Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
Dean's Crispy Bloggette (avec Fromage)
#31
I've snuck in a couple of small extra things, very peripheral so they aren't dangerous. This late in the game I had to test them extra carefully so they were time consuming relative to what they provide.

- I added string list field I/O in the CML simple field I/O class, and fixed the writing of time fields which was messed up.
- Got the remote IP address into more of the server side ORB error messages, so that when he complains about somebody, we know which client that is.
- And got the host name back into the successful test connection to the master server in the installer.
- I got the VSX-84T, Concerto, and Z-Wave driver docs updated (though those won't show up until the release and the upload of the new web site.)
- I put up an 'advanced installation 1' tutorial video
- Got my local copy of the web site updated in preperation for the release.


I'm going to give it a couple more days to make sure no problems crop up, and get some more feedback, then I think that we can declare this one baked and put it out there. I'd encourage a few more old timers to give this one a try since it seems to be doing well and I'd like to get as much feedback on it as possible. In the meantime I'm just going to work on some tests and do some investigation of some things that I need to look into for future purposes.
Dean Roddey
Software Geek Extraordinaire
Reply
#32
A couple folks have had troubles in the new beta in their kiosk systems, which auto-start. The new system starts up fast enough now that there can be some issues with the IV trying to start too soon. So you may need to use the /Delay=xxx command line parameter on the IV to delay its startup a bit, giving the OS time to get up, get it's DHCP info, get all the services started up, let CQC's drivers get connected to their devices, and so forth.

The 'xxx' is the number of milliseconds. So you can do something like /Delay=5000 to have the IV delay for 5 seconds before it starts trying to talk to background services and such.

You may not need it, but if you find in the new beta that the IV is having problems during startup now, it may just be getting up before everything is in place.
Dean Roddey
Software Geek Extraordinaire
Reply
#33
Another thing that I forgot to mention that was added is that you can now set, for remote access or for a wireless client on a slow/bad connection, the following environmental variable:

CID_ORBTIMEOUTADJUST=xxxx

xxxx is a value in milliseconds. This is added to all of the timeouts in the ORB (network access stuff) remote calls. So what I've done is that, instead of just cranking up the timeouts to support such clients, I've set the timeouts to something appropriate for a local network, and provided this variable to allow you to make up for the slower access from remote machines or slower wireless clients.

So something like:

CID_ORBTIMEOUTADJUST=3000

Would add three extra seconds to the timeouts (which are normally five seconds.)

This is not the whole time that a client will wait for a server to respond to something, but the time that it will wait, once a packet header has arrived, for the subsequent packet data to show up. As rule, five seconds is way more than is required on the local LAN, and is just there to handle the occasional disruption of a server by some other high priority process or something like that, or a pathological spike in simultaneous access to a server and so forth.

But for a remote machine or a wireless that has a pretty bad connection, you might need to crank it up a bit. Of course if the connection is really bad, this will just make things worse. This is really intended to help deal with sporadic access problems, so that it doesn't timeout. If it's really taking more than 5 seconds to get most or ever packet sent, the machine is already going to be unusuable as a CQC client really.

For truely remote access, over a lower bandwidth connection like an ISDN line, you may need to bump it so that when it hits a bigger packet like an image, that it doesn't timeout. In that case, the smaller packets will probably be just fine at 5 seconds (though slow of course) and the line should be reliable and not have dropouts and such, but a bigger background image might take more than the available 5 seconds to get across.
Dean Roddey
Software Geek Extraordinaire
Reply
#34
Doh! That /Delay= parameter above takes seconds, not milliseconds. So 5000 would certainly give your machine plenty of time to get up and running, and for you to go lunch probably. Sorry about that.
Dean Roddey
Software Geek Extraordinaire
Reply
#35
I've started poking around with the Sony XLB1 changer tonight. I got it hooked up and pulled the powerfil.sys driver from Vista and put it onto my XP development machine. When I power it on, it get's identified as a PowerFile CD200, which is effectively what it is, just rebadged, so I think that's OK. It comes up and takes a drive letter and shows up in the device manager and drive list. I can't really access any data yet since no disk mounts by default. I'll have to get enough control over it first to get that to happen.

