Official 5.3 Release Thread

[Dean Roddey]

5.2.918

• Update the Schlage BE469 device info to reflect the notification ids from traces we got, so that it will see the status changes correctly (hopefully.)
  
• The publish/subscribe stuff introduced an issue in the attribute editor when you invoke some of the editing dialogs for the more complex attributes. If we made a copy of a widget, it would still publish change reports, but it's no longer part of the template, it's a copy of a template member. The template would get an error because it was getting info from a widget it doesn't know about.
  
• The publish/subscribe stuff has a fundamental issue in that it only supports async delivery. But for certain types of things, where one thing is being kept in sync with another, this is an issue since more than once change can occur in the source before the target sees the msg. So we need to rework to make it synchronous, then we can build the old async scheme on top of that, so that we can support both. Update a number of subscribers to use the sync mechanism where appropriate. Others can stay with the async, which is fine for notifications and commands. This is quite a bit of work.
  
• When you do triggered events in the IV, the event action target isn't being added to the target list, so you get an error if you try to use it to extract event info. That is now fixed.
  
• The RCS driver is getting an error when it tries to parse the incoming thermo names prompt because of a failure to adjust for the fact that our list is one larger than the number of actual thermo slots (so that we can use the 1 based indices of the RCS protocol directly.)
  
• Our Z-Wave driver disables controllers automatically because we don't need to deal with them as units that the user sees. But, it's going by the basic type and is including (and hence disabling) scene controllers (and possibly other things that we don't support yet but might later.) So we need to use the generic/specific types for this instead, to pick out specifically those that are master or static controllers or installer tools and such.
  
• Somewhere along the line during the 5.x transition, the Intf Editor's use of the status bar changed from showing info for the widget the mouse is over to showing info when there is a single widget selected. The former is much more useful so change back to that. The above pub/sub changes makes that fairly straightforward to do now.
  
• Add widget area to the widget palette. The above change means you can't see info for covered up widget. You can drill down and select it, but the mouse over doesn't understand that difference. So the widget palette becomes the means to see basic info on such covered up widgets.
  

[Dean Roddey]

5.2.919

• The update of the Schlage BE469 device info in the last update to get the notification ids in place (so it would recognize lock/unlock events from the lock itself) got them backwards (locked/unlocked were flipped) so flip them back the other way.
  
• Try another approach with WebRIVA for scaling. Get rid of nopixelscale and replace it with vpscale=x, where x is 1, 2, 1.5, ect... So lets the user set the scaling they want. Defaults to 1.0 if not set. Also get rid of the other options which could prevent user scaling via zoom (in theory) so it should now allow for zooming.
  

[Dean Roddey]

5.2.920

• Make the ongoing activity timeouts in the new Z-Wave driver progressive, so that they start fairly quick so that we can get going quickly in the face of perhaps a glitch or two. Then they get longer so that we can recover reasonably quickly from a temporary outage. Then they get longer to avoid holding up other stuff while these unresponsive units fail to respond. Make them get longer quicker for frequent listeners. Do a lot other work related to this to try to support this goal.
  
• The RA2 driver's thermostat high/low set point fields got their range limits reversed, so high is using low's and vice versa.
  
• The RA2 driver isn't taking into account that it can get floating point values for temps.
  

[Dean Roddey]

5.2.921

• The Lutron RA2/Caseta/HWQS driver's thermo support wasn't dealing with the possibility of floating point values, which in theory could happen. And it had transposed the high/low SP limit ranges. And it wasn't ignoring the reporting of temp/SP values for the other temp range, and some send both. That could cause out of range errors.
  
• In the triggered event filter editor, when you invoke the popup menu for an entry field and select to insert a device moniker, it actually invokes the field selection dialog instead and inserts the field name instead of the device moniker.
  
• When I changed all of the interface editor stuff to use the new publish/subscribe mechanism for updating displayed widget info I forgot to have the attribute editor subscribe to widget changes so that it can update the area when you move a widget directly (keyboard or mouse.)
  
• When an exclude is done on the new Z-Wave driver it doesn't clear the unit list or the unit state descriptive text below the list (if one was selected.)
  
• The Z-Wave driver should try to give the master controller and Z-Stick controller names automatically. They end up disabled automatically and there's no way to name them since they aren't types that we interact with. It would be nice to have them get immediately recognizable names in the unit list.
  
• The Admin Intf isn't doing the correct sequence of events to restore it's previously stored position. So if you have a tool bar docked on the left or top, every time you run it it will creep in that direction by the size of the docked window. For bottom/right toolbar placement it didn't matter.
  
