Just came back from holidays, back to reality:
If you’re waiting for an answer from me, please don’t desperate, I’ll be answering
in the next few months…
I have been running now kernel 2.6 on my laptop for a few days now, and
everything seems to work perfectly. What made me change to 2.6 was the promise
for more speed, noticeable in the daily desktop usage, and I can now confirm that
this is true. Everything goes much quicker.
This morning I woke up to find my desktop machine broken, with what seems to
be a hard disk problem. Although it’s quite weird, since e2fsck says the
partitions are ok, but even though I can boot, none of the boot files
/etc/mtab, /etc/fstab, /etc/init.d...) seem to
be read. Now trying to recover my home directory, where yesterday’s work
Nice article that outlines
some places where GNOME does a better job than Windows and Mac OS. This simply demonstrates
that we can do much better than others!
I had an interesting talk this morning over IRC with James Cape, of libgtcpsocket/libgnetwork
fame, about how to architect the libgnetwork API. The idea is to provide a very simple
API for all network services, similar to what there is in GNOME-VFS. That is, you would
be able to do stuff like:
ftp_client = gnetwork_client_new (host, GNETWORK_PROTOCOL_FTP); http_client = gnetwork_client_new (host, GNETWORK_PROTOCOL_HTTP);
And then use the same set of functions (the one in
for both the FTP and the HTTP clients.
After the talk, I started having a deep look to OpenSLP,
and even managed to install it in all the machines in my home network. I’ve got now several
services registered (SSH, HTTP, HTTPS, LDAP, etc), and can, via the OpenSLP’s command-line
tools, get the list of registered servers for each service:
$ slptool findsrvs https service:https://192.168.0.2,65535 $ slptool findsrvs http service:http://192.168.0.5,65535 service:http://192.168.0.2,65535 service:http://192.168.0.3,65535 $ slptool findsrvs ssh service:ssh://192.168.0.5,65535 service:ssh://192.168.0.2,65535 service:ssh://192.168.0.3,65535 $ slptool findsrvs ldap service:ldap://192.168.0.5,65535
Now we just need to start using the discovery mechanism in client applications, so that they
can let the user choose from a list of discovered servers instead of having to type
(and remember) host names (or, worst, IP addresses).
SLP seems to be much better suited for service discovery than Zeroconf, specially on large
networks (that’s what some network admins have told me), so it seems to be the way to go.
Some work on libgnetwork about this is going to take place soon, as soon as we decide how
to integrate all kind of service discoveries (Microsoft’s, SLP, Novell’s, etc) into a
single entry-point API. Once that is done, we’ll start sending patches for the GNOME applications
to start using the service discovery features.
Finally had some time to upload the photos from last weekend
ski in Formigal. It’s been only a year since I skied for the first time,
but I’m really loving it more and more every time I do it. Yes, it’s true, I
am still learning:
But things are really improving:
Mono on Debian
At last someone
has taken over the tasks of maintaining updated Mono packages for Debian. Kudos to the
Debian Mono Group!
Laurent (of GDA Ruby bindings fame)
has added support for converting between GDA and Ruby types, thus allowing the removal
of the GDA:Value class. This is something very nice for Ruby developers, since it
allows them to avoid having to do the conversion and work directly with the Ruby
language types. It’s so clever than it seems many Ruby developers are starting
to look forward to it instead of DBI.
More details in Laurent’s
A couple of newsitems from the European Space Agency, one
new discoveries made in Planet Osiris and about the Rosetta
space probe’s future launch in its trip to meet 67P/Comet Churyumov-Gerasimenko.
Vivien committed this weekend all his pending changes for the new Mergeant version.
Very nice to see that going on, since Mergeant development was a bit halted after
the 1.0 release of libgda/libgnomedb. There is still a lot of work to be done,
but this is another step towards Mergeant 1.0, which will solve the “missing
Access alternative problem”, of course with a better one than Access itself.