cc'ing some more keyboard folks

See also issue 19926. Not sure if they are the same story.

(No comment was entered for this change.)

(No comment was entered for this change.)

I don't know much about how in works under the Carbon event model (Flash 10.0); 
https://bugzilla.mozilla.org/show_bug.cgi?id=357670 has a bunch of discussion that may or may not be 
useful.

In the Cocoa event model (Flash 10.1), it's supposed to work as described at the end of 
<https://wiki.mozilla.org/Mac:NPAPI_Event_Models>; we just haven't wired that up yet.

My suggestion would be to start with the Cocoa event model version, since that's the future, and we can see 
how important the Carbon version is based on when Flash 10.1 comes out and how many other legacy plugins 
really need IME support.

(No comment was entered for this change.)

I've heard WebKit has a BSD-licenced IME input window for the Cocoa event plugin case; we should strongly 
consider using that, both to avoid pointless duplication of work and to give users a consistent interface between 
browsers.

 Issue 37558  has been merged into this issue.

Gecko is close to having something that we can hopefully leverage that supports both 
10.5 and 10.6, building on the code mentioned in comment 7.

Stuart: is it possible to do this for M5?

It depends a lot on when the Mozilla code that we want to share is finished and working; I gather that getting 
something that works on both 10.5 and 10.6 and doesn't use SPI (as Mac WebKit apparently does) has been pretty 
difficult, and we don't want to duplicate that effort.

(No comment was entered for this change.)

(No comment was entered for this change.)

(No comment was entered for this change.)

 Issue 41614  has been merged into this issue.

(No comment was entered for this change.)

 Issue 46425  has been merged into this issue.

Unfortunately, it looks like using the Mozilla code[1] won't be particularly straightforward:
- On 10.6, there's a method overridden in the NSView corresponding to the plugin; we don't have a plugin NSView, and the view is in the browser process, whereas the key handling is in the plugin process. Also, we need work with NSEvents, which the plugin process doesn't have (although we could probably construct fake ones from the structures we do have).
- On 10.5, it involves the existing Gecko TSM code, so it's more extensive than what shows up in the patch. We need EventRefs, although again we might be able to fake them (by creating fake NSEvents, and asking for their eventRef).

1: https://bug512886.bugzilla.mozilla.org/attachment.cgi?id=450261

Does this work in Safari? As this is difficult, suggest pushing to Mstone:7

Safari has an implementation of this spec, yes, and Firefox does now as well.

Things like "opt-u o" work in Safari, but inputting CJK characters with input methods does not work. I just tested it on Snow Leopard with Safari 5.0 (6533.16) and Flash 10.0 r45. So I think it also depends on the Flash version.

See comment 5; Flash 10.0 and 10.1 use different event models with completely different IME handling. It's unlikely we'll ever support IME for the deprecated Carbon event model.

(No comment was entered for this change.)

I'm afraid I have to punt this one more time :( It should definitely happen for M9.

Pulling back to M8 since I have a patch for 10.6 under review.

The following revision refers to this bug:
    http://src.chromium.org/viewvc/chrome?view=rev&revision=62820

------------------------------------------------------------------------
r62820 | stuartmorgan@chromium.org | Fri Oct 15 16:31:55 PDT 2010

Changed paths:
 M http://src.chromium.org/viewvc/chrome/trunk/src/chrome/browser/renderer_host/render_widget_host.cc?r1=62820&r2=62819&pathrev=62820
 M http://src.chromium.org/viewvc/chrome/trunk/src/chrome/browser/renderer_host/render_widget_host.h?r1=62820&r2=62819&pathrev=62820
 M http://src.chromium.org/viewvc/chrome/trunk/src/chrome/browser/renderer_host/test/test_render_view_host.cc?r1=62820&r2=62819&pathrev=62820
 M http://src.chromium.org/viewvc/chrome/trunk/src/chrome/common/plugin_messages_internal.h?r1=62820&r2=62819&pathrev=62820
 M http://src.chromium.org/viewvc/chrome/trunk/src/chrome/browser/renderer_host/test/test_render_view_host.h?r1=62820&r2=62819&pathrev=62820
 M http://src.chromium.org/viewvc/chrome/trunk/src/third_party/mozilla/README.chromium?r1=62820&r2=62819&pathrev=62820
 M http://src.chromium.org/viewvc/chrome/trunk/src/chrome/renderer/render_view.cc?r1=62820&r2=62819&pathrev=62820
 M http://src.chromium.org/viewvc/chrome/trunk/src/chrome/renderer/render_view.h?r1=62820&r2=62819&pathrev=62820
 M http://src.chromium.org/viewvc/chrome/trunk/src/chrome/plugin/webplugin_delegate_stub.cc?r1=62820&r2=62819&pathrev=62820
 M http://src.chromium.org/viewvc/chrome/trunk/src/chrome/browser/renderer_host/render_widget_host_view_mac.h?r1=62820&r2=62819&pathrev=62820
 M http://src.chromium.org/viewvc/chrome/trunk/src/chrome/plugin/webplugin_delegate_stub.h?r1=62820&r2=62819&pathrev=62820
 A http://src.chromium.org/viewvc/chrome/trunk/src/third_party/mozilla/ComplexTextInputPanel.mm?r1=62820&r2=62819&pathrev=62820
 M http://src.chromium.org/viewvc/chrome/trunk/src/webkit/glue/plugins/webplugin.h?r1=62820&r2=62819&pathrev=62820
 M http://src.chromium.org/viewvc/chrome/trunk/src/chrome/plugin/webplugin_proxy.cc?r1=62820&r2=62819&pathrev=62820
 M http://src.chromium.org/viewvc/chrome/trunk/src/chrome/common/render_messages_internal.h?r1=62820&r2=62819&pathrev=62820
 M http://src.chromium.org/viewvc/chrome/trunk/src/chrome/plugin/webplugin_proxy.h?r1=62820&r2=62819&pathrev=62820
 M http://src.chromium.org/viewvc/chrome/trunk/src/chrome/renderer/webplugin_delegate_proxy.cc?r1=62820&r2=62819&pathrev=62820
 M http://src.chromium.org/viewvc/chrome/trunk/src/webkit/glue/plugins/webplugin_delegate_impl.h?r1=62820&r2=62819&pathrev=62820
 M http://src.chromium.org/viewvc/chrome/trunk/src/chrome/renderer/webplugin_delegate_proxy.h?r1=62820&r2=62819&pathrev=62820
 M http://src.chromium.org/viewvc/chrome/trunk/src/webkit/glue/plugins/webplugin_delegate_impl_mac.mm?r1=62820&r2=62819&pathrev=62820
 M http://src.chromium.org/viewvc/chrome/trunk/src/chrome/browser/renderer_host/render_widget_host_view.h?r1=62820&r2=62819&pathrev=62820
 A http://src.chromium.org/viewvc/chrome/trunk/src/third_party/mozilla/ComplexTextInputPanel.h?r1=62820&r2=62819&pathrev=62820
 M http://src.chromium.org/viewvc/chrome/trunk/src/chrome/chrome_browser.gypi?r1=62820&r2=62819&pathrev=62820
 M http://src.chromium.org/viewvc/chrome/trunk/src/chrome/browser/renderer_host/render_widget_host_view_mac.mm?r1=62820&r2=62819&pathrev=62820

Implement IME for Mac plugins using the Cocoa event model on 10.6

Based on the implementation of Gecko's Cocoa event model IME handling on 64-bit, but modified to fit Chrome's cross-process event handling.

Because NSTextInputContext is 10.6+, 10.5 will need a cempletely different implementation.

BUG=30905
TEST=Plugins implementing the updated NPAPI Cocoa IME spec should handle IME on 10.6.

Review URL: http://codereview.chromium.org/3808001
------------------------------------------------------------------------

The following revision refers to this bug:
    http://src.chromium.org/viewvc/chrome?view=rev&revision=62864

------------------------------------------------------------------------
r62864 | stuartmorgan@chromium.org | Sat Oct 16 10:59:25 PDT 2010

