4.5 Official Beta Release Thread

[Dean Roddey]

Version 4.4.929 is posted. This is really just for driver writers, though you may need to upgrade to use their drivers after they have reacted this change. It will be only a handful of drivers at worst though, since it's related to V2 drivers and there are hardly any third party V2 drivers done yet.

After you upgrade to or beyond this version, if you are using any V2 drivers of the multi-zone audio, projector, TV, A/V processor, etc... types, those that allow selecting of a source input, of which there are currently only three or four really, you will have to make a slight change in one of the fields. The affected drivers that I know of are Marantz 8801, Monoprice 5704, Monoprice MZ6, the multi-zone audio device simulator, and the Denon X4000.

Sorry for any inconvenience, but it's hard to foresee these things until you start getting into them, and this is such a useful capability that it was worth the change, now while the number of drivers affected is tiny.

If you are doing a V2 third party driver, be sure to update to conform to this change. Again, sorry for any inconvenience, but long term it will be well worth it.

See the change list for the details but basically, we got rid of the matrix switcher device class and updated the Switcher class to be the sole switching interface. We got rid of the SwitchedOutput semantic field type and now only used the SourceInp type. So the fields in the drivers above that handled the source will be slightly different now, e.g. instead of MTRX#Z1~Output, it would be SWTCH#Z1~Source, and similar types of changes.

• Do a V2 version of the Monoprice 5704. It's not been documented yet, so might as well update it and document it in V2 mode from the start.
  
  
• For 4.5, the default 'do fades' setting for fading popups in and out should just be set to True, so that it is enabled by default. The number of folks who would not want them is far smaller now than those who would, and new customers should see them by default so that the IV looks as modern as possible for first impressions.
  
  
• The fade in/our of popups is awfully fast now, so slow it down a tad. It was changed to be throttled by the vertical blanking period of the monitor, which also makes it nice and smooth as well. But that left it so fast that it's hard to tell the difference between fading and not fading.
  
  
• It become obvious that it had been a mistake to allow device classes like AVProc, TV, Projector, MZAudio, etc... to have their own source input fields. They should have instead implemented the Switcher class. Doing so would allow us to control switching generically for any type of device, a very powerful capability and the type of thing that all of this device class stuff was started in order to allow for. This in turn lead to the realization that having separate matrix switcher and switcher device classes was equally at odds with this goal. So, the matrix switcher class was tossed, and Switcher is what is used now in all cases. Where multiple outputs are available, the sub-unit prefix scheme is used. The device classes that previously had their own source input field were updated to get rid of them. The SwitchedOutput semantic type was removed, and SourceInp is used in that case now.
  
  Too bad this wasn't foreseen, but only hind sight is 20/20. But we needed to change it and it had to happen before 4.5 went out, or it would have been very difficult to do afterwards. Any shipped drivers have been updated to comply with this change. Probably there aren't many or any third party V2 drivers yet that would be affected, since there aren't many at all yet. But, if so, you will need to change your drivers to comply.
  

[Dean Roddey]

I updated the first post with the latest cut at the 4.5 external documentation.

[Dean Roddey]

Version 4.4.930 is posted. It just updates a few more simple drivers updated to V2 compliance (which I could safely do by eye since they are simple ones.) And it contains a couple fixes.

• Update the Carls Jr. digital I/O box to support V2 mode, in addition to the original mode. It supports the Relay and DIO classes.
  
  
• Also, in the Connect method, the Carls Jr drivers gets the version string but doesn't check the returned length, and then subtracts from that length to get to a substring. But if the length is less than that (such as zero if the box is not there), it underflows and tries to allocate a huge buffer. That would cause a non-timeout exeption to be thrown, which would cause the driver to think it needed to panic and cycle the com resource, so it would continually cycle between wait comm res and wait connect.
  
  
• Update the Octava HDMI switcher to support V2 mode, in addition to the original mode. It supports the Power and Switcher classes.
  
  
• The Intf. Designer's grid display and configuration stuff is sort of a mess. It doesn't display correctly if you change the grid side while it's displayed. The show grid menu item doesn't work quite right, etc... Get it fixed up.
  
  
• Randomly ran across some info relevant to the embedded browser, and made some changes that appear to have improved things wrt to some known problematic situations. We'll see if it's just me being opimistic.
  

[Dean Roddey]

Version 4.4.931 is posted. It's some updates and fixes for the new V2 Elk driver and some more improvements on the Web Browser Widget.

• The ZoneStatus semantic field type needs to be expanded to include an Unknown value, since that is one of the values of the standard CML zone status enum, and it's a reasonable field value since sometimes it's not in any of the standard three states. However the driver needs to ignore calls to the zone alarm trigger sending helper method with this state. It's not a valid zone alarm trigger value. This will prevent any breakage, and allow the useful unknown value in the field itself, but avoid allowing it into triggers.
  
