Stuff Michael Meeks is doing

Up early, poked at obs a bit with WuFung. Tried to organise thoughts, tasks & conclusions. Isolated and filed a silly crasher in nautilus-sendto's pidgin plugin.
Call with Egbert. Lunch with Bruce & Anne - good stuff. Call with Ian, chat with Pete, team meeting with JP. More osc hackery.
After poking around a little, discovered my gmarkup parsing speedup didn't actually get committed to glib2 - though seemingly a ChangeLog entry did [ amazing ], re-opened the bug.
Dinner in the evening with Andy & Simon; Alpha video, bed earlyish.

2009-04-29: Wednesday.

Up, quick breakfast, train to London; one set of meetings with a charming lady; lunch, another set of meetings with another set of great guys. Back in reasonable time - tried to catch up with E-mail on the train - wow, commuting to London sucks.
Back late, relaxed, drank wine with Kate & the Wifelet (still scarred from her tooth extraction).

2009-04-28: Tuesday.

Poked at osc build bits; chat with Fridrich wrt. odd T60p BIOS issues; it seems we all suffer them. Lunch.
Plugged away at various things, profiled my X server startup. Call with Guy, then Greg.

2009-04-27: Monday.

Up early, chat with Alex, poked Timo & Christian, phone demo with Jeff & Markus; fun new H/W arrived. Poked at installing stuff on it, prodded Kiwi gingerly; lunch.
More misc. image building, calls with JP, Mark, Paul & Jeremy. Sometimes I wonder if being US president comes with built-in Bush-isms Barack seems (50 seconds in) resolved to halt the rise of privacy in that region.
Un-anticipated prayer meeting; poor J. still scarred by the process of her first tooth extraction. Caught up with some hackery. Clarity.

2009-04-26: Sunday.

Up early, dealt with babes while J. slept, cleaned up the house. Set off for NCC, ran the creche while John spoke. Back for chicken lunch; David off - and back to NCC for a prayer tea - fun.
Move movement of things around the house, cramming stuff into nooks and crannies variously. Glued up a blanket box that appeared to be falling apart; moved some computers downstairs. Gordon sermon in the evening on Genesis.

2009-04-25: Saturday.

Up lateish, looked after everyone, and tidied the house while J. went clothes shopping at the race-course. J. arrived back minus car - the Peetle (M.) appears to love exiting the car via clambering into the front seats [ and switching all the battery draining features on ].
David arrived, and helped jump start the car; great. Back for lunch, then a lot of furniture movement with David' stirling help. Unscrewed some of the stairs, so we could get my desk downstairs - amazingly it went up there quite easily (from what I remember) - why does it not go down again ? re-assembled stuff downstairs.
Out for a spontaneous barbeque with Chris & Claire, which was fun - admired their chickens, new bench etc. Back, packed the babes to bed - up late enjoying David's company.

2009-04-24: Friday.

Up early, took H, to school; to work; poked at initrd again. Dug at OO.o images.zip performance and ordering to get some details for Tora to carry the torch there.
Call with Jeff, then James. More patching of interesting /lib/mkinitrd scriptlets.
Interested by John's call for a foundation for OO.o. Apart from the insinuation that criticising Sun's stewardship is a subtext for wanting to substitute an identically bad regimen with a different steward (which is not the case here, and would be truly dumb), it's encouraging.
For some reason people like to ask me what Oracle's acquisition of Sun means for OO.o - luckily I don't know, so I always have at least that to tell them. What I do know is that if Sun had acted, as advised, in a truly open fashion, and built a project that other companies, large and small could invest in and work with fairly: things would be better. In that (hypothetical) world, it would have produced more service revenue from the better product it fathered, but also would have made the project far more featureful and successful. As a side effect, that would have made individual developers with OO.o skills, and the project itself very much more secure against uncertainty at any single company. The key question for me is - will Oracle have the vision to create a real Free software community around OO.o, that can turn it into something as diverse, dependable and fun as it should be - there are a lot of highly clueful people there, so of course I'm optimistic, lets see what happens. As a great example of extreme clueful bravery, in the face of lots of fashion accessorising, I (personally) love Larry's views on the cloud - It has already achieved dominance in the industry ... complete gibberish ... there are still mainframes ....

