[03:32] *** TheBashar has parted #kamaelia
[09:02] *** Uraeus has joined #kamaelia
[09:16] *** Uraeus has joined #kamaelia
[11:37] *** Uraeus has joined #kamaelia
[13:04] *** TheBashar_web has joined #kamaelia
[15:17] *** TheBashar has joined #kamaelia
[17:02] *** Davbo has joined #kamaelia
[17:04] < TheBashar> Artistic question for anyone listening.
[17:04] < TheBashar> If I want to make a new component that is a little different than an existing one, say a SingleShotHTTPClient with timeout...
[17:05] < TheBashar> Is it better stylstically to achieve the new functionality by subclassing the existing component and hooking to the internals?
[17:05] < TheBashar> Or is it better to encapsulate a mini-kamaelia graphline of other standard components inside a new component?
[17:11] *** TheBashar is now known as TheBashar-afk
[17:15] < Davbo> I'd probably just subclass it if it's not going to cause any problems
[17:26] < Davbo> this might be relevant TheBashar-afk http://code.google.com/p/kamaelia/source/diff?spec=svn6067&r=5266&format=side&path=/branches/private_DK_PaintDev/Kamaelia/Kamaelia/UI/Pygame/Button.py&old_path=/branches/private_DK_PaintDev/Kamaelia/Kamaelia/UI/Pygame/Button.py&old=5058
[17:26] < Davbo> that's a diff of me adding an "ImageButton" to the already existing pygame button class
[17:27] < Davbo> just a button represented by an image rather than a text caption
[17:28] < Davbo> what you're describing sounds more complex so probably different :)
[17:43] *** TheBashar-afk is now known as TheBashar
[17:46] < TheBashar> Davbo: Thank you for the example. I'm trying to get a feel for how far down to scale using the postoffice metaphor when making components. This is a good example.
[17:51] < Davbo> no problem
[18:37] *** vmlemon_ has joined #kamaelia
[18:46] < vmlemon_> Hi
[18:47] < Davbo> Hey vmlemon_
[18:52] *** vmlemon_ is installing Cygwin on Windows 7 at the moment...
[18:52] < vmlemon_> (I ended up installing the beta on my crappy old desktop box, after someone sent me a copy on DVD)
[18:55] < vmlemon_> Surprisingly, it's not that slow, even though I only get a 1.0 on the "performance rating" thing, since there isn't a driver for the ATI Radeon 9000 card in that machine
[18:55] < vmlemon_> The other stats are around 3.5-5.0, if I remember correctly
[18:56] < Davbo> yeah it just tells you the lowest score from the components doesn't it?
[18:56] < Davbo> what do you think?
[18:57] < Davbo> screenshots look like KDE4
[18:57] < vmlemon_> I agree with the looks like KDE4 sentiment
[18:57] < Davbo> not a bad thing, KDE4 is awesome
[18:57] < vmlemon_> Although others disagreed, when I showed them the screenshots from Wikipedia side-by-side
[18:58] < vmlemon_> The taskbar screams "Stolen from KDE 4" ;)
[18:58] < Davbo> hehe
[18:58] < vmlemon_> Not to mention the boxes around desktop icons, when you hover over them
[18:59] < Davbo> hey, we make open source software the way we want it to be. It's a stupid idea to ignore what the users want
[18:59] < Davbo> ;)
[18:59] < vmlemon_> Heh
[18:59] < Davbo> I wonder if the KDE team scared Microsoft with the KDE on windows stuff :)
[19:00] < Davbo> "No, they're bringing KDE4 to windows!! What do we do?!?! Lets beat them to it!"
[19:01] *** Davbo isn't very happy with Microsoft recently.