  
• The new V2 Elk driver wasn't updating the 'in alarm' field for areas, so it would never go true. It was remembering the state internally, but the field wasn't being updated.
  
  
• The new V2 Elk driver wasn't storing the correct value for the area 'arm up' field. It was using the arm mode value to set the field, so it would show a different (incorrect) state depending on the up arm mode value received.
  
  
• In the new V2 Elk driver, fix the long outstanding issue with the zone status. The logical status is now the V2 specific field, but the LogStatus really should be PhysStatus, since it is the physical status. The V1 driver will still be the original way. Sorry for the iconvenience for any folks who have already started using this V2 driver, but that's the Beta thing. I really didn't want this new driver to go out the door with that ancient error still intact.
  
  
• Get some more of the old supported InvokeCmd commands into the new V2 Elk driver, and in the docs for it.
  
  
• Update the Denon X4000 driver to support zone 2 and 3 so that it's a more full featured driver.
  
  
• More improvements on the Web Browser Widget, wrt to potential instability it can cause when displaying some script driven Flash content. Not sure completely this fixes it, we'll see. It seems to have taken care of it for me. Widgets are given a change to 'pre-clean up' before a popup is closed. The redrawing that happens at the end of the popup being closed, and particular the fade out, seems to freak it out somehow. The web widget takes advantage of this to destroy the web browser instance before the cleanup happens.
  

[Dean Roddey]

I updated the 4.5 'external content' web site stuff (drivers, CML, etc...) with all the latest changes, see the first post for the link.

[Dean Roddey]

Version 4.4.932 is posted, which just as a few small fixes and driver updates.

• A new output variable parameter was added to some of the Event target, the ones that let you query out info about an incoming event trigger. These let you get a newly available name value, so that you can get the basic name in addition to having the field name. It is optional, so it can be empty. However, the upgrade that automatically updates those commands to have the new parameter, did make the new value empty, it has the word "Name" in it, so it will be non-empty and an illegal variable command. You can just delete that content, but fix the upgrade to henceforth set it to an empty value when updating existing commands.
  
  
• Update the Yamaha VX1900 A/V receiver driver to support V2 mode.
  
  
• Somehow JSON support never got added to our web server, so add that, since it's trivial.
  
  
• The Web Image widget shouldn't log errors if it fails to download an image, unless the IV is running in verbose logging mode. It won't be that uncommon for failures to occur with servers being restarted, occasional network glitches, etc... You wouldn't want those errors polluting the logs.
  
  
• Get the Barix ExStreamer media renderer driver going again and supporting V2 mode.
  

[Dean Roddey]

I have created a new Media Guide technical document. It'll be on the new 4.5 web site in the usual Learn -> Technical Docs location. In the meantime, here it is:

http://www.charmedquark.com/Web2/Downloa...aGuide.pdf

Let me know if this is helpful, or fails to cover something that should be covered, etc... A lot of folks have trouble in this area, so hopefully it will help.

[Mark Stega]

Nice & sorely needed- I've done a skim read, detail to follow. The ToC is messed up...

[Dean Roddey]

Yeh, I noticed that after I posted it. I fixed it for the next upload.

[Dean Roddey]

Version 4.4.933 is posted, which just adds a few more small fixes that have been reported.

• There was a goober in the media engine where setting the art path for large art would also set it for small art, which could cause wierdness down the line, and considerable overhead if it caused the large art to be returned when thumbs were requested.
  
  
• Looks like My Movies 5.0 has changed their cover art storage scheme. They don't seem to be using the directory they use to use. They also though store art in the media directories, wherever you point the per-disc paths to, so update the driver to also look there if it doesn't find it in the previous location. Hopefully that'll insure it works in both cases.
  
  
• The new optional parameters to the trigger event action targets that let you get load, motion and zone trigger info need to be tolerant of the new name value not being present, because it won't be until the drivers are all updated. It allowed the new parameter (output var name) to be optional (empty), but if you provided it, it was assuming the value to put into it was going to be there.
  
  
• The conversion of the lighting simulator driver to V2 compliance somehow lost it's sending of load change event triggers.
  
  
• There's a bug in the event class' parsing of incoming event triggers if one of the values is empty. The name of the value ends up being stored incorrectly.
  
  
• Make the IV engine's cover art cache half the number of slots it was before, so that, worst case, it can't eat up so much memory. It'll still be large enough to prevent much downloading in most systems. Browsing via large art is going to make the usage worse by a considerable margin, so it's not recommended.