Stuff Michael Meeks is doing
|
|
This is my (in)activity log. You might like to visit my employer
Novell which is an amazing company, and also
Dell who in days of yore provided me with a
free laptop for Gnome development / conferences.
Also if you have the time to read this sort of stuff you could enlighten
yourself by going to Unraveling Wittgenstein's net or if
you are feeling objectionable perhaps here.
Older items:
2009: (
J
F
),
2008: (
J
F
M
A
M
J
J
A
S
O
N
D
),
2007: (
J
F
M
A
M
J
J
A
S
O
N
D
),
2006,
2005,
2004,
2003,
2002,
2001,
2000,
1999,
legacy html
-
Up lateish, played in the road with the children, keeping
our distance from others: Scarlet-Fever-wise. H & N. slowly
learning to ride their bikes. H. got started on the next size up,
Naomi had some success learning to steer her pedal-less bike,
interestingly learning more going up-hill and running to a stop,
than going down hill.
-
Lunch, more playing outside, J. concurrently de-grassing
the front-garden; why is it that grass grows wonderfully in the
borders leaving the lawn the preserve of moss ? Watched
Veggitales' Jonah in the afternoon, tea, bed.
-
Booked coach ticket to Romania; surprisingly civilised wrt.
late booking & cancellations, the coach company - compared to eg.
cheap flights it seems. Wrote some notes up on OO.o.
-
Call with Sean, back to sorting receipts into heaps. Lunch,
filed expenses at some length, while listening to internal Q1
financials call.
-
Poked at doing a dbus-launch from the gnome session shell
script to avoid linking gnome-session twice, and launching b-a-s
early to stop it delaying the panel later. Poked at nautilus'
plugin loading ways. More Clarity.
-
Up early, to work, poked mail, call with Ralf, call with Olaf.
Lunch. Fought the inbox, updated the gmarkup bug for Matthias. Filed
clarity bits.
-
M. off to doctors and then hospital with J. to handle possible
scarlet fever case; leaving me holding the fort; Desktop staff meeting.
Cooked something lame for the remaining babes for dinner, fed and put
them to bed.
-
Back to work; J. arrived back late. Poked at Impress
OpenGL transition color scheme issue (related to cairo somehow), and
poked an odd launcher issue. Extracted a groupwise XML trace for a bug,
worked with JP on a security issue.
-
Up lateish, poked mail, some degree of catchup there. Read
about bugs. Lunch; OO.o team meeting. Up-loaded my OO.o
slides
from Scale.
-
Call with Adrian, wrote up some notes, chat with Kohei. Plugged
away at admin until late. Niculai around in the evening for dinner.
-
A night of flying, slept a little. Iris scan worked like a
charm on my return, nice - no passport required. Plugged away, code
reading on the train. Discovered gnome-session doing some chunk of
startup linking twice - since it execvp's a dbus-launch of itself,
surely better to dbus-launch gnome-session the existing parent
script.
-
Returned to the bosom of the family etc. The peetle's
~20 month old, repeated telephonic: I want you back! ceases
to give that wrenching effect.
-
Up early, call with Guy & co. breakfast, prodded at mail.
Checked out, wrote some mail, poked at gconf some more - did some
cleanups for Mathias and noticed another trivial win - interleaving
ASCII detection into the escaping code for free - to avoid more utf-8
validation.
-
Off to see some hills, and a partially mangled 'HollyWoodLand'
sign, and a hotel or two - interesting enough I suppose; failed to notice
any particularly attractive women: particularly to match those in
Newmarket. Quite pleased with the relatively tasteful abodes of the
film class. Back to the airport.
-
Up early, breakfast again; off to Joe's talk - nice slides,
some good points. Poked at gconf some more, to try to build a separate
simple list of what is touched at login time - which should be a tiny
fraction of the data. Lots of interesting stuff in the trace, mailed
some people to encourage them to pre-load directories where they
are not.
-
Why do we (apparently) load all the thumbnailer keys on startup ?
484 round-trips for that at least: interesting. Some interesting
un-necessary sets happening on login too.
-
Enjoyed the EyeOS talk, back to the booth, talked to people
at considerable length on every topic, got a few people interested
in the build service. Chatted to a chap with some beautiful python
library for generating ODF, helped pack up the booth.
-
Out for dinner with Christian, Jeff, Nathanial & co.
