[OE-core] xinput-calibrator startup question

[Jussi Kukkonen]

On 19 April 2016 at 21:24, Koen Kooi <koen at dominion.thruhere.net> wrote:
>
>> Op 19 apr. 2016, om 17:12 heeft Jussi Kukkonen <jussi.kukkonen at intel.com> het volgende geschreven:
>>
>> Hi Koen,
>>
>> I was looking at xinput-calibrator in preparation for bug 9365 "Remove
>> xtscal in preference of xinput-calibrator", and I'm wondering if you
>> remember the logic behind this commit:
>>
>> commit 6464bcd67d10ab9967ac83c27c413c1014be707e
>> Author: Koen Kooi <koen at dominion.thruhere.net>
>> Date:   Wed Apr 30 11:33:23 2014 +0200
>>
>>    xinput-calibrator: fix XDG launch
>>
>>    In the move from meta-oe to OE-core XDG based launched was dropped
>>    without noting it in the commit message, so fix that regression.
>>
>>    Gnome-session will now launch the calibrator again.
>>
>> The commit installs the app desktop file in the XDG autostart
>> directory. But we already have a XSession script for a similar purpose
>> -- the result in Sato is that xinput-calibrator runs twice (and I
>> would expect that to happen in gnome as well)?
>>
>> I'd like to either not install the desktop file or install it in
>> applications directory, so please let me know if the reasons for the
>> above commit might still be valid.
>
> They are still valid, XDG is still the right place to launch it.

I was hoping for a little more detail... I wonder why is xdg autostart
better than xsession in this case (this does not seem obvious to me)
and if it is, shouldn't the xsession script have been removed when the
autostart desktop file was added? It seems clear that the current
setup is broken since both autostart methods are used.

Jussi