• If you run the AI and the client service isn't running, it gives a fairly cryptic error that doesn't tell you what you might want to do to fix that issue. Now it tells you probably the client service isn't running.
  

[Dean Roddey]

5.2.922

• The old legacy Lutron Homeworks driver needs to allow for a zero in an item address of it's in the 3rd slot of a dimmer. It's considering any zero component an error now and lot loading the config. Also, allow it to 'fix up' any bad addresses on load, and just tell the user it did so, so that it can at least load the config. Worst case, allow it to skip a completely bad item, and just warn about it. Update the in place editing to enforce these rules.
  

[Dean Roddey]

Version 5.2.923 is posted. This is the first formal release candidate. Honestly it's highly likely to be the actual release since it's quite tight by now. But, in the worst case scenario if someone found something that was both quite annoying and safe to fix I could make some more small changes. But most likely this is it. It would be nice to get some folks on this guy in daily usage just to verify that there aren't any remaining niggles that need to be dealt with.

5.2.923

• On the states expression editing dialog, comparison value's ... column is always comparing to the value of a field, so it should look at the selected field and invoke the field value selection dialog to make that a visual operation. Also, hot keys aren't working in the attribute editor in that dialog (see next item.)
  
• Since it is the Win32 way, I implemented accelerator tables in the 5.x transition and have made use of them. But, it was a mistake to use them for anything other than application level menu hot keys, because the Win32 dialog box loop doesn't translate them. So if you use them in any sort of reusable window, and one of those gets used in a dialog box, those hot keys won't work. So far it looks like the attribute editor is the only one that got done that way, just trying to save a little work. So update it to just directly do the keyboard input processing itself.
  
• Some of the little 'add' buttons were using an ellipsis instead of a plus sign as they should. Ellipsis means another dialog is going to be invoked. They may still be some, they'll get dealt with when found.
  
• The old Lutrons HWs driver should use the name AND the type as a uniqueness check. It's ok to have an LED and dimmer with the same name as long as other bits added to the field names for those types are different.
  

[Dean Roddey]

I put up a build for 5.3. I won't post it for a day or so yet. If you guys want to go ahead and upgrade, go for it. If no one has any issues, I'll post it.

http://www.charmedquark.com/Web2/Downloa..._5_3_0.exe

I did some trailing small fixes, which are:

5.3.0

• The attribute editor, when the ... is clicked for multi-line text attributes, was not enabling the handling of CRs in our standard multi-line input dialog. So you couldn't get new lines into the text, it would just invoke the default button when you hit enter.
  
• The interface engine's update to use the new publish/subscribe method was checking to see if the changed widget had a parent container and calling the publish call on the parent. But it needs to also see if the changed widget itself is the container and call it on itself if so, else we don't get changes to the template itself reported, such as drag resizing and whatnot.
  
• There were a number of issues with the template itself redrawing properly in the editor window when resized by mouse or keyboard or attribute editor. It wouldn't redraw the selection handles correctly. It wouldn't adjust the window size in all cases to contains the new template size and so forth.
  
• If the mouse was over the template while you adjusted the size via Ctrl-arrow keys in the attribute editor, the status bar would not update to reflect the size change because sizes (as opposed to areas that child widgets have) changes were not being published as widget changes.
  
• The new Integra NR676 support didn't include Video1 in the valid video sources.
  
• Forgot to update the About dialog's copyright date.
  
• If you resized a widget via the attribute editor using the ctrl-arrows or ctrl-shift-arrows, it wouldn't correctly update the display so that the selection handles wouldn't get updated.
  
• When I stopped using accelerator keys in the attribute editor in the previous drop (they don't work in dialog boxes), Alt-arrow keys stopped working for adjusting the position of widgets in the attribute editor window, because apparently any alt shifted key is reported via WM_SYSKEYDOWN instead of WM_KEYDOWN. So I need to handle both of those in my core dispatching framework.
  
• And now using any alt-shifted key combos in a window without a menu or accelerator causes a system ding sound. There is no way to get around this that really works. So, the attribute editor now uses Ctrl-arrows as before for changing the size of a size or area, and Ctrl-Shift-arrows for changing the origin. Oh well...
  
• The Omni client side driver's documentation was not updated to reflect the 5.x UI changes.
  

[Dean Roddey]

Version 5.3 is now released so this thread is closed and a new one will be opened.