Archive for July, 2009

Guess what…

Monday, July 20th, 2009

… I’ve new “fun” with release-management. This time “make distcheck”, which was recommended to me, is bocking my work.

The amount of work that is needed to “ship a line of code” to users is insane.

notify-osd 0.9.15 is out

Monday, July 20th, 2009

I just made a tarball-release of notify-osd 0.9.15 available here. I’m doing a PPA for karmic as you are reading this, which will be available here. Rest assured that it also works on jaunty and other non-ubuntu distributions. But be warned that the experience is likely to be degrated a bit as not many distributions share the ayatana-team’s view on the way notifications should look and behave (read: this mainly means integration-patches for applications).

The major thing in this release is the surface- and blur-cache allowing for in-bubble blur when the mouse hovers over it, achieving a kind of defocus effect to allow the user to better focus on what’s below the notification-bubble. This works on compiz and metacity+compositor. Yes, it also works on systems as low-spec as a Mini 9.

Work on this is not fully done yet, further optimizations are pending and there is a regression at the moment for over- and undershoot-effects. But I did not want to sit forever on this *g*

For the more curious reader, here’s an excerpt from the ChangeLog:

  • fixed wrong default values for bubble-width, margin- and icon-size
  • fixed regressions introduced by the move to the surface- and blur-cache
  • improved drawing of gauge/indicator
  • tweaked rendering of layout-grid, added display of layout-grid to debug-mode of notify-osd
  • refactored rendering code for bubble
  • migrated rendering of drop-shadow, background, icon, title, body and gauge to surface/blur-cache
  • added surface- and blur-cache
  • added debug-output for font-size, screen-dpi, pixels/EM and bubble-width in pixels
  • fixed rounding errors in pixel-per-em calculation, changed default bubble width to 24 EMs, fixed title- and body-text height, fixes LP: #336110 and LP: #391190

And what would a blog-post as this be without a screencast:

Testing animated bubble resize

Wednesday, July 15th, 2009

With the surface- and blur-cache available now, it is now also possible to animate the resizing (actually only “growing” will be used by notify-osd). Here’s a small unspectacular example screencast:

You can find this example in a branch of notify-osd here.

Good night…

Wednesday, July 8th, 2009

… and with this I’ll give you:

Edit: This of course works under compiz and metacity (+compositor). Under plain metacity you would not get the transparency against the desktop though. You can find this and other examples in a branch of notify-osd here.

Entering uncharted waters

Saturday, July 4th, 2009

Or to put it in another way… my 5 minutes of troll-ism.

I attended the keynotes this morning at the Gran Canaria Desktop Summit. The first one, by Robert Lefkowitz, was the most impressive/informative/entertaining of them in my opinion. “A gentleman would use liberal software.”, what a terrific punch-line to sum up his talk! I like that *g* The keynote by Walter Bender on the sugar interface was also very good. Sorry Quim, I missed your keynote.

I’m not writing this here to praise the good keynotes or to apologize. I’ve had and have an issue with Richard Stallman’s keynote. Just to make this clear, I do not have an issue with him as a person - I don’t know him, thus I’m unwilling to judge him - he appears to be funny though (if by accident or deliberately I leave open for the readers to decide by themselves). More specifically I have an objection with his view on the danger C# opposes to the (GNOME) desktop.

Question: Why is a free implementation of C# and its runtime- and execution environment and supporting libraries worse then C++ (and its standard template libraries that goes with it), which was also initially invented and developed at big “evil” company (AT&T) by Bjarne Stroustrup?

I don’t see a fundamental difference. He says that C# is to be avoided where C++ is to accepted/recommended (speaking within the scope of potential legal strings attached, not talking about language features). Why is this also different to Java? It is just a language with free implementations of its framework (compiler, runtime-environment, libraries).

Sadly I became aware of this question only after the QA session with him. I should have asked him directly on the spot. I invite everybody (also Richard Stallman, should he happen to come across this very blog post) to help me answer this question… either via comments or via eMail.

Thanks in advance for your time and (thoughtful) input!

Hit Las Palmas

Friday, July 3rd, 2009

Arrived at the hotel/summit. Going to head to the venue with a couple of folks soon.

And a small screencast for the day. Testing out the new blur-cache meant for notify-osd:


It’s easy on CPU. So light actually, that I was able to record this very screencast with recordmydesktop on a Dell Mini 9. This is finally also using the subtle text drop-shadow the design folks asked for. Color, font, size and all are just randomly picked by me, as this is a test-program to exercise the small interal APIs I created for implementing the blur-cache.

Edit: I forgot to mention it, that I’m doing this work in a branch currently. It’s lp:~macslow/notify-osd/blur-cache.

Off to DesktopSummit/GUADEC

Friday, July 3rd, 2009

It’s that time of the year again :) I’m about to start my trip to the summit. Uff… 5:00 in the morning and a trip of roughly 14 hours before me. But can’t wait to see all you GNOME-heads again face to face!