wallpaper-webui: Make sure the wallpaper thumbnails are appropriately scaled.

wallpaper-webui: Make sure the wallpaper thumbnails are appropriately scaled.

This gets rid of an implicit ImageSkia -> SkBitmap conversion.

BUG=141146
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=151945

[Ivan Korotkov, 2012-08-15 20:36:03]

lgtm

http://codereview.chromium.org/10857019/diff/1/chrome/browser/ui/webui/option...
File chrome/browser/ui/webui/options2/chromeos/user_image_source.cc (right):

http://codereview.chromium.org/10857019/diff/1/chrome/browser/ui/webui/option...
chrome/browser/ui/webui/options2/chromeos/user_image_source.cc:12: #include
"chrome/browser/ui/webui/web_ui_util.h"
Ouch.

http://codereview.chromium.org/10857019/diff/1/chrome/browser/ui/webui/option...
File chrome/browser/ui/webui/options2/chromeos/wallpaper_thumbnail_source.cc
(right):

http://codereview.chromium.org/10857019/diff/1/chrome/browser/ui/webui/option...
chrome/browser/ui/webui/options2/chromeos/wallpaper_thumbnail_source.cc:160:
ui::ScaleFactor scale_factor;
Does the URL ever include username@domain? In that case, any such occurrences
without trailing @<factor>x part will be parsed improperly.

[sadrul, 2012-08-15 20:53:28]

http://codereview.chromium.org/10857019/diff/1/chrome/browser/ui/webui/option...
File chrome/browser/ui/webui/options2/chromeos/wallpaper_thumbnail_source.cc
(right):

http://codereview.chromium.org/10857019/diff/1/chrome/browser/ui/webui/option...
chrome/browser/ui/webui/options2/chromeos/wallpaper_thumbnail_source.cc:160:
ui::ScaleFactor scale_factor;
On 2012/08/15 20:36:03, Ivan Korotkov wrote:
> Does the URL ever include username@domain? In that case, any such occurrences
> without trailing @<factor>x part will be parsed improperly.

Good point! I discussed with bshe@, and username@domain indeed can be part of
the URL. However, the url is of the format
'chrome://.../custom_user@domain?id=...' etc., and it looks like the parsing
code will return NONE for the scale factor in this case (even if domain starts
with '1x' or '2x'), which is the current behaviour anyway. I think this is safe?

[pkotwicz, 2012-08-15 21:25:21]

lgtm

[sadrul, 2012-08-15 22:28:53]

http://codereview.chromium.org/10857019/diff/1/chrome/browser/ui/webui/option...
File chrome/browser/ui/webui/options2/chromeos/wallpaper_thumbnail_source.cc
(right):

http://codereview.chromium.org/10857019/diff/1/chrome/browser/ui/webui/option...
chrome/browser/ui/webui/options2/chromeos/wallpaper_thumbnail_source.cc:160:
ui::ScaleFactor scale_factor;
On 2012/08/15 20:53:28, sadrul wrote:
> On 2012/08/15 20:36:03, Ivan Korotkov wrote:
> > Does the URL ever include username@domain? In that case, any such
occurrences
> > without trailing @<factor>x part will be parsed improperly.
> 
> Good point! I discussed with bshe@, and username@domain indeed can be part of
> the URL. However, the url is of the format
> 'chrome://.../custom_user@domain?id=...' etc., and it looks like the parsing
> code will return NONE for the scale factor in this case (even if domain starts
> with '1x' or '2x'), which is the current behaviour anyway. I think this is
safe?

I have also submitted http://codereview.chromium.org/10837270/ that should help
avoiding similar problem in the future.

[Ivan Korotkov, 2012-08-16 08:44:17]

http://codereview.chromium.org/10857019/diff/1/chrome/browser/ui/webui/option...
File chrome/browser/ui/webui/options2/chromeos/wallpaper_thumbnail_source.cc
(right):

http://codereview.chromium.org/10857019/diff/1/chrome/browser/ui/webui/option...
chrome/browser/ui/webui/options2/chromeos/wallpaper_thumbnail_source.cc:160:
ui::ScaleFactor scale_factor;
On 2012/08/15 22:28:53, sadrul wrote:
> On 2012/08/15 20:53:28, sadrul wrote:
> > On 2012/08/15 20:36:03, Ivan Korotkov wrote:
> > > Does the URL ever include username@domain? In that case, any such
> occurrences
> > > without trailing @<factor>x part will be parsed improperly.
> > 
> > Good point! I discussed with bshe@, and username@domain indeed can be part
of
> > the URL. However, the url is of the format
> > 'chrome://.../custom_user@domain?id=...' etc., and it looks like the parsing
> > code will return NONE for the scale factor in this case (even if domain
starts
> > with '1x' or '2x'), which is the current behaviour anyway. I think this is
> safe?
> 
> I have also submitted <crxref style="background-image:
url(chrome-extension://ppnlfijacbhfpjgkafbeiaklmgoljooo/res/type/codereview.png);
">http://codereview.chromium.org/10837270/</crxref> that should help
> avoiding similar problem in the future.

Thanks! With CL 10837270 submitted it should be safe to use.  Prior to it,
|path| would only contain username, as @domain would be cut in absence of
@factor.