Blog
Page: 0 ... 5 ... 10 ... 15 ... 20 ... 25 30 31 32 33 34 35 36 37 38 ... 40 ... 45 ... 50 ... 55 ... 60 ... 65 ... 70 ... 75
Phase 2 In The Grand Green Earth Plan
Date: 4/9/2007
After a successful Phase 1 of the Grand Green Earth Plan for my house hold. Phase 2 went into effect last week with the purchase of green(er) family transport:



We are with you.

For those who care about numbers, and it does all come down to numbers, here they are: It's not as good as I was hoping but when budget, size, drivability, features and safety collide it's always going to get messy. I really like that we have 4 air bags now as well as the improved economy. I've seen a few trucks run reds and the messy aftermath... and I for one want a little bit more between me, my family and the truck's bull bar.

Phase 3 perhaps...?
(0) Comments | Add Comment

Scribe v1.89 Stable
Date: 23/8/2007
I've made a branch in svn for v1.89 stable based on test20. Apart from one user I've had no reports of major issues. I'll update the translations one last time and release v1.89 stable.

So in the trunk code I'm trying to move towards a cleaner storage architecture so that I can incorporate a new file system based mail/contacts storage. At the moment my idea is that email gets downloaded to a single file (.eml) and then at some point later gets "compacted" into a blob of mails. Which is typically maybe a MB of email in one file. This means you can delete an email and reclaim the disk space used without having to copy more than a MB of data. The mail store directory would mostly be 10's or 100's of "blob" files where all the email reside, some .eml files that havn't been packed into a blob yet and some indexes into the whole thing for fast searching and so on. It's kind of a mix of the MBOX and MDIR ideas, with I think the benefits of both.

The current storage API is not going to map onto that easily so I'm trying to work out the route of least resistance to get that working. I still havn't figured out how the indexing will work. I'd love to get rid of the flaky mime parser the current system uses and start using the new parser for both reading and writing email. But I might leave that so that I don't have too much going on in the code.

That said, I still have trouble with my RSI these days and can't type much. It's not that I don't want to work on the code, but more of a case that I can't do much without endangering my health. I wish I could move faster on these issues but I can't. Sorry.
(6) Comments | Add Comment

Wanted To Buy
Date: 15/8/2007
This is more so that Google indexes it than anything else so here goes. One day I'd like to get my hands on the following guitar amplifiers:

Marshall JMP800 2204 Head
Early 80's model, in good condition, 50w head, unmodified a bonus.

Mesa Boogie Single Rectifier
'90s model series 1, 50w head.


I'd prefer a local sale (Sydney, Australia) but shipping from somewhere in Australia is a possibility. The amps must be local voltage units (240v) not US voltage (110v). They don't need to have new tubes but must be in working order. Please let me know if you are interested in selling such an amp.
(0) Comments | Add Comment

Teh new Apple gear.
Date: 8/8/2007
802.11n AND gigabit!
After ranting about the lack of gigabit wired access on the Apple Airport Extreme, they have finally updated the device to include gigabit ethernet. Oh sweet joy :D

iLife08/iWork08
Oh what have we here, the next rev of iLife and iWork! We've only been waiting for what 8 months now? Finally.

Must... resist... clicking... buy now... ;)
(0) Comments | Add Comment

Don't Lose Data
Date: 7/8/2007
I read Jeff Atwood's post about What's Worse Than Crashing? this morning and I whole heartedly agree on his assessment. However I have not been as diligent as him, in that my software has caused user's to lose data (and I'm very sorry about that if it was you!), however it's getting better, in fact I released Scribe v1.89-test20 this morning to try and prevent users from corrupting their settings file (yay... go me *cringe*).

Anyway I opened up my phone to look for some info I had stored in the calendar. And recent events were in there but as I scrolled back through the months there was no old events at all. It seems all the events from Jun 2007 back were gone. Typical I thought. It's crashed or something and lost some of my data. (The phone has well and truly lost any credibility with me). Anyway so I start googling around for info on the problem hoping to bump into some other frustrated users and maybe get a solution... and got pretty much nothing. So I noodled around for a firmware update. And idly ran through the settings on the phone looking for anything that might shed some light on it. Then I saw it... in the Calendar setup there is an option called "Delete After:" and the value was "4 weeks".

A moment of stunned silence followed. Crickets chirped. Huh? The default mode of the phone is to f*** with the users data? Niiiiice. Good one Motorola... thats just sweet, you really nailed that setting. Oh of course there is a "Never" option, which perchance is what the user would expect to be the default, but thats not the default is it? Nooooo. Unbelievable.... unbelievable.

The default setting!

Unbelievable... *sigh*
(2) Comments | Add Comment

Scribe/Mac Lack Of Progress
Date: 18/7/2007
Since the recent Win32 releases of Scribe I've been working hard on the Mac port in an attempt to make a similar release. Currently there are a number of really nasty issues in the codebase that I'm struggling to fix. The sort of things that make using the software really annoying. Btw I only have access to my Mac at the moment. So PC dev work is not happening till next week.

The first major issue I've got is that the PostEventToQueue function doesn't always deliver a message. Sometimes it silently fails. This causes the message download code to never get the mail from the worker thread and then the worker thread hangs waiting for the GUI thread to process the downloaded email. It's so dumb. I really hope it's not a bug in Carbon.

Then there is the locking issue, where somewhere in my code I lock the semaphore controlling the options structure and then fail to unlock it at which point things go downhill. I'm slowly getting closer to finding the cause of the issue but it's hard to reproduce so I just have to wait for it to happen... look at the logging then re-instrument the code and re-run it. Wait for it to happen (hours/days?) and then repeat. It's slow and annoying.

So in the name of quality I can't release a Mac build just yet. :(
(0) Comments | Add Comment