If a toolkit/devtools/debugger/tests xpcshell test causes dbg-server or dbg-client code to throw an error, the test still passes. This is because: 1) both client and server contain 'catch' clauses which report the error via Cu.reportError; and 2) we never register a console listener. We should register a console listener, and if it gets an nsIScriptError, we should throw.

Try: https://tbpl.mozilla.org/?tree=Try&rev=73d997fe8c3d

(The patch also includes a bunch of minor changes like changing throw 'foo' to throw Error('foo') so we'll get backtraces, and printing stacks so Emacs can pick up the filenames and line numbers.)

Comment on attachment 618709 [details] [diff] [review] Have toolkit/devtools/debugger xpcshell tests register a listener for Components.utils.reportError, so tests fail when they throw an exception. Review of attachment 618709 [details] [diff] [review]: ----------------------------------------------------------------- ::: toolkit/devtools/debugger/server/dbg-server.js @@ +63,5 @@ > } > } > > +/* Turn the error e into a string, without fail. */ > +function safeErrorString(e) { We use aError or somesuch in browser code. ::: toolkit/devtools/debugger/tests/unit/head_dbg.js @@ +16,5 @@ > Cu.import("resource:///modules/devtools/dbg-server.jsm"); > Cu.import("resource:///modules/devtools/dbg-client.jsm"); > > +// Convert an nsIScriptError 'flags' value into an appropriate string. > +function scriptErrorFlagsToKind(flags) { aFlags

We should land both of these at the same time.

Backed out the latest fx-team merge whole-sale because of Ts regressions: https://hg.mozilla.org/mozilla-central/rev/b0200dab0ccc Please reland only after investigating and fixing the regression. Thanks!

I messed up and backed out the wrong range, sorry about that. I backed out my backout, so this is still on mozilla-central: https://hg.mozilla.org/mozilla-central/rev/dec5b367c421