Utter clusterfuck on the email situtation and trying to cope with a one day turn around from vacation to US trip. If you’ve mailed me something important, just remember you’re probably X of about 8000 mails right now that I’ll be working through on the 17 hour flight to SFO.
Getting totally excited about about the Solaris Desktop Summit we’re having. It’s going to rock.
In preparation for the Desktop Summit that we’re having internally inside Sun, I figured it was about time to start reading over the DTrace stuff so I don’t look like a complete idiot when I attend which would be rather embarassing.
So there I was just playing around with scripts figuring out how the probes and providers, and the predicates and aggregations work – still mostly in kernel land where WHOOOOOSH everything goes over my head.
Then I stumbled across my next example, and I had an ‘OH MY GOD, THIS IS FUCKING UNBELIEVABLE’ epiphany –
# dtrace -n 'pid$target:gedit::entry { }' -c 'gedit'
dtrace: description 'pid$target:gedit::entry ' matched 1359 probes
dtrace: buffer size lowered to 2m
CPU ID FUNCTION:NAME
0 74350 _start:entry
0 74351 __fsr:entry
0 74356 main:entry
0 74423 gedit_debug_init:entry
0 74424 gedit_debug_message:entry
0 74424 gedit_debug_message:entry
....
And I started watching it, and more information spewed out at the terminal as I interacted with the user interface. I’m sure I’m only touching the iceberg of how useful this is going to be in the future. Sure, it’s only as useful as the hands of the developer, and I still suck at writing code.
But any developer who hasn’t had a serious play with this is still in a world of black and white. Find your colour folks. Pimp your Trace!
Update: Alan takes things a step further.
Hooray, finally got Darren’s Solaris Desktop Gap Analysis document published on opensolaris.org. It’ll be nice to have something to point at now, and it’s a pretty good list of potential projects that we need to look at going forward if we’re to compete in the desktop space. All feedback to desktop-discuss [at] opensolaris [dot] org.
Moved into my new place on Monday. Telecom fucked up the transfer again, but this time on the normal phone side. Apparently my broadband connection was operational all the time. Bugger. Managed to lock myself out of the old place. Bugger. However, it all turned out well in the end, and I’m more or less moved in, except for the boxes piled around the place. This one is a lot bigger [2 bedrooms – one of which I’ll use for an office], with a nice outside balcony stretching the full length of the flat, complete with outdoor furniture, so I imagine it’ll become a nice evening work place and party venue.
Met up with Marek for TVIC, which turned into beers at the local. All good, and formulated some plans for climbing tomorrow. Hard to believe I’m off for 2 weeks holiday on Saturday to Thailand – really looking forward to getting away, relaxing, and leaving stuff behind in NZ, despite not really feeling all that prepared for 2 weeks of climbing. Kind of feel like I’ve slipped my training for the last week or so, but hopefully that’ll just be a small break before I hit the overhands in Krabi. Rock on!
And the great news from our family is that my brother, Dunc, got engaged. Wonderful stuff, and absolutely delighted for himself and Denise. I’m sure it’s going to be one hell of a wedding party!!
Finally, after 5 years of being at sun, the ‘gman’ alias is MINE, ALL MINE! Not only can you reach me at ‘theface at sun dot com’, but now ‘gman at sun dot com’. Happy days – it’s the simple things in life.
I finally got my flights booked for a trip to the US in late November for John’s excellent idea of having an internal only [sorry folks] Solaris Desktop Usability and Performance Summit. The summit will be a 10 day hackfest with various different teams around Sun [JDS, XServer, Approachability, Sun Ray, Compiler, DTrace, and Kernel]. It’s going to be fricken awesome, and great things will come of this. I know a heap of us will be blogging live from the summit.
So, with a trip to Thailand planned in a weeks time, I’ll have a 1 day turn around [insane] when I get back before I head out to the states. I, and many other JDS team members, will be in Menlo Park from Mon 28th Nov until Sat 10th Dec. Hopefully I’ll get to meet a whole heap of OpenSolaris and GNOME fans over there – drop me a mail if you’re around and I’ll come and say hi! Bonus points if it involves beer.
You know Jeff, if this Ubuntu gig does’t work out for you, there’s always a contract with the village people available.
Philip Lindsay set up Planet NZTech, an aggregation of people who are doing stuff in the tech industry which is cool to see. Hopefully it’ll be a natural progression for our ‘The Valley In Christchurch’ meets.
This story was interesting, if a little on the bitter side. While I’ll always have a fond memory of our time developing a Linux based product, despite being fraught with difficulty, especially when your base distribution vendor gets bought by another company – do the math. Doing the ‘Ximian Desktop’ thing is hard, and without any decent control over the platform you’re stuck in a world of ‘will not fix’, hacks and workarounds. It’s not pleasant from an engineering point of view, and from a business point of view, I can appreciate the decisions being made too. Tom seemed to unfortunately get caught up in that frustration.
The good news is that the code is available for most of the core components of the desktop [GNOME 2.10 currently, GNOME 2.12 to follow real soon]. With our build environment as it is, there’s a collection of RPM spec files for you to build any packages that you need to. The rest of the components will all be available in due course as announced last week.
However, we’re following all the same engineering principles that we had on JDS Linux with our OpenSolaris work, to create an easy-to-use desktop, and I personally think we owe it to ourselves to see if this gig works for a while. I for one welcome our new evil overlords – at least there’s a better chance of being able to work with them and get the right projects started from a desktop point of view. I sure as hell haven’t seen the ‘holy war’ fanatics that are mentioned – everyone have been nothing but hugely helpful towards the desktop team. Don’t judge a project based on one or two people – dealing with Linux should have taught you that by now.
