Add support for receiving text from system services on Mac....

Add support for receiving text from system services on Mac.

BUG=20868
TEST=manual

Review: http://codereview.chromium.org/3158026


Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=57087

[Nico, 2010-08-23 00:58:54]

LG

insertText: is ime-related and hence inherently tricky, so maybe James can take
a look, too :-)

http://codereview.chromium.org/3158026/diff/2001/3002
File chrome/browser/renderer_host/render_widget_host_view_mac.mm (right):

http://codereview.chromium.org/3158026/diff/2001/3002#newcode2309
chrome/browser/renderer_host/render_widget_host_view_mac.mm:2309: }
can you match the style in -writeSelectionToPasteboard:types:?

(i.e.

if (!string) return NO;
return [self insertText:string];
)?

[gmorrita, 2010-08-23 01:34:11]

LGTM. Thank you for doing this!
Testing PGP service manually would be great 
because the bug reporter requested it to work.

[James Su, 2010-08-23 04:45:20]

http://codereview.chromium.org/3158026/diff/2001/3002
File chrome/browser/renderer_host/render_widget_host_view_mac.mm (right):

http://codereview.chromium.org/3158026/diff/2001/3002#newcode2307
chrome/browser/renderer_host/render_widget_host_view_mac.mm:2307: [self
insertText:string];
Calling [self insertText:] while an input method is activated and a text is
being composed may cause problem. But we should not encounter this situation at
all, so it's probably better to add a DCHECK(!hasMarkedText_) here to ensure it.

[dmac, 2010-08-23 16:15:34]

On 2010/08/23 00:58:54, Nico wrote:
> LG
> 
> insertText: is ime-related and hence inherently tricky, so maybe James can
take
> a look, too :-)
> 
> http://codereview.chromium.org/3158026/diff/2001/3002
> File chrome/browser/renderer_host/render_widget_host_view_mac.mm (right):
> 
> http://codereview.chromium.org/3158026/diff/2001/3002#newcode2309
> chrome/browser/renderer_host/render_widget_host_view_mac.mm:2309: }
> can you match the style in -writeSelectionToPasteboard:types:?
> 
> (i.e.
> 
> if (!string) return NO;
> return [self insertText:string];
> )?

Done

[dmac, 2010-08-23 16:15:54]

On 2010/08/23 01:34:11, morrita wrote:
> LGTM. Thank you for doing this!
> Testing PGP service manually would be great 
> because the bug reporter requested it to work.

I've tested a couple of different services. Input, Output, and IO.

[dmac, 2010-08-23 16:17:03]

On 2010/08/23 04:45:20, James Su wrote:
> http://codereview.chromium.org/3158026/diff/2001/3002
> File chrome/browser/renderer_host/render_widget_host_view_mac.mm (right):
> 
> http://codereview.chromium.org/3158026/diff/2001/3002#newcode2307
> chrome/browser/renderer_host/render_widget_host_view_mac.mm:2307: [self
> insertText:string];
> Calling [self insertText:] while an input method is activated and a text is
> being composed may cause problem. But we should not encounter this situation
at
> all, so it's probably better to add a DCHECK(!hasMarkedText_) here to ensure
it.

Actually it is possible. I entered some text in Hiragana and while I was
composing I triggered the service. It did a replace of the text which was
different than textedit or safari. I have updated the patch to imitate how
TextEdit and Safari handle it. Please take a look and see what you think.

[James Su, 2010-08-23 18:39:36]

LGTM.

On 2010/08/23 16:17:03, dmac wrote:
> On 2010/08/23 04:45:20, James Su wrote:
> > http://codereview.chromium.org/3158026/diff/2001/3002
> > File chrome/browser/renderer_host/render_widget_host_view_mac.mm (right):
> > 
> > http://codereview.chromium.org/3158026/diff/2001/3002#newcode2307
> > chrome/browser/renderer_host/render_widget_host_view_mac.mm:2307: [self
> > insertText:string];
> > Calling [self insertText:] while an input method is activated and a text is
> > being composed may cause problem. But we should not encounter this situation
> at
> > all, so it's probably better to add a DCHECK(!hasMarkedText_) here to ensure
> it.
> 
> Actually it is possible. I entered some text in Hiragana and while I was
> composing I triggered the service. It did a replace of the text which was
> different than textedit or safari. I have updated the patch to imitate how
> TextEdit and Safari handle it. Please take a look and see what you think.