However this one may be a little different, at least initially. I was all set to build 2.20.3 to fix some more problems, but I have decided, due to the amount of changes, that I need to give a bit of a test run first. The only way I can do that is via the beta channels, but the versioning scheme does not allow for beta's of point/patch releases. So 2.21 Beta 1 may actually become a patch update of 2.20, released as 2.20.3 IF nothing significant is found during testing. Ive been very active in the last few weeks, but my life has caught up with me somewhat and the next 3 weeks or so I will only have a little time in the evenings after work - which probably means I wont get much done to XMBC.
I hope the changes I have made that I was going to put out as 2.20.3 are all working and nothing bad has been introduced, but as I'm not 100% confident (especially after having to release two minor releases to 2.20 already!) I figure this is the next best thing.
Thus, 2.21 Beta 1 will contain no new GUI work (no new translations needed) and hopefully no new bugs but of course, generally I would expect beta versions more likely to contain bugs (and sometimes more serious bugs) - although in this case, I rather hope not (famous last words!).
Changes since 2.20.2:
- #804 - Modified versions ox XMBC should log the fact, and some strings in the binaries should be protected from modification.
- #803 - Fixed another Firefox scrolling issue, when remapping to WHEEL UP/WHEEL DOWN. Consolidated scrolling code so they all call the same function to do the window class/style checks.
- #802 - Add further logging around any call to set cursor speed, also log the cursor speed set in Windows when hooking the mouse (e.g. resume from sleep) to diagnose random cursor speed changes.
- #800 - Add ability to customise the Enhanced Pointer Precision threshold values used in Windows to affect acceleration (manually editing the XML settings until 2.21)
- #791 - Further structural changes to try and prevent XMBC from crashing in the audio/volume manager when resuming from sleep, under certain (rare) circumstances.
Please keep this beta thread on topic. It is for announcements and reporting bugs in the current beta, NOT for requesting new features and/or asking about existing (non-beta) functionality!
NOTE: If you have check for beta versions enabled, you should get notified of a new version and prompted to update in the next day or so. This is the most efficient method (bandwidth wise) as the updates are only a fraction of the size of the full install. Otherwise, you can get the full installation beta HERE. Note that this link will always get you the latest beta version!
The two major areas to focus testing are:
- Scrolling, especially with "Scroll window under cursor" enabled. Particularly in browsers (Chrome, Edge and Firefox) and Windows UI (Modern/Store apps, like Files and PowerToys). I have re-worked the code completely to have one place to check the window properties etc, rather than copies (of sorts but clearly not 100%) in at least three different places (nightmare!). Going forward it should make things much easier, but there is a chance I have screwed up things that previously have been working (Ive been checking locally for a few days prior, and so far haven't found anything of concern).
- Volume management (including changing volume by scroll wheel, simulated keys etc). There has been a rare, but persistent crash happening since I released 2.20 to a small number of users (none of whom have contacted me after sending the bug report!). I've been unable to completely track it down, but a prime suspect evolved throughout the 2.20.x changes as I narrowed it down. So I have reworked the volume/audio management code, particularly across threads, to hopefully squish this crash completely or, if not, provide more information on when, where and why it happens. Incidentally, so far it has only happened for users who also run AVolute SonicStucio3 which I believe is a driver package distributed with some sound cards/motherboard on board audio devices. I am unable to install that package here so I have no idea it it is a coincidence or actually related to the crash!