On Wed, Jun 1, 2016 at 1:05 AM, Benjamin Smedberg <benja...@smedbergs.us> wrote: > You shouldn't need to annotate the file/line separately, because that is > (or at least should be!) the top of the stack.
Yes. Don't get hung up on the lack of annotations. It isn't much of a problem; you can click through easily enough. I have filed bug 1277104 to fix the handful of instances that are showing up in practice, but it'll only be a minor improvement. And adding more detail can be a hindrance. As others mentioned, if you add the line number that will change over time, which means that related crashes will be grouped separately in this search. A more extreme example of this is NS_RUNTIME_ABORT's AbortMessage, which includes the *Process ID*. From a crash aggregation POV this is a disaster, because it all but guarantees that no related crashes will get clustered in this kind of search. > FWIW, we are currently working on changing the signature for crashes with > an AbortMessage (those using NS_RUNTIMEABORT) so that the abort message is > part of the signature. After that works, we should probably do the same > thing or something similar for the new MozCrashReason field. (I'm not sure > why we used different field names for these.) I tried doing a similar search for AbortMessage but I couldn't get it to work. For example: https://crash-stats.mozilla.com/search/?product=Firefox&abort_message=ABORT&_facets=abort_message&_columns=date&_columns=signature&_columns=product&_columns=version&_columns=build_id&_columns=platform&_columns=abort_message#facet-abort_message Gives a table that looks like this: Rank Abort message Count % 1 abort 20461 100.00 % 2 file 20148 98.47 % 3 line 13681 66.86 % 4 build 6974 34.08 % 5 builds 6744 32.96 % Even ignoring the anti-clustering problems caused with the Process ID (described above), it seems to be bucketing according to individual words in the AbortMessage, rather than the entire AbortMessage. I don't understand what's happening. ---- Anyway, can we just get rid of NS_RUNTIMEABORT? I count 308 occurrences of it in the code vs. 4064 for MOZ_CRASH. Nick _______________________________________________ dev-platform mailing list dev-platform@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-platform
