Print out the error when loading an extension fails.

Print out the error when loading an extension fails.

This is primarily useful for extension developers in helping them diagnose what's wrong in their extension manifest.


Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=134408

[Wez, 2012-04-11 23:09:07]

PTAL

[Yoyo Zhou, 2012-04-11 23:20:06]

What's the use case for this? Where you've put it, it only runs at profile
startup.

[Wez, 2012-04-11 23:29:20]

On 2012/04/11 23:20:06, Yoyo Zhou wrote:
> What's the use case for this? Where you've put it, it only runs at profile
> startup.

To provide some feedback on why a package won't load, for package developers.

It would make sense to show an error for loads in general, if we're happy that
that won't result in users seeing spammy warning output.

[Yoyo Zhou, 2012-04-11 23:40:17]

What I mean is, what kind of errors is this going to catch? It seems like it
would only affect extensions that are already installed, which would mean a
backwards-incompatible change to manifest parsing. Or, I suppose it happens if
you have an unpacked extension that's changed on disk and has become invalid.
But in that case, the extension goes away and one has to attempt to load it
unpacked again.

Anyway, I guess this LGTM, but for developers it's probably more useful that
there is validation that prints the error when they try to load an extension
unpacked.

[Wez, 2012-04-26 21:05:16]

On 2012/04/11 23:40:17, Yoyo Zhou wrote:
> What I mean is, what kind of errors is this going to catch? It seems like it
> would only affect extensions that are already installed, which would mean a
> backwards-incompatible change to manifest parsing. Or, I suppose it happens if
> you have an unpacked extension that's changed on disk and has become invalid.
> But in that case, the extension goes away and one has to attempt to load it
> unpacked again.
> 
> Anyway, I guess this LGTM, but for developers it's probably more useful that
> there is validation that prints the error when they try to load an extension
> unpacked.

Would it make more sense to log the error from LoadExtension() itself, then,
rather than patching up call-sites?

[Wez, 2012-04-26 21:43:38]

On 2012/04/11 23:40:17, Yoyo Zhou wrote:
> What I mean is, what kind of errors is this going to catch? It seems like it
> would only affect extensions that are already installed, which would mean a
> backwards-incompatible change to manifest parsing. Or, I suppose it happens if
> you have an unpacked extension that's changed on disk and has become invalid.
> But in that case, the extension goes away and one has to attempt to load it
> unpacked again.
> 
> Anyway, I guess this LGTM, but for developers it's probably more useful that
> there is validation that prints the error when they try to load an extension
> unpacked.

The other call-sites use ReportExtensionLoadError(), so I've updated this one to
match - extension developers can then run Chrome --v=1 to see detailed load
errors.

PTAL!

[Yoyo Zhou, 2012-04-27 21:07:32]

LGTM

[commit-bot: I haz the power, 2012-04-27 22:46:46]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/wez@chromium.org/10050032/7001

[commit-bot: I haz the power, 2012-04-28 00:38:35]

Change committed as 134408