Hi,
On Mon, 29 Apr 2013 09:17:15 +0200, Oswald Buddenhagen <[email protected]>
wrote:
On Thu, Apr 25, 2013 at 03:11:25PM +0200, Martin Bříza wrote:
To provide reasons why to do it:
- The current state is nearly unmaintainable
that tiny file? you have a low threshold. ;)
Yeah, that's true :) But thinking of the future it'd be nice to have it
nice and clean.
and having all three (for now) ways of session handling in one file
doesn't feel very clean.
it was the best at the time of creation due to the amount of shared
code (the old gdm and kdm socket code was identical except for the
string literals and a few conditionals). the ck code was tacked on later.
at this point the old gdm code can be probably purged, making the kdm
code independent. you should do that in a separate first commit.
the ck and systemd paths may share some code, though. maybe add a
DBusUtils class.
That seems reasonable, I will keep it in mind.
I would leave creating the CK module to somebody who is experienced
with how
exactly the whole system works and knows if it is safe.
good luck with that. i'll do reviews, not a bit more.
note that partial works (be it regressions or just a highly asymmetrical
code structure) will not be accepted. if you don't find somebody to do
the refactoring for you, you need to take the route of minimal
modification.
It will work the same as it does now but I'll split the systemd code away
from the main (let's say "legacy") part of the code. Basically, all the
functions would be virtual but the inheriting classes would implement the
functionality on their own.
Until the CK support is split (if it is done at all), it will stay in the
"legacy" part to not break backwards compatibility.
Kind regards,
Martin
