4.6 Official Beta Discussions Thread

[Dean Roddey]

[Version 4.6 is released, so this thread is closed and a new one will be created for 4.7]

This thread is for discussions related to 4.6. 4.6 will not be a long release cycle. We just needed to get 4.5 out because it had SO many improvements and features that we couldn't keep holding out, though there was still plenty to do. We will pick that up again now.

Primarily what will happen for 4.6 is:

• More driver conversions. We will continue to convert drivers to V2 compatibility
  
• Auto-gen improvements. The auto-generation system will taken up another notch.
  
• Preliminary work for 5.0, which will be primarily off on the edges and won't affect the stability of the product, and there may not even be any workable bits to play with within the 4.6 time frame.
  
• And of course fix any bugs that got through the 4.5 release cycle.
  

We may even do a 4.7 and 4.8 as we go along, just to get new driver conversions and auto-gen improvements out, as long as the bigger 5.0 work hasn't interfered with the stability of the basics in any way. As long as there's no need to split the code base, which there won't be as long as 5.0 work stays out on the edges, then we can put out new drops along the way without much drama.

[Dean Roddey]

As a sort of ongoing side thing, I've come up with an XML format that describes the syntactical rules for device class interfaces. This will be used in the driver IDE to allow it to automatically validate the fields defined by a driver (which has to be done while it is running of course) to make sure it matches the device classes it claims to implement. Of course it can only check syntax, not semantics, but that will still be a big help to insure compliance.

[Dean Roddey]

Well, I spent the entire day pretty much on exploring why some folks are getting crashes when they use Flash within a web widget. I've tried more things today than I want to remember. I completely replaced the OLE site embedding code with another one, and it still happens.

So I'm starting to wonder whether I'm looking in the right place. The fact that I can put one in a popup, and open and close that popup apparently as many times as I want without issues, but it fails when in an overlay and that overlay is reloaded from 1 to 10'ish times, that is starting to seem suspicious but it's hard to imagine what the real difference is.

But I guess there has to be something going on in the reloading of overlays that isn't good. I'll have to dig into that next. Oh well, I learned a lot of COM'isms in the process I guess. And the new site embedding code seems to be of higher quality as well.

[Dean Roddey]

I'm starting to think that the problem is in the flash code of the weather site. So far only folks having an issue are having it using a flash weather animation. I can set it up to load a Youtube video and reload it as many times as I want. Nor does any other content seem to be an issue.

Something is causing a heap corruption, but it can't be my OLE embedding stuff since any other content can be reloaded over and over without issues. And, even if I set up the embedded window to not auto-load anything when loading up the overlay, and use a button on the overlay to manually load the flash animation, it will pretty quickly cause an issue.

So I'm beginning to wonder if I'm the fall guy here and it's really something wrong with their flash code.

Has anyone else had the web widget cause the IV to fall over on anything else besides a weather.com flash animation?

[daddyd]

Is this the same sort of issue as when I leave the web-widget open with my CCTV cameras on? If I leave it for a bit it (on any IV) the entire system locks up and I need to restart the server that is hosting the cameras. I copied with pride (stole) the code for my camera html page from IVB - it even has his title label on the top of my viewing page.

[Dean Roddey]

No, I don't think that's anything related to this. The server shouldn't really be able to tell if the client viewing its output is us or a regular browser. Does this lockup not happen if you just leave it running in a regular browser?

[daddyd]

No, only when viewing through the IV. Shuts down the computer that is hosting the cameras and 4 of my other devices.

[Dean Roddey]

Do you have the web browser set to auto-refresh perhaps? If so, that would be a big difference from a regular web browser. Perhaps they have some sort of resource leak or something and the ongoing refreshing, which would take a really long time doing it by hand but not too long with an auto-refreshing viewer, could be bringing out that problem that you'd otherwise not likely encounter.

[Dean Roddey]

I got a good ways into a new V2 compatible DSC driver today. It's a bit of a slog since it's a complicated device with a less than optimal control protocol, and I'm redoing it fully to get it up to speed with the other big drivers (Elk, Omni, etc...) in terms of the driver architecture, taking advantage of all of the latest CML capabilities and such.

Initially I'm just going to mostly get the V2 specific stuff in place so that it's usuable with the auto-generation system, plus a few other bits. But I'm not going to necessarily try to replicate all of the functionality of the original driver at this time. I have to move on to other things too quickly.

[dgage]

Dean,

In terms of security capabilities via the driver, is the DSC pretty much on par with the security side of say Elk or HAI? I'm not asking if it is a good security system as I'm pretty sure it is. I'm more asking about the capabilities of the driver and what DSC exposes to CQC that is then consumed. Just wondering how much of an alternative it is to the Elk that I've had my eye on. Thanks.

David