Re: GTK+ modules loading changes
- From: Piñeiro <apinheiro igalia com>
- To: hadess hadess net
- Cc: gnome-accessibility-devel gnome org, desktop-devel-list gnome org
- Subject: Re: GTK+ modules loading changes
- Date: Tue, 26 Oct 2010 18:31:02 +0200 (CEST)
From: Bastien Nocera <hadess hadess net>
>> Ok, so explaining that in a rough way, this is equivalent to set the
>> env var GTK_MODULES to the modules you want.
>
> The changes I documented are unrelated to the env var.
>
>> > This will make the libsupermodule.so GTK+ module be loaded by GTK+.
>> >
>> > If you need the module to be conditionally loaded, you'd add those 2
>> > lines to your .desktop file:
>> > X-GTK-Module-Enabled-Schema=org.gnome.desktop.interface
>> > X-GTK-Module-Enabled-Key=accessibility
>>
>> And going on with the rough way, this is equivalent to fill
>> GTK_MODULES in a conditional way.
>
> This is equivalent to using the GConf stuff we had before, not the
> GTK_MODULES env var.
Sorry I used a bad example to explain this, but I get the point.
>> In order to support this, it is used the envvars NO_GAIL and
>> NO_AT_BRIDGE. In order to support this atk-bridge and gail has code in
>> his module_init code preventing this.
>>
>> As far as I see this should be still working. But to be sure... could
>> you confirm it? Any advice to the current procedure ? (the current
>> procedure is a hack). This question arised after read this "Adds the
>> disabled-gtk-modules..." [7]
>>
>> More info about the Firefox loading modules mess here [5]
>
> Those will still work as well, or as badly, as before.
>
> The only thing we changed is where the GTK modules list, which is
> exported through XSettings, is loaded from. I haven't changed anything
> in GTK+ itself.
Ok.
>> I think that at-spi2-atk bugs are still on freedesktop [8], I will ask
>> Mike Gorse about it.
>
> OK.
>
> The only big change is that enabling a11y in GSettings will now change
> the GTK+ modules list on the fly. I'd expect quite a few bits of code
> breaking if they monitor the list of GTK+ modules at run-time.
Do you have any specific "breaks" on mind, or you are talking in a
abstract way? Breaks like activating the a11y module loading without
at-spi2 being running?
BR
===
API (apinheiro igalia com)
[
Date Prev][
Date Next] [
Thread Prev][
Thread Next]
[
Thread Index]
[
Date Index]
[
Author Index]