Daniel Philipps came for desert which was tasty, a most pleasant
evening.
-
Up very early, tried to sleep, failed. Breakfast, started on
the mail, spammed people with a write-up, started slide creation;
deeply tempted by the prospect of removing some more fat from gconf.
Read some analysts output.
-
Poked at libmapi exported symbols, wow; it badly needs a map
file, or better - visibility work, and an agreed symbol prefix / namespace
to boot
GetTable or SetColumns is not going
to fly in a big application without that.
-
Back to slides; unbelievably irritated by DreamHost, talked to
a RackSpace sales chap in some interactive chat thing - that's a good,
sticky sales method - particularly when coming from something far lamer.
-
Remarkably successful slide-ware creation, on and on like a
machine; skipped lunch - the joys of a huge breakfast, and no exercise.
Out for dinner with Stephan, Trevor & Will.
-
Up rather early; bid 'bye to the lovely wife &
babes. Set off on the train. Hacked GMarkup a tad, made it
another chunk faster by avoiding a couple of allocations per
element and attribute, now ~35% faster, and more to come.
-
Knocked up some in-place un-escaping logic, and
callgrind reports an overall 45% win, good. Ran over the
nice, simple regression test suite, looking good, somewhat
reduced quality error reporting but - seems reasonable.
-
Watched misc. movies on the plane, the latest Cohen
Brothers' Burn After Reading seems to me a somewhat
scorching comment on the ethics of extreme positive thinking,
interesting film.
-
Finally arrived, found the hotel, managed to persuade
the network to like me, filed my
gmarkup patch. Chat with Guy, Gary, JP. Bed early (for LA).
-
Prepping crazily for absence at SCALE from tomorrow; dug at
mail. Chat with Srini & Tambet, call with Mark, wrote notes,
filed misc. bugs.
-
Lunch, OO.o team meeting, call with misc. Novell guys;
beta customer call. Packed in the evening, got the laptop setup,
built the latest OO.o.
-
Up early, holiday today - a day off in half term; off
into Cambridge with all the babes - via the Park & Ride - to
enjoy a double-decker bus with the kids. On to the Zoology Museum,
admired various skeletons, the kidlets got some sketching in.
-
On to a Restaurant near Magdalene for lunch, back to the
Whipple Museum of the history of science later - much (free) fun
had by all small girls. Misc. shopping - how I loathe buying printer
ink at a vast price in a small package; home variously. Watched
Bedknobs and Broomsticks with the little girls.
-
Managed to get second round of quotes for building work
printed and sent out.
-
Up lateish, the School holidays, time off tomorrow. Poked at
bugs, chased a beagled issue swallowing my disk, poked at KCachegrind
output some more. Lots of mail chew.
-
More chasing of beagle file leakage into compressed mail
attachments, clarity. Poked at cmake quickly. Discovered the Kumppa
screensaver is a rival to the washing machine for baby entertainment,
if only it made a <whirr-clunk> sort of noise too.
-
Bathed babes, put them to bed, pizza dinner - back to poking at
the GMarkup performance problems: there are a load of places that do
odd things like:
if (*p == '\n') p = g_utf_next_char (p);
that are just daft - one of the joys of utf-8 is that if you're looking
for an ASCII '<' then there is no way some clown can hide that in a more
complex multi-byte sequence; ergo we can parse the XML, and then
validate it rather than pre-validating. Implemented that - throwing a
ton of logic away, and saved 8%, good; now to reduce the heap thrash.
-
Lie in, off to NCC, Tony spoke met a new couple from Nigeria
Solomon & Peace - back for lunch; some form of slugging, out to
Hannah & Nick's in the afternoon, had a lovely meal there too;
wow. Back, babes to bed, relaxed.
-
Up early, rapid Valentines day card creation from emergency
spares laid down in a time of plenty. J. had a lie-in, fed and packed
all babes into the pram, set off for the market. Flowers, heart-shaped
cake purchasing action - along with tons of vegetables.
-
Back, lunch. Disassembled the R40 and replaced the old fan
assembly with a new one - laptop now working, played with misc
open-source games for a while. H. out to a party, played with
the other girls, and Benjy for a bit - bed early.
-
Up early; disappointed to see the Government's inept handling of Wilders.
