After an evaluation, GNOME has moved from Bugzilla to GitLab. Learn more about GitLab.
No new issues can be reported in GNOME Bugzilla anymore.
To report an issue in a GNOME project, go to GNOME GitLab.
Do not go to GNOME Gitlab for: Bluefish, Doxygen, GnuCash, GStreamer, java-gnome, LDTP, NetworkManager, Tomboy.
Bug 678125 - at-spi2-atk still requires NO_AT_BRIDGE
at-spi2-atk still requires NO_AT_BRIDGE
Status: RESOLVED FIXED
Product: at-spi
Classification: Platform
Component: at-spi2-atk
unspecified
Other Linux
: Normal normal
: ---
Assigned To: Li Yuan
At-spi maintainer(s)
Depends on:
Blocks: 677491
 
 
Reported: 2012-06-14 19:53 UTC by Alejandro Piñeiro Iglesias (IRC: infapi00)
Modified: 2012-06-15 16:07 UTC
See Also:
GNOME target: ---
GNOME version: ---


Attachments
Setting a NO_AT_BRIDGE on atk_bridge_adaptor_init (831 bytes, patch)
2012-06-14 19:55 UTC, Alejandro Piñeiro Iglesias (IRC: infapi00)
reviewed Details | Review
Getting back the NO_AT_BRIDGE (1.47 KB, patch)
2012-06-15 11:03 UTC, Alejandro Piñeiro Iglesias (IRC: infapi00)
reviewed Details | Review

Description Alejandro Piñeiro Iglesias (IRC: infapi00) 2012-06-14 19:53:14 UTC
During all the revamp moving the bridge from a module to a library, the NO_AT_BRIDGE test was removed (as was part of the module loading code). But it is required for multi-toolkit apps like GNOME Shell, initially it was readded to gtk, but that belongs to at-spi2-atk
Comment 1 Alejandro Piñeiro Iglesias (IRC: infapi00) 2012-06-14 19:55:03 UTC
Created attachment 216455 [details] [review]
Setting a NO_AT_BRIDGE on atk_bridge_adaptor_init

Before the revamp, this test was done prior the call to the method. This patch re-add the test to _init itself
Comment 2 Bastien Nocera 2012-06-15 09:44:09 UTC
Review of attachment 216455 [details] [review]:

At least that.
Better would be a way to make the bridge into an object, and set a root on it.

::: atk-adaptor/bridge.c
@@ +723,3 @@
     return 0;
 
+  load_bridge = g_getenv ("NO_AT_BRIDGE");

load_bridge = check_envvar ();
if (inited && !load_bridge)
  {
    g_warning("ATK Bridge is disabled but a11y has already been enabled.");
    return;
  }
Comment 3 Alejandro Piñeiro Iglesias (IRC: infapi00) 2012-06-15 09:57:16 UTC
(In reply to comment #2)
> Review of attachment 216455 [details] [review]:
> 
> At least that.
> Better would be a way to make the bridge into an object, and set a root on it.

Yes, this is an old discussion:

https://mail.gnome.org/archives/gnome-accessibility-list/2009-August/msg00027.html

Get a way to get rid of part or most of AtkUtil interface, and provide a way to set/get the root. Probably also a signal to notify that the root has changed.

Now that the bridge is a library, it could be easier to implement, but for the moment I think that a good first step would be make this library movement.
Comment 4 Alejandro Piñeiro Iglesias (IRC: infapi00) 2012-06-15 11:03:41 UTC
Created attachment 216504 [details] [review]
Getting back the NO_AT_BRIDGE

Updated patch after Bastien review. Just a small change. inited && !load_bridge is valid for that warning. But the test to know if the function should return is inited || !load_bridge
Comment 5 Bastien Nocera 2012-06-15 11:27:02 UTC
Review of attachment 216504 [details] [review]:

Rest looks good.

You'll need another patch to remove those tests from the GTK+ modules though.

::: atk-adaptor/bridge.c
@@ +708,3 @@
+ */
+static gboolean
+check_envvar ()

void
Comment 6 Bastien Nocera 2012-06-15 16:07:34 UTC
Fixed the review bug, and made the changes to the GTK+ modules.