2023 week #43: SL TPVD meeting summary

Goatswood, August 2023, blog post

The following notes were taken from my chat log transcript and audio recording of the Third Party Viewer Developer (TPVD) meeting held on Friday, October 27th, 2023. Afraid no video for this month, Pantera has a lot going on.

Meeting Overview

  • The TPV Developer meeting provides an opportunity for discussion about the development of, and features for, the Second Life viewer, and for Linden Lab viewer developers and third-party viewer (TPV) / open-source code contributors to discuss general viewer development.
  • As a rule, these meetings are:
    • Generally held once a month  the third or fourth Friday, at 13:00 SLT at the Hippotropolis Theatre. See the SL Public Calendar for specific meeting dates.
    • Open to all with an interest in viewer development.
    • Conducted in a mix of text and voice.
  • The notes herein are drawn from a mix of my own chat log and audio recording of the meeting, and are not intended to be a full transcript.

Viewer Updates

The Github Actions (GHA) RC viewer, version 6.6.16.6566955269 was promoted to de facto release status on October 25th. This marks the first release viewer built using the Github Actions (indicated by the extended version number). This viewer also includes a significant update to the Chrome Embedded Framework (CEF) for media handling.

The rest of the current crop of official viewers remains as:

  • Release channel cohorts:
  • Project viewers:

General Viewer Notes:

  • Upcoming official viewer can be viewed on the Github Actions build page, so what is coming down the line can be seen; however artefacts within the list are not publicly linked to downloadable items such as installers. This may change in the future.
  • The Emoji RC viewer is defined as being “in test”, with the other RC viewers in a maintenance cycle to gain parity with the new release viewer.

PBR Materials Update

  • The back-end support for PBR Materials is liable to be deployed to both the BlueSteel and LeTigre RC channels (so approx 3,000 regions) in week #44 (commencing Monday, October 30th), pending the code clearing QA.
  • Core work remains in “maintenance mode” with bug fixing and a focus on driving the project forward towards release.
    • It is believed the only significant blockers remaining are those of performance issues on the Mac version of the viewer – and LL are happy to take suggestions from TPV developers on how to get the Mac variant of OpenGL perform better.
  • One additional issue which had been reported was that of non-PBR enabled viewers appearing to suffer high pack losses on PBR regions (BUG-234550). However:
    • Investigations appear to have shown this as being a “sensor ghost” – the viewer reporting messages it receives but which are not in its message template (a would be the case with a non-PBR viewer receiving PBR specific messages) as a packet loss, which is not really accurate.
    • To deal with the specific issue with non-PBR viewers on PBR regions, and as an interim measure until the PBR viewer is the de facto release viewer, the updated message template for the PBR viewer is to be incorporated into the current RC viewers, together with a suppressor to prevent non-PBR viewer generating false packet loss updates. The code will then be available to all official “pre-PBR” viewers as they are updated with the RC code.
    • TPVs wishing also wishing to eliminate the false reporting of packet loss in their non-PBR viewer releases are encouraged to also pull the PBR viewer message template commit from Github and incorporate it in any upcoming pre-PBR updates they may have in the works.
  • Overall, and allowing for any list-minute issues + clearing the Mac viewer issues, the plan is to pretty much release PBR Materials on the basis of what is currently incorporated in the simulator code and the current RC viewer.
    • This means there will be a range of known issues within the release; however, it is believed that none of them will break any existing content on the Main grid, nor will any fixes for them which are introduced following the initial release.
  • One of the outcomes of the glTF  / PBR project has been a closing of the “divide” between the simulator and viewer engineering teams, with people working on both sides as the PBR project has progressed. This will continue through further iterations of glTF support (e.g. through elements such as the Interest List, which is both viewer-side and simulator-side, and which is not currently geared towards optimal handling scene-wide updates, as will be required as additional aspects of the glTF specification are incorporated into SL).

In Brief

  • There was a general discussion on the internals of viewer development (general message handling, viewer asset throttling, location of “official” masters of things like the message template, the status of the Lab’s Libcurl implementation in light of recent libcurl security issues (in short: “stuck” on an earlier version, so not directly affected), HTTP2 vs. HTTP3 updates,  etc.).
  • A general commentary on Windows 32-bit usage (not no longer officially supported) – with Firestorm stats reporting around 1.9% of their user base still appear to be running 32-bit windows – but the majority of this figure might be down to an Intel bug reporting 64-bit users running 32-bit.
  • A general discussion on the viewer + time-zones, providing a means to add the day of the week to the viewer log-in screen, rather than those in time zones other than California have to check the local day via the viewer clock tool-tip once logged-in.
    • In this, the fact that the time reported in the viewer is “PDT” (and/or PST? I’ve never actually noticed if the viewer reflects changes between PDT and PST notation) rather than “SLT” – which is the acronym everyone in SL understands – can also be a source of confusion. As can the fact that elsewhere within Second Life properties UTC (GMT) can be used.
    • during the discussion it was indicated that LL will likely try to be more consistent in referencing SLT rather than physical-world time zones.
  • Reference was made to a forum discussion on event message polling (particularly viewer side region start-up) which can have an impact (/unintended consequence) in terms of things like message execution. Monty Linden is involved in these discussions and in investigations, and improvements will be appearing in upcoming simulator releases.
  • The above led to a general conversation on reliable SL documentation sources (the broad answer in case of technical documentation being “the code”), and the historic reasons for public-facing documentation becoming fragmented (e.g. the old issue of SL wiki pages being repeatedly vandalised & so locked from editing by users who might otherwise be well-placed to update them), etc.

Next Meeting

† The header images included in these summaries are not intended to represent anything discussed at the meetings; they are simply here to avoid a repeated image of a gathering of people every week. They are taken from my list of region visits, with a link to the post for those interested.

2 thoughts on “2023 week #43: SL TPVD meeting summary

Comments are closed.