329882 - bc has no properties console.js

Status: RESOLVED FIXED

(Toolkit Graveyard :: Error Console, defect)

Description

[Henrik Gemal]

using latest nightly build I'm seeing:

Error: bc has no properties
Source File: chrome://global/content/console.js
Line: 61

Comment 1

[Bob Clary [:bc] (inactive)]

/me resents the implication. ;-)

Comment 2

[Simon Bünzli]

Henrik: Have you ever seen this again? Any STR?

One known way to get this error is to install Console², apply some filters, then uninstall the extension and reopen the original Error Console. Clicking one of the four filter buttons fixes that (non-)issue.

Comment 3

[Simon Bünzli]

Henrik: Without any answer to comment #2, there isn't much we can do here.

Please reopen this bug, if you can indicate Steps to Reproduce (if possible on a clean profile).

Comment 5

[Mike Kaply [:mkaply]]

What can't we just add an if statement since we know where the problem occurs?

Comment 6

[Philip Chee]

> What can't we just add an if statement since we know where the problem occurs?

You know the drill. Attach a patch, ask for review. etc.

Comment 7

[Mike Kaply [:mkaply]]

OK, so the bug here has to do with the console mode being in a strange stage.

Instead of the mode being "All", the mode is "Errors Warnings Messages"

So when the code is executed:

var bc = document.getElementById("Console:mode" + aMode);

It's trying to get the element:

"Console:modeErrors Warnings Messages"

which is obviously wrong.

So the question is how we get a mode set to "Errors Warnings Messages" instead of "All"

Comment 8

[Mike Kaply [:mkaply]]

Definitely the fault of Console2.

Because it's XUL files are named the same, it's overwriting XUL data in localstore.rdf.

Console2 uses states for each type of messages (not all or nothing like us). So after running console2, in localstore.rdf we have:

<RDF:Description RDF:about="chrome://global/content/console.xul#ConsoleBox"
                   modeLang="JS CSS XML"
                   modeCtx="Chrome Content"
                   mode="Errors Warnings Messages" />

That mode is what screws us up

That's an invalid mode for the regular JS console.

It sucks that Console2 is doing this, but I think we should have some sort of workaround so that error doesn't pop up.

Incidentally, easiest way to fix the error locally is to just click on All to reset the value.

Comment 9

[Mike Kaply [:mkaply]]

Attachment: Attempted fix

If we fail to get bc, default to "All"

The only other thing we could do here is before attempting to get bc, check to see if aMode is one of the four valid values, but that seems like a lot of work to me for an edge case.

Comment 10

[Philip Chee]

How about:

var bc = document.getElementById("Console:mode" + aMode);
bc? null: bc = document.getElementById("Console:modeAll");
bc.setAttribute("checked", true);

Comment 11

[Simon Bünzli]

(In reply to comment #10)
Cleanest would be
- var bc = document.getElementById("Console:mode" + aMode);
+ var bc = document.getElementById("Console:mode" + aMode) || document.getElementById("Console:modeAll");

And please add a comment noting why the mode could get invalid at all...

(In reply to comment #8)
> It sucks that Console2 is doing this

Well, it was meant to be so great that you wouldn't ever think of uninstalling it before it would finally get merged into Toolkit proper... ;-)

But anyway: Why add code for an edge case with a simple work-around?

Comment 12

[Mike Kaply [:mkaply]]

> (In reply to comment #8)
> > It sucks that Console2 is doing this
> 
> Well, it was meant to be so great that you wouldn't ever think of uninstalling
> it before it would finally get merged into Toolkit proper... ;-)

I hope you didn't take offense to my sucks - didn't mean that Console2 sucks :)

And sometimes you have to disable your extensions. That's just life :)

> But anyway: Why add code for an edge case with a simple work-around?

Manly because, if we can help it, browser code shouldn't put errors in the JS console if it can help it. (even if only extension developers will see it)

Comment 13

[Mike Kaply [:mkaply]]

Attachment: Better patch

I like Simon's code better.

Comment 14

[Reed Loden [:reed]]

http://hg.mozilla.org/mozilla-central/rev/2367bee47379