Similarly amazed by the Muslim council's attitude; at least from what I've
seen Fitna the movie is apparently an attack on the tiny minority
of Muslims that apply a selective and barbaric hermeneutic to the Qu'ran and
act on that: empirically such people exist, and use the texts he highlights.
Given that context, surely a moderate Islamic response is to use the opportunity
here, to not only condemn Geert Wilders (for his lack of balance), but to
strenuously separate the peace-loving,
vast-majority of Muslims from the pernicious and frankly cowardly minority -
that apparently lauds the beheading of defenceless captives in cold blood (as
an example). It is mildly distressing to see Geert being roundly condemned as
a fascist, but an apparent silent condoning of the fanatical Ideologues, the
likes of al-Zarqawi, to whom the term might more accurately apply. Demanding
censorship of such a film, while simultaneously not robustly mirroring his
message of condemnation of such views is really concerning. People will
see it, and they desperately need to know know that the vast majority of
Muslims do not share this interpretation, and are no risk to them; who
will tell them ? Who will persuade all Muslims to avoid anti-semitism
and intolerance if their spokesmen will not ?
-
Dug at mail. Tried to fix a yast2 bug, and had all manner of grief
with both sudo and su - eventually realised it was down to stracing my
whole session; good to fix that. Fixed a nasty entity handling issue in
the yast2-gtk HTML -> XHTML conversion, HTML is truly nasty.
-
Spent some time analysing our packages, patch counts and so on,
requested FTO for Tuesday, worked on documents, not code.
-
Dug at mail, wrote some unit tests, poked at gconf some more.
Discovered a chunk of time spend building 11k error messages during
XML parsing, that we subsequently threw away: gettext, g_strdup_printf
etc. committed the obvious patch. Filed the inherited attributes thing,
seemingly most of the remaining win is to remove the un-necessary l10n
strings in the main defaults/schemas file.
-
Staff meeting with Ralf, more mail thrash, poked at nautilus
startup - wow, waiting for b-a-s to start is a pain. Dinner with pretty
wifelet. Back to dig at b-a-s, grief the translation problem has
exploded in the .server files: 96% of the CPU time is spent parsing
XML; g_markup could clealy use a little optimisation love it seems:
~2Mb of .server file parsed in 240million pseudo-ops, 120 pseudo-ops
per character parsed seems a little excessive.
-
Poked mail, lots of bug mail - wandered around reading and
fixing misc. issues. Played with new laptop from the postman. Noticed
a nomination
for 'Geek of the Week'; what a horrible designation Geek is, nevermind
the tie.
-
My thinkpad overheated & failed; backed up data, and tried
again. Installed real software variously on the new machine. Poked at
gconf some more.
-
Up too early, helped get the babes to school variously. Tried
to catch up on sleep, no success - to work instead. Poked mail, rather
a backlog from the weekend somehow.
-
Pleased to see Alex Larson's artifact
free scrolling work in gtk+. Sorry to see Hub's news. Lunch.
-
More mail, filed queued bug reports from being away from
the net. OPS all-hands call. Amused that the gconf regression test:
'test-persistence' takes ever so long to execute.
-
Up early, breakfast & off to the Eurostar. Poked at gconf
performance some more. Several oddnesses. Firstly it seems that
%gconf-tree.xml is 50% composed of the space
character (pretty-printing, with 8 stop space indents). Knocked up
a trivial patch
to accelerate writing and save ~2Mb (of 5) here.
-
It seems that the split out localisation code doesn't apply
to keys with no default - leaving the localisations inline, adding
25% (700K) to the size of the main tree we parse. Then of course we
have lots of redundancy from duplicate mtime and schema nodes that
could be cascaded into their parents, then there are un-necessary
tags for eg. boolean default values, and empty descriptions. A quick
breakdown might be:
| Size/Kb | Description |
|---|
| 5221 | size of my pre-existing %gconf-tree.xml |
| 2717 | literal white 'space' |
| 729 | l10n for items with no default key |
| 332 | mostly redundant mtime and schema attributes |
| 1234 | (magic) size after removing all of the above |
-
Of course, there is still more fat that can be removed there,
not least in the in-memory storage structures - and clearly in terms of
reducing CPU-usage, removing redundant attributes and elements should be
a priority.