[19:02] < Davbo> University project is making a "Workload Manager" for Engineering department, their current system is a massive excel spreadsheet :(
[19:02] < vmlemon_> Urgh
[19:03] < Davbo> Using xlrd to read it into nice Python structures then using sqlalchemy to get it into a database (where it should have been in the first place)
[19:04] *** vmlemon_ sees that Microsoft's "support" for UDF on non-optical volumes in Windows Vista and Windows 7 really isn't (it's a bastardised version that isn't supported by other OSes natively, until they get around to reverse-engineering the "enhancements", and W7/WV won't allow use of UDF volumes created on non WV/W7 OSes, no matter what version is used)
[19:05] < Davbo> MS are good at doing stuff like that
[19:06] < vmlemon_> I spent about 4 hours trying to mount a UDF-formatted test partition on a USB disk, and finally came to that conclusion
[19:06] < vmlemon_> (The UDF volume was created with mkudffs on Linux)
[19:07] < vmlemon_> "Microsoft: We Support (Abuse of ) Open Standards" ;)
[19:14] < Davbo> hehe
[19:23] < vmlemon_> I can hear the chainsaws revving up, to cut down the trees to make coffins for Wine and ReactOS, after seeing a new clause in the W7 EULA stating that components from it can't be used outside of the product to run software intended for Windows, or words to that effect
[19:23] < vmlemon_> *intended for Windows on non-Windows platforms
[19:26] < vmlemon_> Still, I'm surprised to see that they've managed to produce a working NFS client
[19:26] < vmlemon_> http://picasaweb.google.com/tyson.key/Screenshots#5305748874342946994 - screenshot of the EULA dialogue, showing the said clause
[19:27] < vmlemon_> "You may not: * use components of the software to run applications not running on the software"
[19:46] < vmlemon_> Hah, "No. A quick search at the United States Patent and Trademark Office for "XWin" turns up one dead record and one live record. The live record is for a logo belonging to a rock crusher manufacturer based out of Belgium. Neither trademark affects Cygwin/X, as the dead record is no longer enforceable, while the live record is in an unrelated and distinct industry; there cannot be confusion between rock crushers and computer programs."
[21:09] < TheBashar> Anybody know if there is a way to hard kill a component without killing the scheduler?
[21:09] < TheBashar> I have an HTTPServer component that sometimes does not exit cleanly.
[21:10] < TheBashar> I wrapped in with a component with a time-bomb that sends a shutdownMicroprocess() if it doesn't exit after a time period.
[21:10] < TheBashar> But this still does not make it quit.
[21:11] < TheBashar> Actually, the timebomb sends it shutdownMicroprocess and serverShutdown. Neither work when it gets stuck.
[21:11] < TheBashar> Can I somehow get the scheduler to kill it?
[21:23] < vmlemon_> Hmm, not sure :(
[21:23] < vmlemon_> You could always wait and see if mhrd or MS- come here, tomorrow
[21:23] < vmlemon_> (They're the oracles on Kamaelia, since they built it) ;)
[21:24] < vmlemon_> *built most of it
[21:24] < TheBashar> Thanks, I might have to if I want to figure out why the httpserver is getting stuck.
[21:25] < TheBashar> But I'm going to keep digging to find the right shovel to clobber the stuck microprocess with.
[21:25] < vmlemon_> There was a recent drive to get a relatively coherent method of shutting systems down implemented, although I don't know what became of it
[21:26] < vmlemon_> There are race conditions and other things to worry about, if I vaguely remember correctly
[21:26] < TheBashar> I'm all for standardizing on shutdownMicroprocess for normal shutdowns.
[21:27] < TheBashar> But I guess we need a standard clobber shutdown as well.
[21:36] *** vmlemon_ has parted #kamaelia
[22:06] < Davbo> TheBashar, I'm not 100% but this might help http://groups.google.co.uk/group/kamaelia/browse_thread/thread/56e69d5f1cc7d43f/
[22:06] < Davbo> a few techniques for shutdown are discussed there
[22:08] < TheBashar> Thanks Davbo. Some good techniques.
[22:10] < Davbo> also; http://code.google.com/p/kamaelia/source/browse/#svn/branches/private_JMB_HTTPSupport might be a relevant branch for you
[22:10] < Davbo> http://groups.google.co.uk/group/kamaelia/browse_thread/thread/8d6f049ab9402e43
[22:10] < Davbo> has the info for that branch
[22:11] < Davbo> Kamaelia/Protocol/HTTP/HTTPRequestHandler.py: Added case- sensitivity handling and fixed issue with deadlocking on component shutdown.
[22:11] < TheBashar> Oooohh.... now that does sound good.
[22:12] < Davbo> hopefully that might help
[22:13] < TheBashar> Thanks a bunch!
[22:13] < Davbo> no problem
[23:21] *** Davbo wishes his Beechams cold relief tablets were not "non-drowsy"
[23:22] < Davbo> finally got my excel file parsed, well part of one sheet
[23:22] < Davbo> the "main" bit
[23:52] *** Davbo has joined #kamaelia