2009-04-23: Thursday.

Up early, chat with Kendy, long sync with Guy. Plugged away at the mail backlog, and profiling startup - it seems our initrd does some incredibly slow things, wow - who would have thought you could boot a whole kernel in the time it waits for blogd to quit. Spent time tweaking it left & right.
Worked late, while J. went out to baby-sit. Chat with Rich, Alan and Greg.

2009-04-22: Wednesday.

Up very late. Call with Pete & Srini. Chat with Zonker & Patrick, chat with Miguel, caught up with E-mail, more crazy hacking on boot time - started a dung-out of our Xresources yielding some surreal internal flameage.
Today E. learned to say 'DaDaDaDadee' or something to that effect, and was instantly deluged with acclamation and hugs. Structural engineer also apologised and provided new drawings complete with a new Beam 8 as suggested. The roof will shortly be full of wire it seems.

2009-04-21: Tuesday.

Up early, poked at some boot time issue - fsck being run on every boot. Turns out if your kernel mounts root with different data ordering flags to that in your fstab; then when you unmount the partition you get: EXT3-fs: cannot change data mode on remount which seems to be code for - bad luck we didn't really unmount it; workaround add rootflags=data=writeback to the kernel command line.
Discovered we spend 3 extra seconds at boot, before hitting the initrd if we have a resume=/dev/sda1 type kernel command-line; magic - hopefully Greg can fix it. Worked insanely late.

2009-04-20: Monday.

Seven years of marriage today; good stuff - like a fine wine, it just gets better.
Google trends has some suitably useless graph on revision control: git vs. mercurial - of course, if people didn't call their new project by some clever household name eg. 'foot' - the "Free Object Oriented Toolkit" we might be able to better track trends - oh, and actually find what we're looking for. It's obvious that naming your next child 'gzdrqvg' will not only guarentee them the account name of their choice ~everywhere, but also make them instantly google-able.
Filed missing Clarity time-sheets, poked at udev packaging - with some crazy idea. Dug at aaa_base - why not. Chat / sync call with JP. Poked at OO.o icon themeing and performance, call with Peter.
Out in the evening for a very lovely anniversary dinner with the lovely wifelt in Ely; feeling rather under the weather, so drove. Detour on the way back. Parents left very late to get a clear run home.

2009-04-19: Sunday.

Off to NCC in the morning, back for lunch. Moved our log pile with Dad's help (and an army of small, trotting girls) from the side-way that is soon to be a lean-too (DV).
Took a break to examine the loft - strange, the wall the structural engineer is balancing 1/4 of the roof on doesn't appear to extend to the attic. Drilled a hole into it, appears to be made of cinder-block fluff - hmm. Looks like another wonderful opportunity for a large metalic beam.
Finally got the phones re-routed re-routed such that they are less likely to be 'Buildered' to death; good. Up late with the parents, had a fire to burn legacy receipts.

2009-04-18: Saturday.

Up early, cleaned house, helped set the table. Mum & Dad out all day; Sandy dropped in, then the Bancrofts, arrived: Tania, Charlie, and lots of more fun-sized people.
Large passover-style meal, slugging in the Sun, fine company, good food, lots of fun. Mum & Dad back later, slugged with them.

2009-04-17: Friday.

Naomi's first violin lesson with Anne; good stuff, learned the names of the strings fairly quickly with animal mnemonics.
Took up the landing carpet - apparently it is glued down with some re-usable tacky glue - which is great. Lifted the floor, discovered a nest of phone cable, much of it un-used, considered re-routing. Lunch.
Out to Bury St Edmonds, went to the music shop, poked around - looks great. Wandered through the Cathedral gardens. Back.
More telephonic endeavour, managed to get the floor down. Bert over for dinner, good to see him, up late.

2009-04-16: Thursday.