Changed paths:
 M http://src.chromium.org/viewvc/chrome/branches/552/src/chrome/browser/renderer_host/test/test_render_view_host.cc?r1=62864&r2=62863&pathrev=62864
 M http://src.chromium.org/viewvc/chrome/branches/552/src/chrome/common/plugin_messages_internal.h?r1=62864&r2=62863&pathrev=62864
 M http://src.chromium.org/viewvc/chrome/branches/552/src/chrome/browser/renderer_host/test/test_render_view_host.h?r1=62864&r2=62863&pathrev=62864
 M http://src.chromium.org/viewvc/chrome/branches/552/src/third_party/mozilla/README.chromium?r1=62864&r2=62863&pathrev=62864
 M http://src.chromium.org/viewvc/chrome/branches/552/src/chrome/renderer/render_view.cc?r1=62864&r2=62863&pathrev=62864
 M http://src.chromium.org/viewvc/chrome/branches/552/src/chrome/renderer/render_view.h?r1=62864&r2=62863&pathrev=62864
 M http://src.chromium.org/viewvc/chrome/branches/552/src/chrome/plugin/webplugin_delegate_stub.cc?r1=62864&r2=62863&pathrev=62864
 M http://src.chromium.org/viewvc/chrome/branches/552/src/chrome/browser/renderer_host/render_widget_host_view_mac.h?r1=62864&r2=62863&pathrev=62864
 M http://src.chromium.org/viewvc/chrome/branches/552/src/chrome/plugin/webplugin_delegate_stub.h?r1=62864&r2=62863&pathrev=62864
 A http://src.chromium.org/viewvc/chrome/branches/552/src/third_party/mozilla/ComplexTextInputPanel.mm?r1=62864&r2=62863&pathrev=62864 (from /trunk/src/third_party/mozilla/ComplexTextInputPanel.mm revision 62820)
 M http://src.chromium.org/viewvc/chrome/branches/552/src/webkit/glue/plugins/webplugin.h?r1=62864&r2=62863&pathrev=62864
 M http://src.chromium.org/viewvc/chrome/branches/552/src/chrome/plugin/webplugin_proxy.cc?r1=62864&r2=62863&pathrev=62864
 M http://src.chromium.org/viewvc/chrome/branches/552/src/chrome/common/render_messages_internal.h?r1=62864&r2=62863&pathrev=62864
 M http://src.chromium.org/viewvc/chrome/branches/552/src/chrome/plugin/webplugin_proxy.h?r1=62864&r2=62863&pathrev=62864
 M http://src.chromium.org/viewvc/chrome/branches/552/src/chrome/renderer/webplugin_delegate_proxy.cc?r1=62864&r2=62863&pathrev=62864
 M http://src.chromium.org/viewvc/chrome/branches/552/src/webkit/glue/plugins/webplugin_delegate_impl.h?r1=62864&r2=62863&pathrev=62864
 M http://src.chromium.org/viewvc/chrome/branches/552/src/chrome/renderer/webplugin_delegate_proxy.h?r1=62864&r2=62863&pathrev=62864
 M http://src.chromium.org/viewvc/chrome/branches/552/src/webkit/glue/plugins/webplugin_delegate_impl_mac.mm?r1=62864&r2=62863&pathrev=62864
 M http://src.chromium.org/viewvc/chrome/branches/552/src/chrome/browser/renderer_host/render_widget_host_view.h?r1=62864&r2=62863&pathrev=62864
 A http://src.chromium.org/viewvc/chrome/branches/552/src/third_party/mozilla/ComplexTextInputPanel.h?r1=62864&r2=62863&pathrev=62864 (from /trunk/src/third_party/mozilla/ComplexTextInputPanel.h revision 62820)
 M http://src.chromium.org/viewvc/chrome/branches/552/src/chrome/chrome_browser.gypi?r1=62864&r2=62863&pathrev=62864
 M http://src.chromium.org/viewvc/chrome/branches/552/src/chrome/browser/renderer_host/render_widget_host_view_mac.mm?r1=62864&r2=62863&pathrev=62864
 M http://src.chromium.org/viewvc/chrome/branches/552/src/chrome/browser/renderer_host/render_widget_host.cc?r1=62864&r2=62863&pathrev=62864
 M http://src.chromium.org/viewvc/chrome/branches/552/src/chrome/browser/renderer_host/render_widget_host.h?r1=62864&r2=62863&pathrev=62864

Merge 62820 - Implement IME for Mac plugins using the Cocoa event model on 10.6

Based on the implementation of Gecko's Cocoa event model IME handling on 64-bit, but modified to fit Chrome's cross-process event handling.

Because NSTextInputContext is 10.6+, 10.5 will need a cempletely different implementation.

BUG=30905
TEST=Plugins implementing the updated NPAPI Cocoa IME spec should handle IME on 10.6.

