Fix pepper plugin description regression

Fix pepper plugin description regression

CL http://codereview.chromium.org/6162008 broke descriptions for pepper plugins.
This puts it back to the previous state (we don't have a way to separately give
mime type descriptions vs plugin descriptions)

BUG=cros:11298
TEST=Flash on CNN

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=72569

[evanm, 2011-01-25 23:32:08]

I don't understand this.  How did the referenced change break this?

[evanm, 2011-01-25 23:34:30]

Ohhh, I see the old code did this:


416     info.file_description = ASCIIToWide(plugins[i].description);	   417    
info.file_extensions = ASCIIToWide(plugins[i].file_extensions);
418     info.file_description = ASCIIToWide(plugins[i].type_descriptions);

Note 416 and 418 touched the same variable
http://codereview.chromium.org/6162008/diff/13001/chrome/browser/plugin_servi...


PS: this API is super weird -- we sorta have to twist things to support the
stuff PPAPI provides.  It would be nice to map things into WebPluginInfo more
directly.

[evanm, 2011-01-25 23:37:32]

er, and LGTM too

[piman, 2011-01-25 23:44:26]

On Tue, Jan 25, 2011 at 3:34 PM, <evanm@google.com> wrote:

> Ohhh, I see the old code did this:
>
>
> 416     info.file_description = ASCIIToWide(plugins[i].description);
> 417
> info.file_extensions = ASCIIToWide(plugins[i].file_extensions);
> 418     info.file_description = ASCIIToWide(plugins[i].type_descriptions);
>
> Note 416 and 418 touched the same variable
>
>
http://codereview.chromium.org/6162008/diff/13001/chrome/browser/plugin_servi...
>
>
Yes...


>
> PS: this API is super weird -- we sorta have to twist things to support the
> stuff PPAPI provides.  It would be nice to map things into WebPluginInfo
> more
> directly.


Agreed, it needs refactoring. We ought to be able to specify both the plugin
description and the per-mime-type description on the command line, but
changing the format would break other things right now.
I want to first fix the regression (which is blocking Chrome OS, and is a
trivial fix), then look at refactoring.


>
>
> http://codereview.chromium.org/6327019/
>