291374 - RFE: better error reporting for "object has no method named" in XPConnect calls to JS-impl'd ifaces

Status: RESOLVED INACTIVE

(Core :: XPConnect, enhancement)

Description

[Mike Shaver (:shaver emeritus)]

Right now, you don't find out anything but the fact that we tried to call
nsIObserver::observe on _something_, and it didn't have the attribute.

I have a patch to leverage the __LOCATION__ property I added to
component-globals and improve the error reporting to at least include the
filename.  It is not a pretty patch, and I would want to refactor it a bit to
avoid the somewhat-nasty overload of aData-as-stack it introduces in
XPCConvert::ConstructException, but it really makes debugging out-of-whack
components a fair sight better.

I think I'd also like to take a swing at including the object's
.constructor.name property, if present, to give a still-better hint to the
developer. 

Comments very much desired.  (To test this patch, I recommend writing a simple
component that registers against "xpcom-startup" as an observer, and doesn't
implement an |observe| function.  Makes testing in xpcshell a breeze.)

Comment 1

[Mike Shaver (:shaver emeritus)]

Attachment: mildly, but usefully, better reporting

I'm not even going to request review on this, but don't let that keep you from
providing comments!

Comment 2

[Zook Valem]

Severity --> enhancement due to partical summary "REF: "

Comment 3

[timeless]

Comment on attachment 181469 [details] [diff] [review]
mildly, but usefully, better reporting

>Index: xpcwrappedjsclass.cpp
>@@ -1358,9 +1359,51 @@ pre_call_clean_up:
i believe this needs a null check: (and rooting?) (and should you consider
atomizing?)
>+            JSString *str = JS_NewStringCopyZ(cx, "__LOCATION__");
>+            if (JS_ValueToId(cx, STRING_TO_JSVAL(str), &locId) &&
>+                OBJ_GET_PROPERTY(cx, outermostScope, locId, &locval) && 
>+                !XPCConvert::JSData2Native(ccx, &location,
>+                                           locval, type, JS_FALSE,
>+                                           &NS_GET_IID(nsILocalFile), &retval)) {

Comment 4

[Mike Shaver (:shaver emeritus)]

Yes, yes, there are all sorts of robustness issues there; I'm looking for
comments on the general approach, not on my punctuation.

Comment 5

[timeless]

seems ok to me. *shrug* :)

Comment 6

[Mike Shaver (:shaver emeritus)]

Mass abdication.

Comment 7

[BMO Automation]

Per policy at https://wiki.mozilla.org/Bug_Triage/Projects/Bug_Handling/Bug_Husbandry#Inactive_Bugs. If this bug is not an enhancement request or a bug not present in a supported release of Firefox, then it may be reopened.