When enabled by the fix for bug 745980, the crash reporter on Windows identifies the app by its origin rather than its name, which makes its verbiage read strangely; and on Mac the reporter presumably identifies the app as "Webapp Runtime"; since it seems to get the name from nsXREAppData::name, which is the origin on Windows (because of the fix for bug 747409) and "Webapp Runtime" on Mac <http://mxr.mozilla.org/mozilla-central/source/webapprt/application.ini.in#12>. The reporter should identify the app by its name.
We could fix this by overriding the app name on all platforms (now that the profile directory contains the webapp name).
(Note that we're already using the profile name on Windows and Linux, only Mac is missing).
Ok, I inserted a crasher, and my build of Firefox crashes into the crash reporter when I run it, but the crash reporter doesn't run when I crash the same app running under webrt on the mac :(
For sanity's sake I installed Linux (Ubuntu 13.10) and compiled my crasher there. Again, the crasher brings up the crash reporter when I cause the crash in Firefox, but not when I cause the crash in the app running from the web app runtime. I am going to install Windows and learn how to compile there next so I can test that too. Am I doing something wrong or do we need to open new bugs for the crash reporter not running for web app runtime crashes?
We should identify apps by their origin not their name (or by both, the but the origin is more important). Names are more variable and can be localized, while origins aren't. In terms of data analysis, origin is far better. And yes, there are issues with the crash reporter. I'm not sure what they are yet, though. I inject a crash on Windows and nothing happens.
(In reply to Donovan Preston [:fzzzy] from comment #4) > For sanity's sake I installed Linux (Ubuntu 13.10) and compiled my crasher > there. Again, the crasher brings up the crash reporter when I cause the > crash in Firefox, but not when I cause the crash in the app running from the > web app runtime. > > I am going to install Windows and learn how to compile there next so I can > test that too. > > Am I doing something wrong or do we need to open new bugs for the crash > reporter not running for web app runtime crashes? Commenting out some DLL blocklist code (that prevents the crash reporter from running, see bug 958344), the crash reporter is correctly executed on Windows.
Per bug 1238079, we're going to disable the desktop web runtime and remove it from the codebase, so we won't fix these bugs in it.