More clearing up, recovering from yesterday; threw out a load of ISA cards, ancient 486 motherboards, and my first 10Mbit Ethernet hub (a gift from Simon Tennant).
Tried to reduce the weight of my filing cabinet by throwing away tons of obsolete receipts and guarentees. Mum & Dad arrived in the evening, up late talking to them.

2009-04-15: Wednesday.

Up early, off to Wicksteed Park. Fantastic day out, lots of little people running from ride to ride, often the same one repeatedly. The roller-coaster a big hit with M. Hannah terrified, but fascinated by a giant swinging Pirate ship, N. and others enraptured by a car track. Lots of fun, back late, quick dinner, bed.

2009-04-14: Tuesday.

Up very late; slugged, more packing and moving of things, amazing how much time it takes to move fairly little stuff.

2009-04-13: Monday.

Lie in for J. knocked up new shelf for the kitchen, routing MDF produces some ugly swarf. R & I. up, had breakfast and left.
Did a bit of work, caught up with critical E-mail. Out for a swim with everyone - E. surprisingly happy in the pool, but rather easy to accidentally dunk at the 'normal' holding height. Fun stuff, back, tested new kernel, committed, worked late.

2009-04-12: Sunday.

Off to NCC, nice to see everyone again. Back for a quick chicken dinner. Off to Thetford Forest with the children and R&I. Suitable amounts of rushing around, climbing everything. Hot cross buns & tea; E.'s first chocolate - it is Easter after all - I suspect further confirming her suspicion that everyone else is eating something nicer than she is.
Up late by the fire, talking & enjoying Flanders & Swann.

2009-04-11: Saturday.

Up early, dealt with the babes while J. slept. Off to the market to buy fruit - got stuck into the Lyons, free Easter egg hunt with the kiddies.
Back; J. awoke, spent time cleaning the house, tidying bedrooms, making beds, getting things ready for Robert & Ilona. They arrived later, had our own easter-egg hunt in the garden. Stayed up talking to Robert & Ilona by the fire.

2009-04-10: Friday.

Bank Holiday; lie in - spent lots of time packing things away under the floor from the loft, moving various things out in preparation for impending building work.

2009-04-09: Thursday.

Chased down a twisty set of bugs causing icecream not to work properly inside the build service, or at least - when I'm using 'osc build' locally. Finally got my kernel builds to run in parallel, by not running the scheduler on the same machine as the main compile - good.
Mail, and DRI playing - interesting stuff poking at Mesa, some fun stuff to allow all manner of creative display corruption.

2009-04-08: Wednesday.

Up, wrote LXF column, poked Vuntz. Read mail, interested by the dconf thread, I look forward to seeing how that pans out. David Neary has an insightful write-up of the problems of 'joint' copyright assignment; companies don’t believe. They don’t have morals. - some of course try harder than others, but at root their morals are primarily aligned with their fiduciary duty to generate as much value as possible for their shareholders.
Dug out another trivial to fix performance silly caused in part by a bad impl. over-used by the UNO threading environment code. Call about a11y with Mike, Mark, Brady and Jared. Wrestled with the weird kernel build stuff - oh for fewer .spec file lines and more shared logic across distros.
Drove home, lifted babes into their beds; sweet little baby creatures; prodded snail mail, and the answer-phone. Bed.

2009-04-07: Tuesday.

Cooked breakfast, and to work ... found another OO.o silly taking, ~6% of import time, some heavy lifting manual RTTI check for a leaf type (SfxVoidItem) that could be a single comparison; and some multiple of 17 million calls to it, 1 line of diff.
Noticed that ooo-build has hit 1000 patches, and this blew the mind of apply.pl such that it wouldn't incrementally re-patch properly; tweaked that a little to suit. Poked at / filed a gdb failure with simple, non-threaded apps.
Most amused by Thorsten's spoof OO.o DSCM decision. I guess, it's a bit like the Irish referendum results on the Lisbon treaty: continue polling using the method de jour until your preferred outcome comes top; then stop.
Various meetings, call with Jared, got my kernel to compile finally.

2009-04-06: Monday.

