[Webkit-unassigned] [Bug 137297] Allowing eval'ed code to be named in stack traces

bugzilla-daemon at webkit.org bugzilla-daemon at webkit.org
Wed Jul 5 11:33:35 PDT 2017


https://bugs.webkit.org/show_bug.cgi?id=137297

--- Comment #5 from Joseph Pecoraro <joepeck at webkit.org> ---
> Safari doesn't support //# sourceURL= in script tags either. Would you
> recommend that I create another bug for that?

Safari supports sourceURL in <script>, Function, and eval for locations logged to Web Inspector. For example a `console.trace()` or uncaught exception case you will see the source URL location:

Global Code (this is missing sourceURL):

    [Trace] Trace
        Global Code (test-page.html:25)
    [Error] Error
        Global Code (test-page.html:26)


Function: (sourceURL shows up for the anonymous function)

    [Trace] Trace
        anonymous (namedFunction.js:3)
        Global Code (test-page.html:36:202)
    [Error] Error
        anonymous (namedFunction.js:4)
        Global Code (test-page.html:36:194)

eval: (sourceURL shows up for the eval code)

    [Trace] Trace
        Eval Code (namedEval.js:2)
        eval
        Global Code (test-page.html:39)
    [Error] Error
        Eval Code (namedEval.js:3)
        eval
        Global Code (test-page.html:39)

JavaScriptCore just doesn't use those locations by default in its Error.stack reporting.

It looks like that comes from Interpreter::stackTraceAsString which ultimately gets the ScriptExecutable's sourceURL which is the SourceProvider's url, not the SourceProvider's sourceURL. I think it would be reasonable to use the best name available in all error stacks.

So there are two issues here:

    1. sourceURL directive should show up in Error.stack.
    2. sourceURL directive in <script> global code wasn't displayed properly in Web Inspector.

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-unassigned/attachments/20170705/81c4c3c9/attachment.html>


More information about the webkit-unassigned mailing list