Older blog entries for iDunno (starting at number 61)


Last week I broke my palm by installing something on there that I knew may require a hard reset... wasn't till after I'd done it that I realised I'd not got a recent full backup... and I can't boot windows to get palm desktop to install some of the stuff I want, so sod it, I think I'm going to resort to a clean install using jpilot some time this week (first I need to rebuild my PC though, as I pulled lots of the cards out for a resource conflict that only happens in windows). It's a shame, means that I might not be able to get back on there some of the software that was on the 'windows' cd that came with it... OK - fair enough, it was non-free software, but still, would have been handy to have some of it (Photos springs to mind, but I think I've found something that will do that for me :)

Still looking around for a nice Palm OS 5 Free Software web browser, and am still rapidly coming to the conclusion that one's not been written yet, so, what do we need, a set of libraries for parsing files (html, xhtml, css), a graphics library (jpg, gif, png) and a frontend that can then use the libraries in a sane manner. It'd be nice to have a 'zoom' function on a palm device, too, in my opinion.

What I don't follow is why the only web browser I've found thus far that works is EudoraWeb, and that is text only (sigh).

Anyways - never mind, back to the grindstone, or something.

Played with a new CSS hack, bottom aligning text inside a div. The only solution I found to this (now documented here) was to use a fixed height div and some absolute positioning on the text, contained in another div. If anyone has better solutions to this problem, I'd love to hear them.

Also recently been playing with center alignment of some things, this is apparently easier said than done, as usual MS have not followed the CSS specification and so the easiest way doesn't work, but with an extra line, it does (hooray!). There's quite a few 'hacks' to center things, I'll document them all when I have time.

In other news, I haven't done bugger all on my python rss aggregation script recently, but hope to pick that up some time in the future.

Well, it's been a funny old afternoon, one of the other side of the companies clients have has a Remote Shadow failure, first it was stealing all available CPU, then they restarted it, then it died :/ (fortunately, there's a DR machine for just this case :) So I've been on the phone to them trying to work out what's going wrong, it appears one of the Shadow sets has disappeared in to non-existence and won't recreate because of the following error "Overlapping Partition Error", I'm waiting on a reply from the software manglers to see what we can do, might have to kill the disk, init a new shadow set on it, and then bring the data back across from the DR machine (the data has to come from there anyways, as the DR machine is now live).

Anyways - if all goes badly, there's a trip to london for me, if it can all be sorted over the phone, there's less of a problem, and I don't get to go to london.

That is all.

W00t! Well, went back to my dad's for the weekend, played lots of pool with my brother friday night, spent *far* too much money on beer, but never mind :) Saturday went in to town with dad, went into WHSmith and purchased Finding Nemo, fantastic film :) Sat night was spent mostly at another pub, where I spent very very little, random people bought me drinks most of the night, damned good it was too :)

Sunday, went to the ALUG meeting, much random talking, more beer, all good :)

Monday so far: work, work, work, compile old subversion on a FreeBSD box to recover the archive, now restored in to the new version on that box (rah). Just working out what to do with the rest of the evening, ho hum.

*ARGH* I *WISH* sandisk would make up their minds about the SD WiFi support for PalmOS devices, it appears to disappear at random from their site, it was due last year, then they removed anything relating to it from the website, it was then due again in the first quarter of this year, guess what, it's lost all mention again. This is silly. Very silly. If one *did* come out, and it *did* work in my Palm Tungsten E, I might even be tempted to buy one, and then convert home to wireless notworking, but while they're still umming and ahhing, there's just no real convincing stance on it all.

Right - back to the working thing again.

OK - completely random post having seen yet another BT advert...

"Why are so many people coming back to BT?" well, it's like this here, most of them are after broadband, they can't get ADSL without a bloody BT line (yay for monopolies), NTL doesn't get everywhere, or even anywhere near the coverage that BT has, is it any suprise people are going to BT? They've got a locked userbase.

Right - sorry, just had to get that out of my system, it's been bugging me since seeing the first one of those damned adverts.

Back to the telly and the film.

Coo - something I did appears to be working quite well. Last night we sent an e-mail out to our customers telling them of a new service we had set up called "Cle@nMail", basically this uses SpamAssassin and ClamAV to filter the users e-mail in to 2 seperate boxes, .spam and .non-spam, the bit I wrote is the front end (currently in php, planning to do a rewrite elsewhere in python sometime) which allows the user to log in to the website and view the .spam and .non-spam mailboxes, and use sa-train to train SpamAssassin. Our customers appear to have wanted this for a while, we've setup a few today already, we weren't actually going to launch it till the 1st March, but hey, it's there, it works, now just to get the customer feed back :)

It uses clammailfilter.pl to do the virus scanning, it seems to work quite well (it's a smallish perl script I cobbled together, still needs working on, but doesn't everything?), and it seems to have stopped me from getting any crap :)

All in all, it's not been a bad day, and we're stil getting calls from people after our service :)

On a different note, now that the panicing is over, I'll be playing with my bloggy code on sommitrealweird some more, so that I can (eventually) actually have something nice and usable and light to aggegrate other feeds :)

Right - that's it for now - I'm going back to working on the signup for Cle@nMail, it needs to be in place for monday ;)

Finally fixed my blog to be XHTML compliant using the mx.Tidy python module (thanks for the pointers thom.)

Now, to stop it from downloading the rss feed when it doesn't need to... hrm... I'm hoping I can get away with messing with just checking headers, but I'm not sure yet. Currently it just processes my advogato stream, but I expect to add another stream for less technical stuff RSN (just as soon as I get round to writing something to do it :)

Discovered a problem with the ClamMailFilter tar ball I created yesterday, apparently, when creating these type things on BSD, be careful and use the --portability flag for tar, or you end up with a POSIX tar file that 90% of the world can't use ;)

Right - time for bed, me thinks.

pphaneuf: It might be worth taking a look at PgAccess, which will connect to postgres and give you all the fun of forms, and a front end. Hope that helps.

Released a version of ClamMailFilter last night, it's still not completely finished, but it does the job (for me at any rate). It does appear to sometimes not clean up after itself which I'm currently working on, and should be fixed in the coming week or so.

Currently still working on the spam management system for work, will probably do a complete rewrite in python when that's working, and publish that version instead (current version is php4, which although it works, isn't the most pretty code in the world to work with).

Anyways - back to writing the subscription front end for the spam management so that our customers can sign up at the beginning of next month (nothing like tight deadlines, is there ;). I think that it'll be a mostly manual setup for now, but that shouldn't be too bad.

52 older entries...

New Advogato Features

New HTML Parser: The long-awaited libxml2 based HTML parser code is live. It needs further work but already handles most markup better than the original parser.

Keep up with the latest Advogato features by reading the Advogato status blog.

If you're a C programmer with some spare time, take a look at the mod_virgule project page and help us with one of the tasks on the ToDo list!