Review URL: http://codereview.chromium.org/3808001

TBR=stuartmorgan@chromium.org
Review URL: http://codereview.chromium.org/3782012
------------------------------------------------------------------------

10.6 support merged to M8; pushing the remaining work (10.5) out.

Note that no plugin that I know of currently implements the updated Cocoa IME spec, so plugins still won't support IME until they are updated. Since both Chrome and Firefox are shortly going to support it, hopefully that will start to happen soon.

The following revision refers to this bug:
    http://src.chromium.org/viewvc/chrome?view=rev&revision=62931

------------------------------------------------------------------------
r62931 | stuartmorgan@chromium.org | Mon Oct 18 09:21:48 PDT 2010

Changed paths:
 M http://src.chromium.org/viewvc/chrome/branches/552/src/chrome/browser/renderer_host/render_widget_host_view_mac.mm?r1=62931&r2=62930&pathrev=62931

Merge 62930 - Fix one-time version check for Mac plugin IME

One-time checks work better when you remember that you've done them.

BUG=30905
TEST=None

Review URL: http://codereview.chromium.org/3779011

TBR=stuartmorgan@chromium.org
Review URL: http://codereview.chromium.org/3770016
------------------------------------------------------------------------

 Issue 62284  has been merged into this issue.

P2 bugs with an owner that are not marked as started are being automatically moved to mstone:11.

(No comment was entered for this change.)

Gecko did some swizzling to support 10.5 without using TSM directly, in order to implement the updated spec (which requires information TSM doesn't provide); we have have to go that route as well.

Pushing 10.5 support to M12. Since no plugin currently implements the updated IME spec, there's not much time pressure on getting the 10.5 part of this implemented.

I'm not actively working on this; if someone wants to pick it up, I can provide pointers for comment 33.

Stuart, I was about to dupe  bug 72810  to this one before realizing that it's only for 10.5
However, as I wrote in  bug 53643  comment 3 and it's reported in  bug 72810 , IME does not work on Mac 10.6, either. 

Is it due to a recent change to sandbox flash plugin? 

(No comment was entered for this change.)

> However, as [...] reported in  bug 72810 , IME does not work on Mac 10.6, either.

No, as reported in  bug 72810 , IME doesn't work *in Flash* on Mac 10.6. Comment 2 in that bug clearly explains why, and continues to be true. If you want to discuss that in more detail, please use that bug.

This bug is tracking Chome-side support for the Cocoa IME spec on 10.5, and has nothing to do with Flash specifically.

(No comment was entered for this change.)

Sorry that I missed your comment 28, which should have made everything clear if I had read it. 

It looks like Adobe has pushed a change to 10.3 which is now enabled for Chrome Canary on Mac that will enable IME on flash based pages. However, the location of the input options are not immediately below the text (see screenshot, the IME input options show up on the bottom left of the page).

Additionally, if you navigate away from a flash based page and then go back to the original page, typing is disabled. You can try these steps:
1. go to weathernews.jp
2. type a city
3. open a new tab (command+t)
4. command+1 back to the original tab
5. try to type in or delete the city, IME will not register.

	weathernewsime.png 337 KB   View   Download

Please file two new bugs; this bug was (and is still on 10.5) about implementing support for the Cocoa IME spec, not a catch-all for all possible bugs in the implementation.

Moving out of M12.

Moving out of M12.

This is working on weathernews.jp now. Should we mark this fixed?

It works on 10.5 even though we haven't implemented any of the Cocoa NPAPI IME support on 10.5? That seems pretty unlikely.

I may have misinterpreted the bug--I'm on 10.6.4 and IME just started working today for flash based pages with the push of flash 10.3. We can keep this open for Mac 10.5

That's the only reason this bug has been open for quite some time. See comments 28, 34, 36, 38, and 42 (and the bug summary).

(No comment was entered for this change.)

(No comment was entered for this change.)

Moving !type=meta|regression and !releaseblocker to next mstone

(No comment was entered for this change.)

Punting out non-critical bugs.  Please move back to 14 if you believe this was done in error.

moving non-essential bugs from 15 to 16. Please feel free to move back if this is an error and your bug is a blocker for 15.

(No comment was entered for this change.)

Removing milestone on all bugs marked available and targeted at m17.  Please re-triage to a new milestone if desired.