Stuff Michael Meeks is doing

• Prodded mail, rather a nasty sore throat coming on. Visited Bert to reset his circuit-breakers. Call with Christian & Kendy. Poked at and fixed gtk/broadway so it doesn't leak / jam modifier key state with v7 websockets, submitted to openSUSE:12.1:Update:Test.
• Pleased to see Stephan's lovely configmgr API cleanup, that should let us make configmgr access even more efficient in the future, as well as being much simpler and more readable now; nice. Of course, also an easy-ish task to help out with: dunging out much less pleasant, old code in this area.
• Horribly frustrated by cups, not only does it insist on pausing the (network) print queue whenever something prints, but finding the un-pause setting [ incidentally hidden in one of two combo-boxes in the printer maintenance page ] was extremely non-intuitive. Filed misc. bugs, eventually got something working, it seems adding ipp://.../ipp?waitjob=false&waitprinter=false is a good idea.
• Clobbered some gtk3 theme color issue. Poked Lowell, Ciaran, Gerald. Sync. with Martyn. Dinner with the parents. Hacked on this & that, sat by the fire chatting to the parents and poking android emulators until I got some.

• Up, practises, babes off to school. Chewed mail, out to a funeral of Jane Hancock (Dave's wife), back to mail, tripple (and more) patch review etc. Admin / status report writing. Lunch.
• Poked at the ClamAV signature databases, found main.avd, then dug around for the source for them. Poked Ciaran, chat with Simon.
• Dinner, Dave around for Bible study & catch-up, good chap.

• Woken by babes being noisy in the morning; off to NCC, Tony speaking. Had Justin & Karen back for lunch & sat by the fire with them for the afternoon.
• Babes had a Marx (Engels, Groucho, or & Spencer?) movie while we tidied up, played with & then put babes to bed. Hyperactive slugging in the evening.

• Up late, breakfast. Pair of Jehovah's witnesses around for an hour of interesting, and widely ranging dialog & friendly discussion, on the 144,000 Jewish, male, virgins for example. Or why the inspired authors of scripture translated the tetragramaton as kurios: lord etc. and whether 'name' is rather more of a profound concept of identity and goodness than even a profound label can be.
• Checked mail, Sue, Clive & family arrived, then Bruce & Anne on much improved form, eventually Auntie Louise. Big family, birthday-time lunch.
• Anthony & Tim arrived from the cricket later, in time for H, N, and M's trio. Good to catch up with Tim a little afterwards.
• Sandy arrived to baby sit, out to the Pregnacny Crisis Quiz - happily heavily over-subscribed; 65 participants, had to use the creche tables ourselves; much fun (quiz, and puddings) had by all. Bed late.

• Up, mail chew; pleased to see three guys interested in porting LibreOffice to OpenIndiana (using gcc), poked at misc. bugs. Started working on talk about Easy Hacks for FOSDEM, cleaned up the wiki page. Sudden inspiration for android - pushing all rendering to the main thread; failed - still inexplicably hangs in ANativeWindow_lock.
• Switched to use EGL instead eglLockSurface - runs beautifully, returns no errors, and a NULL pointer from eglQuerySurface(...EGL_BITMAP_POINTER..) - stymied again. On to trying glTexImage2D - surely that will get a lot of pixels to the screen fast, lots of error free gl calls, with a blank screen to match; hmm.
• Out for a run, read stories, fish soup for dinner (to assist the brain), started adding sample rendering code through the startup process to see where it goes wrong, and reviewing patches in the idle cycles.
• Up extremely late, eventually discovered that the problem seemed to be caused by not processing (unexpected) events on the thread's ALooper - which causes aforementioned lockup; nice - finally an up-side-down, wrong color, no font VCL error dialog windows.

• Up early; practise with the babes, back to the window locking oddness. Nice post from Fridrich on making the best of FOSDEM.
• Managed to get some pixels onto the screen eventually, team meeting, ESC call, Vojtech's staff, pixels no longer going to the screen: meetings are like that it seems ! Dinner.
• Switched to EGL rendering to attempt to retrieve pixels from locked up system, no joy there either but more interesting errors. Setup a new cppunit git repo on freedesktop for Markus, apparently it is unmaintained & we really need more features.

• Chewed mail, quick call with Vojtech, then Charles. Finally got around to submitting a LinuxTag paper or two. Lunch. More mail, patch pieces.
• J. out for Rosemary's leaving pizza party. Up extremely late poking android's wedging on ANativeWindow_lock - sadly the debugger gives no trace: an thread un-attached to the VM ?

• Up early, misc. mail chew, question processing, patch review, re-building action etc. Inched through more startup problems, Lunch.
• Chat with Kendy, more mail cleanout. Lydia over for dinner. Up late hacking android main-loop pieces with Tor.

• Mail chew, read the git commits over the weekend. Call with Simon, improved the LibreOffice donation page to include a nice image rotation.
• Misc. android hackery - got past several unpleasant roadblocks in the UNO bootstrapping. Reviewed slideware. Dinner, babes to bed. J. under the weather, but out to a meeting. Back to the hackery - started on the first-start, user-installation creation code.

• J. dropped me to NCC to practise with the band, service, Thea spoke. Back for lunch with Keziah over. Out to a service of Christian Unity in the town. Back. Played games, lazed on the sofa. Tea, told stories to babes, put them to bed & read more stories. Sermon from Hugh Palmer, silly Naked Gun movie, bed.

• Up earlyish, H. and N. off to Bury to do music & book buying. Cleaned the house up, hacked a bit at some androidish pieces: discovered some problems with unit tests not being compiled.
• Lunch, Mary Rogers over in afternoon, sat by the fire and played with babes. Lydia over in the evening - more hacking at sal/ stopped readLine corrupting/writing to it's input buffer and crashing and fixed misc. build issues.

• Chewed mail and misc. vcl fixing - and finally calc unit test runs to completion on Android (great work from Tor); getting the process slowly better documented in README.Android and no pixels yet of course. Chat with Simon, then Charles.
• The LibreOffice FOSDEM Devroom Schedule went live - a really great set of shortish talks (to get the most grist we can into eight hours) and some great speakers, I'm really looking forward to it.
• Planned my day, interspersing the tedious stuff with fun hackery, so that at least -some- tedious things get done. Lunch.
• Filed a few more easy-hacks around cleaning up the horrible old tools/ - a duplicate system abstraction that still malingers underneath LibreOffice. Hid a few more unused locking methods in SvStream, and made the FSysRedirector more obviously a no-op. There are big blocks of easy-to remove cruft in tools needing a beginner or two.

• Poked mail, fixed misc. build problems, poked at and spent the morning extending the tools/ stream abstraction also fixing some build issues. Nice to finally get some hacking done.
• Chat with Jonathan, lunch, SUSE team meeting, LibreOffice ESC call, Vojtech's staff, discovered I'm late at my travel budget planning; bother.
• Pondered the LibreOffice team. There is one set of very skilled hackers that perhaps people don't notice. As of today, we have quite a chunk of people working full-time on LibreOffice that used to be on Sun's OpenOffice.org team (in order of migration): seven guys: Caolan McNamara (RedHat), Noel Power (SUSE), Thorsten Behrens (SUSE), Bjoern Michaelsen (Canonical), Stephan Bergmann (RedHat), Eike Rathke (Redhat), Michael Stahl (RedHat) - making (I think) the largest concentration of full-time ex. StarDivision hackers on any project with a nice cluster in Hamburg still. It'd be great to grow that list of course.
• More hackery and build fixing; late call with Camilo. Read babes stories, J. out for a run, final emulator hackery and bed.

• Up lateish, call with Vojtech. Dug through the mail, until finally, in the afternoon - got to a little Android hacking; fun.
• In connection with the somewhat irritating MS Office 2010 issue I mentioned last Tuesday, I was somewhat startled to see Rob Weir change tack to a new attitude about standards:

  I have no wish for the ODF standard, like the US Constitution or the Bible, being used as an excuse to justify stupidity. ODF is a specification for document exchange. If you are using it in a way that decreases interoperability then you really need to step back and ask yourself if your literal interpretation really makes sense.

  Of course, amazingly the implication is that it would be 'stupidity' to follow the spec. and produce documents that are valid ODF 1.2 (as LibreOffice 3.5, and the Apache OpenOffice 3.4 pre-release builds do.
  Then we have this gem:

  If a program does not meet user expectations then it is a bug. If you want to be compatible with Microsoft Office then you need to play by their rules. ...
  In any case Seeing responses like this from LibreOffice makes be very optimistic about the future of Apache OpenOffice. Whatever the cause, the fact that LibreOffice ships with this problem shows either a woefully inadequate QA program, or total indifference to real world requirements. Even testing a single LibreOffice document in Office 2007 would have shown this bug. Is that too much to expect?

  This is really a deep & rich lasagne of irony, I'm really trying to work out which bit is most tasty, could it be - first the aggressive, purist, open-standards champion advocating deliberatly writing non-conforming output, and making ODF 'play by' Microsoft's rules ? Or - could it be the fact that (apparently) the TC chair hadn't bothered to validate or test changes to his standard in 'real world' office suites, but rather prefers to deflect attention at 'woefully inadequate QA' to a single implementor: LibreOffice. Or finally could it be that he hasn't noticed that his own Apache OpenOffice implementation actually does the same thing. Hard to choose really; mystifying; checked colour of the moon to make sure: apparently not blue.
• Of course, personally, I'd love us to have a good solution that ensures maximum interoperability while conforming, no doubt we'll find some way to work out what that is in the end. It is clearly not just as simple as removing that single version attribute: forwards compatibility is somewhat tough - but starting with backwards-compatibility is prolly sensible. Since ODF 1.2 is not completely backwards compatible with 1.1, knowing what the version is is rather useful for correct interpretation.
• Emily & Sarah around to read stories, share dinner & sit up late to chat - lovely to spend time with them.

• Crawled from bed, practise with babes, rapid mail triage, into Cambridge for interview at Sticky Beaks, dropping obsolete computers & electronics off at reboot on the way - somehow I feel like I lost something; the CPU count is substantially down, but cat swinging is possible. Lunch & home.
• Dug at some bugs, Roger W. kindly dug out a PPTX test document to isolate a performance regression. Put babes to bed, Lydia over for dinner. Chewed through getting the LibreOffice dev-room schedule into Pentabarf, phew. Poked mail, worked rather late.

• Up early, mail chew, patch review, license statement collation in the wiki. Chat with Holger. Lunch. Split out multi-screen / display fixes and back-ported to -3-5. Poked at detecting old file-systems to avoid fsyncing on them.
• Dinner, put babes to bed, call with Pete , worked late.

• Up; off to NCC, Tony speaking on Truth. Back for lunch, lazed around, out to a Baptism (or two) of Charlotte & Jane. Back for more slugging, watched some Marx brothers comedy.
• Put babes to bed, tidied the office with J. while listening to a Rico Tice sermon on Psalm 1. Dunged out lots of obsolete electronics and boxes of things no longer needed in the modern world.

• Up, took H. to try out some choral and brass music lesson / practise goodness in Bury; hacked away at multihead work while there. If only we didn't have two types of integer indexed screens, both called 'screen' in the code, with tangled nomenclature everywhere. Started to push through some big cleanups.
• Home for lunch, tidying and prepping for Naomi's party. Helped games: guessing things in pots from the smell, pass-the-parcel, party food, and expended a four-year-old box of fireworks from my parents - lots of smoke.
• Put babes to bed, back to hacking multi-display issues.

• Up early; walked babes to school. Chewed mail, watched Bryan M. Cantrill's Lisa talk on Illumos, it's always nice to listen to someone that gets it - Solaris users look like they'd do well to switch to Illumos (to me). Particularly amused by his take on Apache as a template (27:00), Oracle (34:00), the closing OpenSolaris (41:30) and more; wow.
• Mail, patch and bug bits, lunch. Spent the afternoon working away at a vile multihead bug of my own creation around the gtk3 port busting gtk2 as well. Worked late.

• Mail chew; poked at systemd socket cleanup again, a new attempt at beautification, more mail, patch review and misc. bugs. Estate Agent came over to value the house.
• More mail, patches. Team meeting, TSC meeting, wrote up minutes, worked away at FOSDEM dev-room scheduling - an encouraging set of speakers this year, really looking forward to meeting up with them.
• Emily kindly over to baby sit for us; J. out to swimming with Miriam; out for a drink & an Indian meal with J. - lovely.

• Mail, call with Vojtech, another with Pierre-Yves, dug at my systemd tmpfile cleanup issue sent off a patch. Lunch. Call with Pierre-Francois - exciting to meet so many interesting & helpful new French people.
• Out to cell group in the evening, good to catch up with the crew after a long break.

• More mail, and admin. Surprised and saddened to see that MS Office 2010 has a built-in allergy to ODF 1.2. Worked through a FOSDEM interview.
• Lunch, call with Tor, a little hacking, chat with Philippe Desmaison. Dinner, Lydia & Janice over. Spent some time struggling once more - trying to get simple VCL samples to run: once again the endless life-sucking experience of UNO bootstrapping.

• Chewed mail, poked at the perennial fsync issue, chased my sudden socket death issue - it seems our compile doesn't kill the sockets. Wrote status report & other monday admin.
• Dumped graph, booked FOSDEM travel arriving midday on the 2nd, leaving midday on the 6th. Out for a run, dinner, J's pregnancy crisis group over. Prodded at the android sdk/ndk with some success, and tried to get Eclipse installed, despaired of my inkjet, and bought a color laser printer.

One of the unfortunate things that LibreOffice inherited, as part of the several decades worth of unpaid technical debt, is unused code that has been left lying around indefinitely. This was particularly unhelpful when mingled with the weight and depth of the useful code we have around the place. Caolan McNamara of RedHat wrote a beautiful tool callcatcher that identified these unused methods, and in recent times in LibreOffice we've had an unusedcode.easy file in our toplevel with a list of methods that should be removed. It's pretty easy to find and expunge a method or two, with a quick git grep, and dropping a patch to the developers mailing list. To escape from a pile of administration recently, I knocked up a pretty nasty perl script to parse the git numstat output, to see how we're doing. That produces a fun graph:

It seems that over half of our unused code has now bitten the dust. Uunfortunately as we remove more, more wasteage tends to be revealed, which explains some of the upward jumps in the graph, nevertheless the trend is clearly down. One of the side benefits of the unsung heros working at the conversion of our old-style macro driven generics to modern STL is that this looses us several unused methods per class converted.

If you want to get involved with LibreOffice development, it doesn't get much easier than this - please do check out the code and have a go. For the more adventurous finding an unused destructor, without a matching unused constructor is proof of a leak that needs chasing, of which there are a handful.

Failing that, why not run Caolan's callcatcher over your project to see which nooks and crannies are surplus to requirements.

• NCC, Tony spoke on the Way - back home for lunch with Claire, Simon, Phoebe & Tally good to see them. Andy & John Madden around for dinner (the latter sadly moving away to Exeter tomorrow), spent some time writing up some chunks of his interesting biography.

• Lazy day, slugging, and lots of house cleaning preparing for a valuation; amazing amounts of 'stuff' get stuffed everywhere over time. Ruthlessly pruned baby bits, sorted many workshop items cleaned all those potential metal splinters out of the zone. Out to Newmarket Open Door to drop things off, business until late.

• Up early; mail chew, into Cambridge for Lunch, back, wrote LXF column, chased some tango / artwork licensing. Dinner. Interested to read about the future of CouchDB. Built some tooling to generate statistics on our dead code removal.

• Up early; more mail reading and backlog unblocking. Some patch review, commit account creation, and account management. Lunch.
• More slog; team meeting, TSC meeting, posted minutes. Chat with Ross, built some stats. Dinner.
• Tried to buy Deutche Bank Ethical ETF's (listed on the LSE interestingly) using Barclay's wunder-trading website. Seemingly not possible - it appears they list only some subset of funds they happen to like - que ?

• Up rather early - the iron routine of school is back again; to work. Waded through mail. Bitten once again by finding my gnome-keyring totally non-functional, filed a bug - it seems systemd's over-enthusiastic cleaning up is to blame; after 10 days it appears to do the moral equivalent of rm -Rf /tmp/* destroying the gnome-keyring socket in the process. In the meantime:

  sudo systemctl disable systemd-tmpfiles-clean.timer
  sudo systemctl stop systemd-tmpfiles-clean.timer

  No doubt just a mis-configuration that can be cleaned up; I wonder if it trashes ORBit2's sockets too - though perhaps it ignores sockets.
• Chat with Frederic, caught up with Norbert & others on IRC; started digging through the mail backlog with vigour. Read the git commit logs over Christmas - lots of nice work and fixing. Lunch.
• Noticed a new and rather sexy General Polygon Clipper library with the boost license. That looks lovely for repsnapper's slicing code; and Martin "Hurzl" Dieringer already uses it, neat - roll on faster, better slicing.
• Prodded the bug lists variously; set some fresh from-clean builds going.
• Saw that Luis (good to his word) and Gerv finally got the MPLv2 out - which is lovely. That gives us an AL2 compatible (and more importantly) GPL* compatible weak-copy-left license; more in the FAQ. A crisp read too. Hopefully there is no longer a need for CDDLs in today's world.
• Dave & Emily over, put babes to bed, dinner, sat around and talked by the fire. Finished the spreadsheet and filed tax return.

• Last day of the holidays; frantic tidying, cleaned out an unpleasantly blocked U-bend, fixed the heated towel rail, re-attached a malingering door-handle etc. Lunch. Tried to dig a tiny metal splinter out of Hannah's foot (most odd) with no real joy.
• Pollyanna in the evening, and tax filing afterwards.

• Up earlyish, off to Bruce & Anne's, nice to see them. Lunch, out for a walk on Aldeburgh beach, unseasonably temperate, and queues for fish & chips - amazing. H. feeling unwell. Caught up with David M. in the evening.

• Up earlyish, drove to Hertford, preached, talked afterwards with some interesting people; back to Helena & James' for lunch with Rose, Mike & Anne - lovely food, fine company - home late.