-
Stopped off at the British Library briefly, lunch with Karen,
happily passing through London. More gconf hacking on the train,
saw the news in Cambridge - seemingly a large fire in Beijing fire,
apparently rather nearby both RH & Novell's offices.
-
Knocked up a patch to inherit mtime from a parent node - to
drastically reduce the number of attributes we need to parse; got to
work on collapsing the schema attributes similarly.
-
Up early; off to the Xorg room, got my key-signing goodness
well underway. Enjoyed Matthew's talk, wandered the hall-ways some
more, lunch with Keith & Eric A. Saw Eric B's nice 3D transitions
on OS/X. Caught up with Zonker.
-
Back to the hotel with Doko, starting to understand the Brussels
topology; out to the subtle death place to enjoy the acutely rude waiter
and great company there. Bed late.
-
Up earlyish, breakfast with some Gent-oos, off to the
conference. Wandered the halls catching up with fascinating
people eg. David Axmark, Keith P, Bdale G, Ted T, Danese C,
Michael T, Simon P. Again - FOSDEM has an insane number of
exactly the right sort of people around the place,
interesting & fun.
-
Missed the talks I most wanted to attend due to rooms
being entirely packed full of people - wow. Out for dinner in
the evening with some Gnome & SIL guys.
-
On to the Gnome beer event, caught up with Thorsten,
Matthew & Eric, back to bed.
-
Up early, packed, synched mail, swept drive, off to the
station; to FOSDEM. Poked at odd fglrx issue on the train -
somehow it turned up and broke my almost-perfect radeonhd
experience, bother.
-
Chased an evolution off-line / attachments bug. Missed
the Intel guys on the train, luckily hit the laptop-power
carriage and, by some co-incidence, Andrew Haley, Alasdair
Kergon, Matthew Garret, and co. Spent a pleasant time at
the bar.
-
Off to the hotel, with Alasdair, out for dinner with
Christian & the GStreamer crew; on to the Delerium thingit,
great to see Simon again, catch up with the OO.o guys, meet up
with all the right sort Free-Software hackers that lurk here,
really good fun.
-
Prodded mail, isolated a yast2 bug. Dug at OO.o OpenXML import
issues. Looked at some boot-charts again, filed noddy gdm patch. Profiled
main-menu, deferred 30% of the startup work, lunch. Calls in the afternoon.
-
Snow still lingering on the ground. Poked at boot time
some more. OO.o team meeting, prodded USB debug device with a
new kernel - after much fiddling with Greg, confirmed it seems
like a new type, as-yet-not-supported-by-linux, wow.
-
Played with system-tap to get early_pci_allowed status
out of the kernel, apparently mis-configured on my system and
failing to compile (urgh), pinged Takashi. Eventually, after
learning a lot about system-tap, from the very helpful hackers
on IRC [ surely there must be a better way of debugging the
kernel ], discovered pci_probe is (now) 8 - interesting at
some level.
-
Alpha in the evening with Simon & Andy.
-
Work, lots of mail, lots of bootchart goodness. Amused by
N. helping J. hang up all the wet nappies at great length, then
when J. left, returning them to the washing-machine, causing some
incredible double-take. Wrote up an interview. Chat with Peter,
bug review.
-
Snow, lots of it - spent some time sweeping the drive to
avoid difficulties later. Chewed mail, packed N. and M. off to
hospital with J. No school for H. today.
-
Poked linker bits for Julian, quick call with Janneke,
Compiz call.
-
Off to NCC, did creche with Claire. Chris, Claire,
Keziah & Jadon back for lunch. It started snowing, as Jadon
played in the garden - settling too.
-
Snuggled by the fire, watched balmy DVD; did some research
for a spoof theology of weather, "ban teaching isobars in school"
etc. Psalm 147 - He spreads the snow like wool, and scatters the
frost like ashes. - "ban the sinful divination on the weather
channel!" [ or something ]. Lots of beautiful imagery in there for
brutal twisting with an excessivly 'literal' (or de-contextualised)
hermenutic; should be fun.
In case it's not painfully obvious: the reflections reflected here are my
own; mine, all mine ! and don't reflect the views of Novell, The
Lithuanian Gov't or Arnold Schwarzenegger. It's also important to
realise that I'm not in on the Swedish Conspiracy.
Occasionally people ask for formal photos for conferences,
bio.
or fun.
Michael Meeks (michael.meeks@novell.com)
