fair to say 20-30% of crashes absolutely need testcases or steps if they are to get fixed?) 

testcases are also useful for automated tests, and to verify a crash is indeed really fixed - fixed from the reporters' perspective not just in theory.  Further, when we don't have at least a couple testcases with live reporter's attached to them we often must wait until a point release to verify a crash is fixed, and that's a long cycle time for feedback and getting things fixed.

The answer to Laura's question perhaps depends on whether 
* We are soon (6 months to a year?) getting a much better "crash fixing process" from moco, mome, or a combination?
* Do we want to wait, or from a quality perspective can we afford to wait, for something better?  Or is it a moot point because we have more work than we can currently handle anyway? :)
* Is results from the manual process (described below) worth the work?
* Is it easy for moco to deliver more addresses per comment 0? 

Under the current "manual" process I email some crash  reporters, and I probably get a 10-20% success rate of getting either a testcase or a reporter signed up to a bug, because :
a) bugzilla reports don't yield much - Thunderbird users don't report crash bugs at the same rate as firefox
b) crash reports often doesn't have the data needed to debug a crash
c) probably less than half of crashes with comments have email addresses
e) but probably an equal number of crashes that don't have comments DO have addresses

This bug is to address item e) above.

Comment 3

[Wayne Mery (:wsmwk)]

(In reply to comment #1)
> Is this now a requirement?

Laura, did I answer your question?  Perhaps this could be done with the crashes listed in the comment tab as having a secondary sort on email address.

Comment 4

[Smokey Ardisson (offline for a while; not following bugs - do not email)]

See also bug 550538 (suggestion 2 in that bug is very similar to comment 3 here); doing something with "sorting" that works generically for comments+email and comments+URL may end up being an workable solution for both.

Comment 5

[Laura Thomson :laura]

Should do this at the query level.  Reassigning to database component.