Bug#524185: ditto
On Thu, Sep 24, 2009 at 10:00:33PM +0200, Julien Cristau wrote:
> On Thu, Sep 24, 2009 at 20:47:35 +0200, Josip Rodin wrote:
> > In any case this is a really bizarre behavior. Surely the program should
> > first successfully detect and enable at least one keyboard and/or mouse
> > device via HAL, and *only if that went well* decide to ignore completely
> > valid xorg.conf settings?
>
> xorg.conf parsing happens before we connect to hal, so no. Also hal can
> be started after the server, and devices can be plugged in later still,
> so there's no way to know that.
Why would someone program the defaults for use cases which can't be
widespread (because they weren't supported in previous releases)?
Do you really think the users care about the technical details of when some
parsing happens in the current implementation? Besides, any previously
obtained state can be cached and interpreted later, at the point when the
HAL connection fails.
*Anything* is better than a system whose entire array of configured input
methods have been intentionally killed without any obvious recourse for the
normal user.
This is just plain wrong, and it saddens me that I'm even having to have
this discussion...
--
2. That which causes joy or happiness.
Reply to: