|
| Greg's diary |
| March 2006 |
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Wednesday, 1 March 2006 | Echunga | |
| Top of page | ||
| next day | ||
| last day |
Finally things are not quite as busy as they have been, and found time to get through most of my mail. Had a long discussion with Elliot about the approach we should be taking. I had thought that my discussion last night represented a complete shift from what we were talking about, but Elliot didn't see it quite like that. Why do I have such problems understanding? Anyway, added stuff to the work log, and it looks like I have a way to “go forward”.
Phone call with a partner in Bangalore in the evening; got more discussed than two weeks ago.
| Thursday, 2 March 2006 | Echunga | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Phone call with a partner in the backup scene this morning (or yesterday afternoon for them). I'm learning lots of new things, mainly buzzwords and acronyms. Most interesting recognition of the call: “XBSA isn't dead” (yet).
My new change of direction involves the design of a streaming interface to the storage engines. At least it'll give me a chance to get my feet wet in the code; after all the general work I've been doing, it's also refreshing to do something concrete.
| Friday, 3 March 2006 | Echunga | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
It seems that I spend at least an hour on the phone every day, maybe closer to 2. Apart from that, finally got some kind of high-level specification for streaming backups. I'm not at all happy, but at least I have something down in writing.
My laser printer has had low magenta toner for a couple of months now—in fact, since before I bought it, since the cartridge came from its predecessor. Today I got mail:
The next message describes the resolution:
| Saturday, 4 March 2006 | Echunga | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Quiet day. Bottled some beer, crushed some grain and pottered around. It's too hot to ride.
| Sunday, 5 March 2006 | Echunga | Images for 5 March 2006 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Another day with little to show for it. Started off in a leisurely way to brew some more beer. I've been having problems with chill haze in some recent brews, so reviewed the mashing methods and decided to mash in the boil kettle, which enabled me to make a much thinner mash. Hopefully that will help.
For some time we've had mushrooms growing in our garden; there seem to be many more now. Last year we examined them, but couldn't find any certainty as to whether they were edible or not. I had more or less come to the conclusion that they were Agaricus Campestris, and it seems that about the only dangerous thing they could be would be Agaricus Xanthodermus or Agaricus Californicus, both of which are only mildly poisonous (cause gastro-intestinal problems), and which some people can eat anyway. They both go yellow when cut, and apparently taste pretty bad. Decided to take the risk, but was rather put off by the way one of the mushrooms went bright pink when cut:
Still, they smelt good, and I fried some of them and finally tried a slice of the pink mushroom. It tastes surprisingly good, and I'm still alive.
| Monday, 6 March 2006 | Echunga | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Lots of different things to do today: a review of a bug report, which apparently still suffers from memory leaks; a showstopper which only occurs on Apple (and my Mac has been down for a couple of months now with a defective power supply, and I couldn't find another one to test on), and more on the backup design issue, more political than technical. Also an interview with a bloke in New South Wales; sounds clever, but he doesn't have much database experience. What didn't I do? My planned design work.
| Tuesday, 7 March 2006 | Echunga | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
We're having irrigation problems again! For some reason, though the dam is full, we're getting only a trickle down at the house. It's a 50 mm pipe, so it doesn't seem likely that it's blocked; the best guess is that the uptake in the dam is clogged, possibly with a film of algae. Fortunately we have a new gardener, Scott, who is prepared to come and take a look tomorrow.
Another day of interruptions, but did finally manage to update my worklog entry for the backup design. That's a load off my mind.
After my first attempt at eating our mushrooms on Sunday, made a complete first course of sautéed mushrooms today. They taste OK, but nothing spectacular. No adverse reactions, so I suppose we can agree that they're Agaricus Campestris.
AUUG board conference call in the evening. I hope we don't have them too often.
| Wednesday, 8 March 2006 | Echunga | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Nine years ago I returned to Australia to decide whether to stay or not. I did, and I've now been living here (in Wantadilla) longer than anywhere else I have ever lived. While on that nostalgia trip, did some investigation and discovered a certain 9 year pattern in my life:
In about March 1961, I was living in Kuala Lumpur and not doing very well at school. My parents decided to send me to England for further education.
In early 1970 I was at Exeter University. The previous (Christmas) vacation had been terrible, since I didn't have anywhere to stay. I learnt out of that mistake and rented a cottage in Bow, a little to the north of Dartmoor. On 3 March 1970 (I was keeping a diary at the time, so the date is correct) I moved in to Comp Cottage, a name that people immediately associated with computers (in fact, it probably relates to Cob, the mixture of mud and straw from which the cottage was built). This was my first real independent home.
In early 1979, my wife Doris and I were threatened with eviction from our rented house in Oberhausen/Rheinland (the owners wanted it back), and we ended up buying a house in the Loefflerstraße in Voerde/Niederrhein, our first own house.
In 1988 things aren't quite so clear: I was living to the north of Frankfurt/Main in Rosbach vor der Höhe with my second wife, Yvonne. At about that time we decided that we wanted a more rural lifestyle, and started looking for appropriate accommodation. That took a long time, though, and it wasn't until January 1991 that we moved in to the house in Schellnhausen.
Finally, on 8 March 1997, I arrived in Melbourne after spending a couple of weeks in Beijing. At the end of the month we found Wantadilla and signed a contract for it; we moved in on 17 July.
Scott along to look at the water supply problem, and fortunately fixed it pretty quickly: the suspension for the water supply from the dam had failed, and it was sucking in weeds from the bottom. Like all good bugs, it was two different bugs which negated each other's symptoms: normally falling off would have sucked lots of mud into the pipe, and we would have recognized the problem. But the weeds didn't just clog up the intake, they also filtered the water so that it came out clear.
I'm off to Sorrento on Saturday, and decided to update my laptop with the latest version of FreeBSD-CURRENT, since I'd been having some issues with the older version. That didn't work well: building xorg didn't work correctly, and I ended up with a number of conflicts. I suppose none of this would happen if I just installed from DVD, but it's frustrating how regularly the problem occurs.
On the work side, spent most of the day looking at the increasing number of bugs assigned to me. That wasn't as easy as I thought: in Uppsala I only have Opteron processors available, and I had difficulty with one particular bug, so decided to do it locally. That, too, wasn't as simple as it should be: autom4te still complains about NFS, which probably is a FreeBSD bug, but I don't have time to look at it, so installed the source trees in my main system, wantadilla.
wantadilla has been doing some funny things lately, including SIGSEGVs and strange behaviour of some programs. I had reset the rather overclocked BIOS parameters to normal a while back, but while compiling MySQL it spontaneously rebooted. Ended up replacing the entire base system with the machine that's been sitting around waiting to be our next DVR (gaining a bit of processor speed in the process), and it now seems stable. I can find out what's wrong with the other system when I get back from Sorrento.
I still don't have a replacement for my phone, which got its ringer burnt out in December. I'd like a DECT phone, but I need one which supports two external lines (office and home). Some have handsets that handle multiple base stations. On a recommendation, I took a look at the Oricom web site. One plus point for them is that they have the instruction manuals online. The manual for the one I was recommended describes how to select base stations on page 65. The description implies that it expects all the bases to be connected to the same POTS line, so that the only purpose of multiple registration is for roaming in a larger area. In particular, there's no way to receive calls from different base stations, or to easily select a base station when making a call.
Sadly, it seems that the DECT manufacturers have lost the plot here. I've followed up other manufacturers as well, but they all (implicitly) support only a single POTS line. I called a Telstra shop, and once I explained to the person on the other end that a phone line was not the same thing as a handset, she went away and checked. They don't have any phones that will handle more than one POTS line.
This is progress? I've had a dual line phone on my desk for 14 years now.
This season's hops are finally ripening. Did a partial harvest of the old Pride of Ringwood and got 140 grams; with any luck I'll get over a kilogram total.
| Thursday, 9 March 2006 | Echunga | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Admin work kept me busy for much of the day, though I got access to an Apple machine and had a chance to investigate some bugs. In the afternoon, procmail stopped working, dumping incredible quantities of spam into my inbox. Looking at /var/log/maillog, I saw:
Fine, but what's “suspicious”? rcs confirmed that the file hadn't changed in nearly two months, and there was nothing obvious. Finally I found the reason: I had changed the permissions on my home directory from 755 (drw-r--r--) to 775 (drw-rw-r--) so that Yvonne could modify some files in my home directory. That seems to be what is worrying procmail. But:
What a misleading error message!
What assumptions it makes about what is secure and insecure! This is on a network which is almost completely firewalled from the outside world, and with only trusted people.
If a program is so picky, it should have the configuration file in its own directory, where it can set the permissions as it sees fit.
| Friday, 10 March 2006 | Echunga | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
I'm leaving for Sorrento tomorrow, but there didn't seem to be much to do to prepare for it. Obviously I'm getting back into the swing of frequent travel again; after all, last month was the first full month that I've been with MySQL where I didn't do any travel.
To town for an ADUUG lunch, at Chopstix in Rundle St. I'm sure I've been there before, but I suspect that they've changed the name. Talking about video again, and the topic of Mac Mini came up. I've had a rather ambivalent relationship with Apple, but the idea of video that Just Works is very compelling. On the way home dropped in to next byte on Glen Osmond Road and had a look at one. It's not completely reassuring; there seems to be no way to set “non-standard” resolutions such as 1280x720 (the standard 16x9 projector resolution), and I'd have to rely on the machine getting the correct information from the projector—after I've bought it. The box also has no expansion slots, so I'd need to use a tuner such as the EyeTV 2, which, they say, can only be controlled by this STUPID toy GUI interface with the mouse, just the thing that I curse about with gmplayer every night. Something as simple as a remote control requires a significantly more expensive tuner, round the $500 mark (but they come free with the sub-$100 PCI cards for PCs). Food for thought, but no instant solution.
Also on the way back, dropped into the SEO in Stirling to vote for next week's state elections. I don't like many of the Liberals' policies, but Isobel Redmond seems to be trying to do the right thing about traffic, so I voted for her anyway, despite her party connections.
| Saturday, 11 March 2006 | Echunga → Melbourne → Singapore –> | Images for 11 March 2006 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Started off on the long journey to Sorrento today, first harvesting some more hops. The new Adelaide terminal is now completely open, and they've changed the Qantas Club as a result. Took at look and discovered RJ45 jacks at the desks:
|
|||||||||||||||||||||||
They even lent me an Ethernet cable, for what use it was: the jacks are for telephones. This is a brand new facility: why have they installed 20th century technology? It's not as if they don't have Ethernet, but it's reserved for a few IBM PCs running Microsoft. Ended up doing what I always end up doing and hijacking a PC connection:
|
|||||||||||||||||||||||
It's helpful that they have a normal DHCP server, so getting access is no problem.
Out first 700 km away from Sorrento, to Melbourne, passing over a good view of the river Murray, where we spent a not completely harmonious weekend last year:
|
|||||||||||||||||||||||
That photo shows the entire distance we went on that weekend, and then some.
Arriving in Melbourne was a sort of “full circle”. I was born in Melbourne, but I left when I was five years old. When I came back three years later, it was after flying round the world, so I was “out of phase” by a day. I only returned in the other direction last year, and this was the first time since then that I have been in Melbourne, so in a certain sense it's a homecoming after over fifty years.
Some homecoming! Tullamarine airport isn't as bad as London Heathrow (nothing could be that bad), nor even Frankfurt/Main, but the impossible signs meant that even though I knew where I was going, I got (slightly) lost. I can't imagine anybody who hasn't been here before transferring from a domestic flight to an international flight without help. The following photo is atypical: it shows a sign, just an ambiguous one. Most of them were either missing or just plain wrong.
|
|||||||||||||||||||||||
On to Singapore, where I ended up at the same desk I visit every time. Previously I connected my own Ethernet cable, which involves grovelling under the desk. This time I learnt from Adelaide and just hijacked the existing cable:
|
|||||||||||||||||||||||
It's amazing that nobody else does this. Most people seem to be happy to use third-party computers and compromise their passwords:
|
|||||||||||||||||||||||
| Sunday, 12 March 2006 | –> London → Roma → Napoli → Sorrento | Images for 12 March 2006 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Uneventful flight to London, where I arrived ahead of time at 5:20. Why do these flights always have to arrive so early? Temperature was -1°, and I was left happy that I didn't have to stay there. The British Airways club has computers with Internet access too, but they make sure that nobody can connect their own computers to them by locking the computers in a box. Only the monitors and keyboards are accessible. There's 802.11 networking too, at a price I could refuse: £6 per hour. Why do people think that this is a good way to treat their guests?
On to Rome via the West of the Alps:
|
|||||||||||||||||||||||
In Rome I had some difficulty finding Shane Bester. His phone was discharged, which didn't really make any difference: he also didn't have international roaming enabled. Down by train first to Napoli and then on to Sorrento. I've had some good views out of my hotel room, but this one must be one of the best:
|
|||||||||||||||||||||||
We were tired, so had dinner in the hotel, and somehow managed to stay up relatively late.
| Monday, 13 March 2006 | Sorrento | Images for 13 March 2006 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
First day of the developer conference today, starting with the “3M” plenary sessions: Maurizo, Mårten and Monty, the latter talking with Patrik about his latest project. After lunch was a little more relaxed. Listened to Kostja talking about the batched range read handler, which required a little more understanding of the optimizer than I have, and Guilhem talking briefly about Valgrind, just enough to whet my appetite.
Then time for team meetings, for which I didn't have much, so took the chance to catch up with mail and also discuss some of the issues of builds on BSD. Similar issues apply to Debian, it seems: there are two different ways to get MySQL binaries for all these systems, and the results are not necessarily the same.
Then Stewart talking about cluster replication, not quite what I was expecting, but quite interesting; I was left wondering whether the presentation techniques of OpenOffice weren't an advantage after all.
Then Brian led a session on showstopper bugs, of which I had one.
Dinner in the hotel in the Ginestre room, which included a large pond:
|
|||||||||||||||||||||||
They also had some musicians, presumably mimicking Neapolitan street musicians, who kept things up for most of the evening:
|
|||||||||||||||||||||||
All in all a very pleasant evening.
| Tuesday, 14 March 2006 | Sorrento | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Up early and in to another show-stopper bug review to give my report that I hadn't done anything. Then more VP reports, followed by other departmental presentations, which kept us busy until lunch. In the afternoon, Kostja did a code walk-through, which evoked a fair amount of interest.
Then LenZ did a presentation about how developers could help the community, which was depressingly badly attended; at the end, I was the only person still discussing the topic. Spent some time with LenZ and Kaj discussing how the community could contribute code which would be worthwhile. Came to some useful ideas, but it went on for a while, and I didn't even get to discuss the build farm idea.
Then a bit of work on my showstopper, which went slowly. It wasn't until the evening review—where Brian was a no-show—that I discovered that the system (Apple Mac OS X) was storing core files in /core; I later found:
=== mysqldev@powermacg5 (/dev/ttyp4) /Users/grog/5.0-Bug-15671/mysql-test 66 -> sysctl kern.corefile
By then I didn't have time to look at the details; people from the local community came along, unfortunately very few (only 4); being away from big cities doesn't make it any easier. Went to dinner with Elliot and the replication team instead, and relatively early to bed.
| Wednesday, 15 March 2006 | Sorrento | Images for 15 March 2006 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
I'm not sleeping enough! Up again before the alarm went off, and woke up some time during breakfast. After the showstopper updates and announcements, a good presentation by Dennis Wolf, who looks surprisingly familiar.
As I discovered back in December, using Italian power connectors is not always easy; as a result, the tables are littered with adaptors and power boards:
|
|||||||||||||||||||||||
Spent some more time talking to Kaj and LenZ about code contributions, and managed to hold Brian down long enough to confirm that he also thought it was a good idea.
In the afternoon, the most interesting talk appeared to be Mats' talk about libtap, a way to write unit tests. I was apparently alone in that opinion: nobody else showed. Like yesterday's lack of interest in the community, this worries me. I don't necessarily agree with the idea of libtap, though I'll give it a go, but the idea of unit tests seems so basic that I can't understand why nobody showed.
Then the next showstopper review, after which a late meeting about backups, where we didn't make even as much progress as I had hoped.
In the evening into town for pizza and karaoke. Much fun was had by many:
|
|||||||||||||||||||||||
Also spoke to Dennis: turns out we knew each other at Tandem; I later found a mail exchange with him from October 1984. How time flies!
| Thursday, 16 March 2006 | Sorrento | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
This showstopper bug review is getting boring. Elliot managed to close one for me (“duplicate”), but we still have more showstoppers than we did at the beginning of the week. We have better things to do this week.
Kaj was on first with his community update. It's interesting how the general feeling (not just from Kaj) is that we should be more open in what we do. We already have the Bugs database open for all to see (so anybody can see the bugs assigned to me), but there are suggestions that we should open up the internal worklog system too, which is currently private. I'm ambivalent about that; some of the content is confidential, and I can't see a way around it. We'd have to find a solution for that.
Then a Frenchman with the unlikely name of Mathias Herberts from the Crédit Mutuel came and held a very good talk about why Crédit Mutuel is using MySQL.
Then Elliot talked about replacements for bitkeeper. The only ones that he is considering seriously are BazaarNG and Mercurial. He finds problems with both of them, but also prefers BazaarNG. Based on what he said and also what I have heard, I'd be more likely to choose Mercurial.
Next Jani with an AutoPush script, still a work in progress. Having seen similar things in the FreeBSD project, I'm ambivalent.
In the afternoon a presentation about VMware, then Ann Harrison on how Firebird and Netfrastructure retrieve data. First they read the indices and build up sparse bit maps for each record that meets individual criteria, then they merge the bit maps and read the data records, thus optimizing I/O.
After that talked with Chad, who has run into an incompatibility with the Apple Intel assembler: it doesn't recognize the clrb and clrl aliases (for xor variants), and it also doesn't recognize the .size and .type directives. Then grabbed Robin Schumacher and Jimmy Guerrero and talked about backups; we're making progress, but we're not there yet.
Then a surprisingly interactive presentation by Stewart on file systems, which went on much longer than intended, making me miss the showstopper review for the first time. I am devastated!
In the evening, people went into town for dinner—team for team. It would have been nicer if more people had chosen to go with people they didn't know so well. I went with the QA and build team. Had quite a pleasant dinner
| Friday, 17 March 2006 | Sorrento | Images for 17 March 2006 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Yet another showstopper meeting this morning; strangely, Brian didn't even mention my bug. A few hours later I found out why: he had decided that it wasn't a show stopper after all, and degraded it. So now I no longer have to participate.
After announcements, the systems admin team gave a status report and waited for brickbats, which weren't forthcoming. I think they're suffering from the paranoia of support people: everything's broken, nothing works. Just ask the customers. In fact, people were relatively happy.
After the break, Timour talked about hash joins, unfortunately leaving me with more open questions than I wanted to ask during the presentation. Probably a case for doing a bit of reading. Jim Starkey after that, with a presentation pretty much as I expected. Monty took exception that he didn't rate Open Source as one of the landmarks in computer history. Jim obviously has a different background from the rest of us, but I also think that Monty and Jim are talking past each other.
After lunch chased up a number of people on different topics, which managed to keep me busy all afternoon. Tried to attend a talk on MySQL Query Browser, but it didn't seem to happen, so spent some (more) time trying to compile it, still with no success.
In the evening to the Ristorante Tasso for dinner with some of the development teams:
|
|||||||||||||||||||||||
Good food. Somehow ended up going to bed far too late again.
| Saturday, 18 March 2006 | Sorrento | Images for 18 March 2006 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
First thing this morning was my presentation on BSD, which was very well attended. Also had a lot of questions, with the result that I overran a 45 minute presentation slot by 15 minutes.
After the break, a talk by Adam Bosworth of Google, telling us what real-life databases need (not the current relational model for one, he says). Interesting talk, but I'm wondering to what extent he's right. On the other hand, Google definitely know what they're doing. Food for thought.
Then Sanja with “Tips & Tricks”, which turned out to be some interesting stuff about testing the server. I knew most of it, but not all.
In the afternoon talking with Jörg Brühe about building under BSD; we still have some work to do there.
Then a talk about reimplementing the binlog as a storage engine. Once again I had the feeling that we were being presented with a potential solution rather than discussing the issues. In particular, I'm concerned that nobody else is concerned about performance.
Farewell dinner in the evening, in the same restaurant that we visited on Tuesday. Spent a lot of time having a heated argument^Wdiscussion with Jim Starkey, with the unexpected result that I drank much more wine than I had intended to. I blame it all on Jim.
| Sunday, 19 March 2006 | Sorrento → Amalfi → Positano → Sorrento | Images for 19 March 2006 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Up about as late as I could this morning, and barely managed to grab a bread roll for breakfast before we set off sightseeing on the Amalfi coast. I've been here before, in August 1983, but we didn't have time to visit the Emerald Grotto on that occasion, so I hoped to make up for that today. Unfortunately, it was closed. Looked around Amalfi, which I didn't do last time either. Nice cathedral:
|
|||||||||||||||||||||||
While walking through the cloisters, managed to drop my camera on the ground. It didn't survive, so quickly down to buy another one, a Nikon “Coolpix” L1. Higher resolution (6 MP), but of course not as good a wide angle lens. Still, maybe the flash works properly.
Back to Positano for lunch, where we were left to our own devices too long, then back to the hotel, where I caught up on my sleep a bit before going to dinner with the Berlin contingent, eating a fagottino mainly because of the name:
|
|||||||||||||||||||||||
It's obviously a far cry from the other fagottino I have, the one made by Guntram Wolf years ago:
|
|
||||||||
| Monday, 20 March 2006 | Sorrento | Images for 20 March 2006 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Am I tired! I would have liked to sleep in, but people make such a noise in the corridors that I was up at a roughly normal time. Down to breakfast, where we ended up having a discussion about virtual white boards, and Mike Lischke came up with the rather surprising claim that he could do a proof of concept in 8 hours. That sounds optimistic, but if he can bring up something relatively quickly, it would save us an incredible amount of money on face-to-face meetings.
Then spent some time with Monty talking about backup, during which time he tore my design to pieces and put it back in a form which really does give us consistent backup. I still need to digest the concepts, but it looks as if we got more progress on the interface design in 90 minutes than in the previous month.
After that, Monty and most other people left, and I went into town for lunch, bumping into Axel Schwenke on the way. Had quite a good lunch in the Enoteca 2000 in the via S. Cesareo:
|
|||||||||||||||||||||||
Then up to the hotel just in time for Axel to get his bus, and I tried with very limited success to get some sleep. Down again to town in the evening to have dinner with Russell, Shane and also Carol and Jenny:
|
|||||||||||||||||||||||
The restaurant, da Gigino in the via degli Archi, was more expensive but not nearly as good as Enoteca 2000. Had a very dubious saltimbocca.
| Tuesday, 21 March 2006 | Sorrento → Rome → London –> | Images for 21 March 2006 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
The hotel had mislaid a cushion I had brought with me (sent it to Naples with the laundry, in fact), and they had promised to have it back by midday, so spent the morning catching up on my mail, somewhat helped by the fact that they had dropped the network yesterday morning and now charged the princely sum of € 22 per day for access.
Down to the lobby at midday and found Shane, who was coming to Rome with me, along with Russell Dyer, but not my cushion. The receptionists were very apologetic, but that didn't help that we missed our train, and the next one was late enough to endanger our departure. Ended up hiring a car, much to Russell's amusement, and off to Rome.
Sorrento to Rome is about 250 km, but after we hit the Rome freeway ring, it took us almost as long as from Naples to Rome (200 km) to get to the airport (another 50 km), not helped by incorrect signposts that led us through a commercial centre being built in the vicinity. Finally got to the airport and discovered no signs whatsoever for rental car return. Even Firenze was better there. Once we found the building, still couldn't find the return for auto europa: they have signs about 5 cm by 15 cm in size, and we literally had to get out of the car to read them:
|
|||||||||||||||||||||||
The arrow (pointing to Hertz return on the fourth level) is irrelevant. Even on the photo, the auto europe sign is barely legible. Click through twice to the large version and you'll see that we had to turn around and go up another level. In real life, it's only legible if you're within a couple of metres of it. At the fifth level, another sign pointed out over the railing to the outside:
|
|||||||||||||||||||||||
It turned out that we were supposed to go into the Hertz return area and drop the car. Nobody was present to examine the condition of the car. Finally got somebody in the rental office to examine it and confirm that the car was undamaged, but also took my own photos to be on the safe side.
Off to London with a lively discussion with a professor of anthropology from North-Western university in Illinois, who thoroughly debunked my story of beer being the foundation of civilization (“beer and baking didn't start until people were sedentary”). I won't change my story; I'll just add a slide giving expert opinion at the end.
In Heathrow, for once, everything went smoothly. No queues at all at the security checkpoint, and I was in Terminal 4 within 20 minutes of landing (at Terminal 1). Got a helpful staff member to change my boarding pass to an emergency exit row, and noted that they've improved the facilities in the lounge. About the only negative thing was that here, too, they don't have free network access for laptops. Asked a rather disinterested staff member about the matter, and he came up with a nonsensical conjectural explanation qualified by the fact that he admitted not to knowing anything about computers. He suggested that if I were interested enough I could fill out a customer feedback form. He didn't see any reason to report the matter himself; after all, he just works for British Airways. I have a choice, which I usually exercise.
| Wednesday, 22 March 2006 | –> Singapore –> | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Got into Singapore at 18:20, and down to look for SD-RAM cards. One of the things about the new Nikon camera is that it can take “movies” for as long as there is space on the card. The prices have certainly come down; even in the airport, I found a high-speed 1 GB card for SGD 133. That's enough for about 25 minutes of video recording at 320x160/30 fps or 640x480/15 fps, which means that the camera can almost replace a video camera. We'll see in due course how well it measures up to the video camera; at the very least, it'll be useful for short sequences of moving horses.
| Thursday, 23 March 2006 | –> Melbourne → Adelaide → Echunga | Images for 23 March 2006 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Back to Adelaide without event, and for the first time arrived in the new airport terminal. Airport design is no new thing, and I was left wondering why we had to walk so far to find our baggage. Once I got there, found a display showing arrivals:
|
|||||||||||||||||||||||
|
|
|||||||||
Helpful, eh? Who cares about the gate number? Which carousel? Hopefully they'll get their act together, but it seems that the general criticism of the new terminal (which opened for domestic traffic 4 months late) shouldn't have finished yet.
Spent the day trying to catch up with mail, with little success.
In the evening, it was brought home to me that we're not in Italy any more. Our rain water pump stopped pumping. That's not serious—I just needed to power cycle it—but on the way out to the pump found the way blocked:
|
|||||||||||||||||||||||
|
|
|||||||||
| Friday, 24 March 2006 | Echunga | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Another day in catch-up mode. I think I should change the way I work after returning from overseas travel: I'm all awake in the morning, when I usually read my mail, but by the afternoon I wane, and during the time I usually do Real Work I'm just too tired.
| Saturday, 25 March 2006 | Echunga | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
If yesterday was slow, today was glacial. I wasn't feeling the best, and spent most of the day catching up on something I've been sorely neglecting: TV programmes. Also did a bit of investigating where I can get a good kitchen slicer. The Stiftung Warentest in Germany tests lots of things (that's their charter), but they don't seem to have tested slicers in 5 years. The big names seem to be Siemens, Krups and Graef. Graef seems to have the best reputation, though it's difficult to see why from the web site: it's full of obfuscatory flash graphics and seems to deliberately make it difficult ot compare things. Apart from a display occupying about 30% of the window (and thus requires automatic scrolling, which you can't turn off), that shows photos and fantasy names (“Futura”, “Classic” or “Master”), it makes it impossible to compare things side by side, even if you can read the tiny text (about 4 pt):
(to get an idea how illegible this is, consider that this image represents a screen full). If you make it larger, it just breaks the display completely:
That's serious: there's a 2 to 1 price difference between the “Classic” and the “Master”, but they look identical, and the specs hardly vary. It seems that the Germans have particularly broken web sites.
| Sunday, 26 March 2006 | Echunga | Images for 26 March 2006 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Yet another quiet day. Bottled some beer in the morning, somewhat slower due to the presence of a sacred ibis in the garden:
|
|||||||||||||||||||||||
Usually they're very gregarious birds, but this one stayed in the garden alone almost all day, until the dogs chased it away. It must be too big for the cats to be interested.
Also tried—yet again—to install some good video software on a machine. This time it was MythTV under Ubuntu. I barely got off the ground before losing my remaining enthusiasm: installing mplayer didn't work right because of font problems. I've seen this before, but why is the problem still there? We have had “Open source” operating systems and environments for nearly 15 years, and still this kind of nonsense occurs.
| Monday, 27 March 2006 | Echunga | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
According to the calendar, daylight savings time should have ended yesterday, but it was extended by a week at relatively short notice due to the Commonwealth Games in Melbourne It caused chaos: lots of people haven't updated their time zone files, and long-running programs don't re-read the time zone files anyway. The result is that lots of things are off by an hour. And the Commonwealth Games finished yesterday anyway (so we can now get back to normal news). Couldn't they have kept the DST transitions as they were and just put all the times for yesterday's games forward by an hour? That wouldn't have been any hardship, since it would have made no difference to the real times.
The weather noticed, though. It has been particularly hot and dry (temperatures in the low 30s), but today the rains set in with a vengeance, and the temperature dropped to about 18°. Looks like autumn is here.
Still catching up today; Mondays tend to be like that, but apart from mail there were also expenses to be done. There's a lot to be said for a per diem.
There are at least three kinds of mail: stuff which can obviously be deleted, which doesn't really get in the way. I can delete that at the rate of dozens of messages per second. Then there are messages like commit messages which can probably be deleted, but require a few seconds to look at to be sure. Finally, there are “real” messages, which can take up to hours each. I've accumulated too many of the latter, and I really don't know how to speed them up.
| Tuesday, 28 March 2006 | Echunga | |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
Up early this morning for a breakfast seminar on goal setting by Peter Daniels, organized by 2IC, my accounting company. Quite interesting, but it makes it clear how much of an outsider I am in this kind of forum.
The plasterers finally came and repaired the damage to the ceiling that I did while setting up the projection system last August. How long these things can take.
On the work front, finally spent some time looking at bugs. We have some strange case of stack corruption in a threaded environment on Apple Mac OS X, not the most run-of-the-mill kind of bug. Spent most of the day investigating that.
| Wednesday, 29 March 2006 | Echunga | Images for 29 March 2006 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
I've had enough of these damned telemarketeers. I've started a black list. Not that I suppose it will help much.
More work on the Apple problem today. It's not pretty:
There's no mysql-test-run, only mysql-test-run.sh and mysql-test-run.pl.
mysql-test-run.sh doesn't work: the server never starts.
mysql-test-run.pl doesn't tell how to start the server.
../sql/mysqld is a shell script. The real server file, at least in my build, is ../sql/.libs/lt-mysqld.
I started by putting the following parameters into .gdbinit:
Currently, though, that doesn't work:
Further discussion on IRC explained how to use the --gdb parameter, which hadn't been working for me: I needed X access, which at 30,000 km distance is slow. At the speed of light, it takes 0.20 seconds there and back; in fact I get a ping RTT of about 0.38 seconds, so there's not much hope of a real improvement). But I did manage to get an X window up—only with the shell version of the tests (mysql-test-run.sh). That wasn't without a couple of problems: first, the script gets the name of the executable wrong:
In fact, it's a shell script. For reasons I haven't investigated, the real executable is called ../sql/.libs/lt-mysqld.
Next, the tests fail with the shell version of the script; I was only able to get it to run with the perl version, but that version doesn't have this ability to start gdb windows.. Looks like part of solving this problem involves (finally) learning perl.
All this long-distance Apple work reminded me that I had an Apple machine about 40 cm from my right elbow; all it needed was a new power supply. I had bought a couple of replacements on eBay a couple of months ago, but the connectors were difficult to fabricate, so it had been left in the “too hard” basket. Today pulled apart the old power supply, expecting to find a dead electrolytic capacitor. I wasn't disappointed:
|
|||||||||||||||||||||||
That should be easy enough to replace. Taking the power supply apart was quite simple, too. It's one of these “yo-yo” style power supplies, and all you need to do is prise it apart. Well, almost. The mains power connector can be prised apart too, but after that you can throw it away:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
The connector fits in the middle; the pins all came apart when being prised. Obviously the connector should have been de-soldered before taking it apart, but since everything else was just pushed together, it wasn't immediately obvious that this was an exception.
| Thursday, 30 March 2006 | Echunga → Sydney | Images for 30 March 2006 |
| Top of page | ||
| previous day | ||
| next day | ||
| last day |
On with my debugging today, making little progress. There are all sorts of irritating things about the test scripts, including:
At least the way we have it set up, the server process owns SIGINTR, so you can't get the debugger's attention with the intr character (^C).
Processes hang and become unstoppable when in the debugger; I suspect this is a Mac OS problem. I think that you first need to stop the debugger; until then, the dependent process won't even go away after a kill -9. There's no associated WCHAN suggesting that it's waiting on resources.
You can't just restart things in the debugger anyway, because some magic assigns a TCP port, and for whatever reason it remains open after the process has been killed.
All this makes for a really slow debugging process, and by the time I left for the airport I hadn't got past it.
At the airport, flew Virgin Blue, not my favourite airline. Their automatic checkin is improving, though it's not there yet. They have touch screens:
|
|||||||||||||||||||||||
I tried to check in at three different machines, all of which didn't react, before I understood that this particular menu didn't have a touch facility: to scan the bar code, you just hold the itinerary printout bar code in the (undocumented) correct position. Then the touch screen facility is enabled, and it prints out a boarding pass which you have to tear out of the too-narrow slot at the bottom; as a result, the boarding pass comes out crumpled.
In Sydney, found my way to the Pacific International, which I had booked via the web as a four star hotel. It used to be the Great Southern hotel, built in 1940 and apparently only superficially modernized since then:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
Off to look for P.J.O'Briens, where Sam Lawrance had organized a BUGS (BSD User Group Sydney) dinner, to which George Neville-Neil and Bruce Evans also showed up. Bruce lives in Sydney, but it's the first time I have seen him on this side of the Pacific. Sam was late (arrived just before I left), and the pub knew nothing of the reservation, but we managed to work around that. Had a good time, though the noise level of a pub is too high for my liking:
|
|||||||||||||||||||||||
|
|||||||||||||||||||||||
| Friday, 31 March 2006 | Sydney → Echunga | Images for 31 March 2006 |
| Top of page | ||
| previous day |
It turns out that the hotel is just round the corner from the Central Railway Station and thus not far from an area where I've been frequently in the past, notably at the AUUG 2001 conference. Then I had found that there was a good place for breakfast round the corner, so went there again today. Then by train to North Sydney, where we had a barely quorate AUUG board meeting.
At lunch, Steve Landers showed me his new Nokia 770 machine. What is it? PDA? Portable computer? I suppose they'll come up with a name for it, but currently it's somewhere in between. It interested me because it's Linux based and runs X. Of course, it doesn't have a keyboard, and the pop-up keyboard proved to be very badly conceived: it doesn't have a Ctrl key, and you need three levels of menu to generate a control sequence. Not exactly what you need for command line editing.
The meeting finished early, and I was at the airport by 16:00 for an 18:30 flight, which at least gave me the opportunity to finally get through some of my mail backlog.
The flight was delayed: the pilot was due in on a flight from Brisbane, which was delayed due to bad weather. The level of information about the delay was unacceptable (“the pilot will land in 6 minutes; there will be a 15 minute delay in boarding.”). In fact, without further explanation, there was a 60 minute delay in takeoff, and we didn't get any information about arrival time until shortly before landing, with the result that Yvonne had to hang around for nearly an hour kicking her feet and leaving multiple voice mail messages on my mobile phone.
Australia used to have some really stupid aviation rules, including prohibition of the use of mobile phones in a plane at any time: “Ladies and gentlemen, [boys and girls], we have just landed in Adelaide, where the time is eight fifty-two (20:53; why can then never get the time right?). We will be refuelling here, so please do not turn on your mobile phones until well inside the terminal building”. The Civil Aviation Safety Authority has now—mercifully—revised that view, and Qantas now tells you that you can use your mobile phone as soon as you have landed. Virgin Blue still sticks to the old view of the world.
Once out of the plane, things still were pretty rough. Adelaide airport took a long time to open, but it's still pretty painful. As in most modern airports, the car park is opposite the entrance; but there's no pay station when you enter from the terminal; they're way down the end of the car park. What idiot thought of that? Also, there are no signs to the exit, and the obvious direction lands you in a dead end down by the hire car return, with the possibility, however, of driving over the kerb and leaving without paying. And the traffic routing outside is also just plain stupid. This really needs revision (and, I suspect, a few planners need firing).
Even then things weren't over and done with. They're working on resurfacing the South-Eastern freeway, day and night. Somehow they're taking forever, in the process reducing traffic speeds to 25 km/h, when they don't block the freeway altogether. I think I should buy a red flag and walk in front of the car in protest. Traffic laws here still seem to be in the middle ages. They've recently taken to putting up signs saying “Speeding? What's your excuse?”. I think I should have some car stickers made saying “Mediæval traffic laws? What's your excuse?”. Frustrated back home.
Do you have a comment about something I have written? This is a diary, not a “blog”, and there is deliberately no provision for directly adding comments. It's also not a vehicle for third-party content. But I welcome feedback and try to reply to all messages I receive. See the diary overview for more details. If you do send me a message relating to something I have written, please indicate whether you'd prefer me not to mention your name. Otherwise I'll assume that it's OK to do so.
| Top of page | Previous month | Greg's home page | Today's diary entry | Next month | Greg's photos | Copyright information |
