The majority of the notes in this update are taken from the TPV Developer meeting held on Friday, October 21st. The video of that meeting is embedded at the end of this update, and references to it are indicated through the use of time stamps in the paragraphs below. My thanks as always to North for recording and providing it.
Server Deployment Recap
- On Tuesday, October 18th, the Main (SLS) channel was updated with the server maintenance package previously deployed to the 3 RC channels, comprising the new Experience Key scripted sitting. Details: Google overview; wiki page.
- On Wednesday, October 19th, the three RC channels were all be updated with the same new server maintenance package, comprising internal simulator changes and a fix for (non-public) BUG-37758 “OBJECT_REZZER_KEY returns creator key of root key with all in-world copied objects instead of the agent who copied it.”
SL Viewer Status
[02:35] On Thursday October 20th, the current RC viewers were both updated for parity with the release viewer:
- Project Bento RC (avatar skeleton extensions), updated to version 220.127.116.110815
- Maintenance RC viewer updated to version 18.104.22.1680804
As per my Bento update #31, there are still issues being addressed within that viewer, which some affecting existing content on the grid, which the Lab is gradually addressing.
There is no “release date” for Bento, as with all of the RC viewers, it will remain in the viewer release channel until such time as the Lab feel confident in promoting to release, something dependent on a number of factors (crash rates, remaining issues, other viewers in the release channel, etc). However, the hope is it could reach release status be US Thanksgiving.
So, if you are a mesh avatar / accessories content maker, and you haven’t tested for issues with your content on the Bento viewer (e.g. possible deformations of avatars) – now is the time to install the viewer and test your products and JIRA issues; this is especially true of rigged mesh. If issues aren’t reported, the Lab cannot investigate them.
[04:46] Work on the 64-bit versions of the viewer is progressing, and they should be appearing “soon”. [11:15] More than half of the required libraries are now built ready for the new build process, out of a total of around 50. When available, the plan remains to offer the viewer on Windows as 32-bit and 64-bit, and on Mac OSX and Linux as 64-bit only.
[06:50] As I’ve previously reported, the Lab is working on a 360-degree photosphere capability built into the viewer. The initial test version of this viewer is progressing and may be appearing in the next week.
It works by capturing 6 images (zenith+ 4×90-degree wide images around the horizon + nadir), which are saved to your hard drive as a .ZIP file, together with a JSON file. [23:46] When issued, the viewer will be an extremely “first cut” version which will have flaws in it, but the idea is to get it in front of the community and work with the community to make the capability something useful to everyone.
I’ll have more on this, hopefully in week #43, once I’ve been able to play with it some more.
The viewer also includes some contributed capabilities, some of which appear to break snapshots under certain conditions – see BUG-40745.
During the TPVD meeting, Oz floated the idea that in order to make Linux support somewhat easier, would be to stop providing an integrated binary, distributed as a TAR file with links to all the required Linux libraries maintained by the Lab directly (some of which are broken or have odd requirements). Instead, the Lab would focus on providing the means to build the viewer specific to one version of Debian Linux, then distribute that package, but not distribute or build all the various dependencies required for the viewer, and instead leave these to TPVs to provide as a part of their Linux platform.
This is seen as beneficial, as it would remove many of the idiosyncrasies involved in building the Linux viewer, and eliminate the overhead of having to build and maintain multiple libraries associated with the viewer, transitioning the build process back towards a more “Debian like” means of providing a basic viewer package.
The downside is, the work would initially require a lot of input and upstream contributions from open-source developers. However, Oz indicated that if sufficient interest is show, the Lab would – other priorities allowing – be willing to work with the viewer community to put such a mechanism for Linux builds in place, providing it doesn’t result in changes with break the Windows or Mac viewer build process or result in a functionally different viewer.
[05:03] The voice updates are awaiting an issue resolution following testing between the Lab and Vivox. These updates will include support for a new codec, includes better error handling and addresses some of the more serious Voice exploits to make it “much, much” harder for Voice to be abused. The latter is seen as a “step on the path” to a larger solution for Voice exploit issues.
[08:00] The Lab is interested in receiving recommendations on how to improve the inventory UI in the viewer. However, such recommendations should be “well-considered and carefully described”, and which should preferably include implementations of the ideas stated in the recommendation, or at least screen shots (captured using the official viewer) and clear descriptions of the recommendation.
Such ideas may then be included in the upcoming round of work the Lab is planning with regards to eliminating more of the common inventory problems people can experience, and make inventory handling more robust. As such, ideas that are simply cosmetic in nature (e.g. ideas for different inventory icons, etc.), are unlikely to be seen as high priority.
Viewer Code Clean-Up
[25:02] An under-the-hood project with the viewer is to clean-up a lot of the code. Some of this work in the current Maintenance viewer, with further work still being carried out. Among other things, it is hoped that these changes will reduce things like shut-down crashes and other issues, as the are released into the viewer.
[29:58-end] There is a general discussion session covering assorted topics, including:
- The fix so that people banned from a group are booted from an active group chat, although ejecting them from the group will not kick them from any active group chat; so ejecting then banning can fail, although the Lab will likely examine this
- The windows version detection issue and Intel HD2000 and HD3000 graphics, which can cause severe issues for users. This is apparently an Intel OpenGL implementation failing to recognise Win 10, but one solution the Lab is considering is have the 32-bit version of their Windows viewer won’t opt-in to the Win 10 features, and thus avoid the issue for 32-bit users.