[Webkit-unassigned] [Bug 25475] Show the filename and first line for "(program)" in the Profiler/Debugger

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=25475





------- Comment #4 from pmuellr at yahoo.com  2009-06-01 20:23 PDT -------
I discovered that FireBug already supports a feature like I suggested in
comment #3.  They look for a line of the format

   //@ sourceURL=<url>

in eval()'d scripts, and use that as the source url for the script.  The dojo
framework has this baked into their code; at development time, they tend to
eval() a lot of the JavaScript code.  

Mention of this feature here:

   http://fbug.googlecode.com/svn/branches/firebug1.1/docs/ReleaseNotes_1.1.txt

I did a quick prototype of this, by adding the following code at the end of the
WebInspector.Script constructor in WebCore/Inspector/front-end/Script.js:

    // if no URL, look for "//@ sourceURL=" decorator     
    if (!sourceURL) {
        var pattern = /.*\s*\/\/\s*@\s*sourceURL\s*=\s*(\S+)\s*/;
        var match = pattern.exec(source);

        if (match) {
            this.sourceURL = "eval:" + match[1];
        }
    }

Seems to work ok.  Although dojo supports this directly, they sniff the browser
and only add this comment when you're running in Firefox.  I "fixed" that code
and all of a sudden 20-30 "(Program)" entries in the scripts pull-down became
useful names.

I'll post some before/after screenshots as attachments.

Questions:

- would this processing be unconditional, for eval()'d scripts?  Looks like
FireBug initially enabled this only via user setting, but I don't see a setting
for it anymore.  I don't think it would cause much harm to do it
unconditionally.

- this code actually sets the sourceURL to the specified value - prefixed with
"eval:".  I figured that we want these in a different "space" than the other
scripts, and it turns out they sort well this way also.  I'm tempted to prefix
with " eval:" or "#eval:" to get them all to sort to the top.

- wondering if having these fake sourceURLs will cause a problem for any other
processing; I'll poke around.

- the regex needs some additional work - it's failing in a graceful way on some
incorrectly specified lines (also need to figure out where to put tests for
this).


-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.