Breakfast with the in-laws; poked mail, fixed up a kernel build in osc - it appears that using osc versioned links is the only sane thing to do (should it be the default ?) - osc link -r 1234 or add <link project="openSUSE.org:Kernel:Vanilla" package="kernel-source" rev="1234">
Quick call with Thorsten. Got my kernel to build; poked at HTML import some more, hacked in some more largeish wins; 24bn cycles down to 5bn including startup to load a smallish / cut-down test document, of course for larger docs - 14 seconds vs. 170secs is worth having.
Chat with JP, Jared. Poked at the latest obs builds; Clarity.

2009-04-05: Sunday.

Up early, dealt with babes while J. slept, packed to travel; off to NCC, and then Aldeburgh. Good to see Bruce & Anne. Read the paper, played with the babes, tried to encourage E. to crawl on the carpet ( softer for falling forward and banging your face on perhaps). Fine dinner, bed late.

2009-04-04: Saturday.

Lie in, breakfast; started emptying the loft of things, took futon and some obsolete testing machines down to SanMarie & Charl's. Big roast lunch.
Dug rotting apples out from under the floor, babes played crawling around under the floor and in the attic, moving boxes down the house with some success. Out for a cycling expedition in the afternoon to the race course. N. doing very well with her pedaling and steering now, real progress.
Got the babes fed, and put to bed; watched Gosford Park - rather a hair raising, and tragic film. Bed.

2009-04-03: Friday.

Up early, poked mail, dug at calc, made HTML import 60% faster or so unfortunately, that's still 55seconds for 64k cells. Lots of fairly un-amusing N^2 type feasts in there. With a small 10k cell test case, 250k linear searches through all cell contents - I guess doing the same search 25 times or so. 8bn cycles plus. Sped up the EditEngine for this corner case (with some trepidation).
Lunch; builders arrived: Clint and Richard to work out which bits to knock down first, somewhat inconclusive. Poked at mail, chewed tasks, looked at optimisation pieces. Quick call with Naresh.

2009-04-02: Thursday.

Off to the dentists with the wifelet and three babes; they're a lovely lot there. Back to the E-mail. Boggled at the calc HTML table import code, mind blowingly limited and slow, re-uses the EditEngine (at least some code re-use), but can't import more than 64k cells, and takes me 170secs to import that many; more thought & analysis required.

2009-04-01: Wednesday.

Poked at mail. Pleased to see Federico's ongoing nailing of exotic multi-head bugs in the internal tracker; 70+ down, 30 to go - that's a deep hole to fill in. Dug at Clarity, expense filing etc.
Really pleased to see, that for the first time ever we have the latest GNOME for both Factory and the latest openSUSE: 11.1 - so it's dead easy to upgrade and play: thanks to Vuntz, Luis Medinas and Magnus. Installation is rumoured to be rather easy:
```
zypper ar http://widehat.opensuse.org/repositories/GNOME:/STABLE:/2.26/openSUSE_11.1/ gnome226
zypper ar http://widehat.opensuse.org/repositories/GNOME:/Backports:/2.26/openSUSE_11.1/ gnomeback
zypper ar http://download.opensuse.org/repositories/X11:/XOrg/openSUSE_11.1/ latestX
zypper dup
```
This of course gives you the very latest X release too, as a side-effect - great to help out with improving things there too. Of course, you need to re-start X (or even re-boot to get the new dbus) and you're a happy Gnome 2.26 user.
Poked at C++ temporary lifetime semantics briefly. Switched to the sexy new ooo-build freedesktop.org git repository using Kendy's nice instructions. Added some strong deprecation warnings, with forward pointers to the old svn repo, configure now errors out there.
Committed a configmgr win - some 1+% of the startup time, spent doing heavy string sorting in a red black etree, where no strong order is required at all; 10 trivial lines of fix. Re-enabled, and updated the up-stream bug for a similar previous silly. Now - why is 3% of startup initialising language guessing fingerprints ?
Poked at go-oo DNS records, OO.o team meeting. Dug at lingucomponent and idle initialized the language guessing code saving 2.5% of the startup CPU pseudo-cycles, and a load of pointless I/O (and some memory no doubt) in a few lines.