It's done a few wierd things so far and I'm not sure if it's me or what. It seems to go off into the ozone sometimes and any access to the drive just hangs up (which would be very bad in a working system.)

I've looked at the IOCTL commands for it and they are fairly opaque to me at this point, since there's no real description of the appropriate way to use them. I'm going to have to do some serious Googling probably. Actually I've done a good bit already but have only found a little bit of info, and none of it related to coding a driver for these IOCTL commands.
Dean Roddey
Software Geek Extraordinaire
Reply
#36
Well I've earned my salary tonight. Hardcore exploration. My brain is about to pop and I didn't really learn all that much that I needed to know, though I learned a lot of other things along the way that might come in handy later. I'll get back into it tomorrow. I'm just working out tonight out to enumerate all the Firewire attached devices and find the ones that are of the type that are compatible with the Powerfile driver. After a lot of poking around, the obvious answer became apparent, though I'm not sure if it's the ultimate-ultimate answer, but it's a more than good enough one for now.

After that I have to figure out how to actually make the thing do something.
Dean Roddey
Software Geek Extraordinaire
Reply
#37
Dean Roddey Wrote:After a lot of poking around, the obvious answer became apparent, though I'm not sure if it's the ultimate-ultimate answer, but it's a more than good enough one for now.
You didn't get to 42 yet, then?

Chris.
(towel day is may 25th)
Reply
#38
I'm testing out a 2.0.17 release now. Since very little has changed but the addition of some drivers and a few small tweaks on the ORB stuff, I don't foresee any problems. If not, I'll get it posted later today and let people move up to this version. Unless someone has a problem (unlikely), it'll become the official release and you won't have to upgrade again once it goes golden.

By my count, including Rohan's offline C-Bus driver, we are going to be at 134 drivers for the upcoming release.
Dean Roddey
Software Geek Extraordinaire
Reply
#39
I spent the day on documentation today, and spent some time in the studio. I noticed that I'd never updated the Zoom or TheaterTek drivers to reflect the 2.0 media scheme, so I got those updated. And I finished writing up the new Z-Wave driver documentation page, trying to provide a lot more guidance and discussion than before, to insure people understand all the issues, and documented more of the new drivers and changed drivers.

On the music front, I discovered that playing drums on the keyboard just wasn't going to cut it. It doesn't work. So I got this thing called a Trigger Finger, a pretty common type of device these days. It's just 16 small velocity sensitive pads on a tablet PC sized thing, each of which sends out a MIDI signal that can trigger a different drum in a drum synth. So it's a little tiny electronic drum kit.

It takes some practice to use and hit it squarely in the middle of the pad with a precise downward motion. So I've been putting in some time practicing on it each night and, just to see if I could do something fairly realistic now, I decided to do another little part of a song. I decided to have fun and do the most bombastic, Arena Rock slash Hair Band slash Power Pop type of thing possible, since that would give me a chance to really play some 'power drums'. Here is the result:

http://www.charmedquark.com/Web2/TmpAudio/BigHair4.wma

Be careful, because it will make you want to run out and get a perm and buy some spandex. I've not put a lead guitar part on it yet, but it's so bombastic as is, I figured I'd go ahead and post it. I'll try to get a lead part on it tomorrow night or the next.
Dean Roddey
Software Geek Extraordinaire
Reply
#40
I'm looking into what is required to deal get the WMP engine wrappered up so that we can use it for both video and audio playback internally within the product. As usual with this stuff, it unfortunately involves COM, which I hate with a passion, and the going is slow. I put in a mind warping day of exploration today and made some progress, but mostly it was just spent wanting to find whoever invented COM and beat him up. And everything is so woefully underdocumented in this area as well. Horrible, horrible, horrible. I hate it, I loath it, I despise it.

But, other than that, everything is going ok.
Dean Roddey
Software Geek Extraordinaire
Reply


Forum Jump:


Users browsing this thread: 2 Guest(s)