[Kde-hardware-devel] status of udisks2 in Solid

[Lukáš Tinkl]

Dne 3.12.2012 23:51, Lisa Vitolo napsal(a):
> Hello,
>
> due to personal commitments I unfortunately had few opportunities of
> keeping track with the most recent events in the KDE community, but I
> noticed there are still some uncertainties about how the switch to
> udisks2 will be managed. I know Solid has been testing its own backend
> for udisks2 from some time, so I just wished to know in few words what
> the state-of-the-art is (you can point me to appropriate links, no need
> to explain everything here in the list).
>
> I'm asking this simply because I'm the student that developed the Solid
> submodule for partitioning operations for this summer's GSoC, so of
> course the switch to udisks2 will affect my work as well. Of course I
> had already started working on it, but I just need some more
> clarifications. Regarding this, if you need any help with the udisks2
> backend, I'm available: I have used udisks for a long time now and I can
> perhaps be of some usefulness to the project :)
> By the way, I talked with Alex in the Solid IRC channel, but it's worth
> mentioning in a more official place: my code is now in a usable status
> and awaiting for a review. I apologise for the huge delay from the end
> of the GSoC project until last week or so, but as I said, other
> commitments kept me away from this for longer than they were supposed
> to. Of course the code is publicly available at my repository
> https://gitorious.org/solid-partitioner/solid-partitioner in the
> partitioner branch, and I am always around for any clarification you may
> need.
>
> Speaking about udisks2 again, I read that most projects intend to employ
> caches to reduce the initial delay caused by udisks2's somewhat slower
> detection. Since Solid has always used a cache to speed up access to
> devices properties, I believe it will continue to do so. While
> developing my module I noticed that under some circumstances some
> notifications where delayed: to be more specific, the cache was re-built
> as a consequence of receiving a deviceChanged signal *after* a
> notification warned the user of a change in the accessibility status of
> mountable partitions, so out-to-date data were read. A (not so elegant)
> trick can be seen in my code that solves this, but I may have missed
> some other cases, so well, be careful about it :)
>
> Good day to everyone,
> Lisa

Hi,
currently kdelibs 4.10 branch contains both udisks and udisks2 backends, 
configurable with a cmake switch (-DWITH_SOLID_UDISKS2:BOOL=ON). In the 
past few days, Dan Vrátil and I did quite some fixes and speedups to the 
latter so the result is now more than pleasant :) Everything is as fast 
(or even faster) than with the older udisks backend.
Don't hesitate to try it out and report bugs if need be.

-- 
Lukáš Tinkl <ltinkl at redhat.com>
Software Engineer - Base Operating Systems Brno
KDE developer <lukas at kde.org>
Red Hat Inc.                               http://cz.redhat.com