2019 SL User Groups 15/3: TPV Developer Meeting

Green Acres; Inara Pey, March 2019, on FlickrGreen Acres blog post

The following notes are taken from the TPV Developer meeting held on Friday, April 12th, 2019. A video of the meeting is embedded below, my thanks as always to North for recording and providing it. This was a relatively short meeting, with several periods of audio silence and text chat. The key points of discussion are provided below.

SL Viewer

  • The Estate Access Management (EAM) RC viewer updated to version 6.2.0.526190 on Friday, April 12th, 2019.
  • As noted in my CCUG summary, the EEP viewer updated to version 6.2.0.526104 on Thursday, April 11th.

All other viewer in the pipelines remain as:

  • Current Release version 6.1.1.525446, formerly the Love Me Render RC viewer, dated March 26, promoted April 2 NEW.
  • Release channel cohorts:
    • Bakes on Mesh RC viewer, version 6.1.1.525409, March 26.
    • Teranino Maintenance RC viewer version 6.1.1.525401, March 20.
  • Project viewers:
  • Linux Spur viewer, version 5.0.9.329906, dated November 17, 2017 and promoted to release status 29 November – offered pending a Linux version of the Alex Ivy viewer code.
  • Obsolete platform viewer, version 3.7.28.300847, May 8, 2015 – provided for users on Windows XP and OS X versions below 10.7. This viewer will remain available for as long as reasonable, but will not be updated with new features or bug fixes.

With the necessary back-end updates now made, it looks likely that the EAM RC viewer will be the next viewer to be promoted to de facto release status.

Teleport Issues

Teleport Disconnects

  • The Lab continues to try to understand the root cause(s) of the teleport disconnects. So far, they can only confirm it is not specific to any single viewer.
  • As previously indicated in my updates, the root cause of the issue is a race condition / timing issue – something that could be exacerbated by the distance of a user from the Lab’s servers (the grater the distance = the more time required for communications = the greater the chance of a timing issue / race condition).
  • Similarly, a high packet loss between viewer and simulator(s) might increase the chances of a disconnect occurring.
  • The problem remains in trying to identify precisely what has changed to cause any race condition / timing issue.
  • As noted in this week’s SUG meeting summary, there was a server RC deployment on Wednesday, April 10th – version 19#19.04.09.526122,. This included additional logging to help capture further data on the teleport disconnects, particularly capturing where the region handshaking may be failing.
  • Multiple regions have been replicated on Aditi, and having further enhanced logging enabled on them. This additional logging capability may be deployed to Agni in week #16, but not necessarily enabled, due to the performance impact it will have. However, the code will be in place should it be required.

Attachment Loss on Teleport

  • The Lab now has a “solid” repro for this issue. As anticipated, much of the problem appears to be related to having too many scripted attachments stacked on the same attachment point (e.g. the right hand, the default attachment point, and often used by default by mesh clothing makers).
  • However, the work on developing a fix has slowed as a result of resources being moved over to the teleport disconnect issue.

Asset UDP Messaging Deprecation

As noted in my SUG summary, it appears the code for the Asset UDP messaging deprecation didn’t work as anticipated, likely the result of one code element failing to merge correct. There will be a be a further RC deployment in week #16 with the corrected code.

A reminder that assuming the deployment is made and the code works as anticipated, anyone using very old viewers still reliant on UDP messaging for assets and which do not have support for HTTP messaging will not be able to receive any updates for the following: animations, body parts, clothing, gestures, landmarks, meshes and sounds whilst connected to regions on any RC channel.

In Brief

  • The are reports of issues with the EEP code interfering with Interest List culling, with Draw Distance having little effect on scene rendering (BUG-226752). This is being investigated.

 

3 thoughts on “2019 SL User Groups 15/3: TPV Developer Meeting

  1. ATTACHMENT LOSS ON TELEPORT
    The Lab now has a “solid” repro for this issue. As anticipated, much of the problem appears to be related to having too many scripted attachments stacked on the same attachment point (e.g. the right hand, the default attachment point, and often used by default by mesh clothing makers).

    — LL is not wholly correct on this. I have had this sporadic ‘ghosting of attachment’ after teleports (where the object is still seen by the wearer, but is invisible to other people) for almost 10 years. It used to happen to my hair, which was attached to my head. Nothing else was attached to my head: the hair was the single attachment. It still happened. And the hair was not scripted. In my experience, this glitch happens to worn objects which are comprised of multiple prims (whether scripted or not). For example, it will happen to an unscripted necklace made up of many prims. Or the old-fashioned hair with sculpties and flexies. The glitch happens on any viewer. It also often happens when a region is busy, or lagged. Nowadays (for the last 8 years!), I have an (inconvenient) workaround: I put all multi-prim worn objects into sub-outfit-folders which I detach and re-attach, pre- and post-teleport. Certainly, I would love LL to fix this, but after almost a decade, I’m not holding my breath. 😉

    Like

    1. There are multiple issues with attachment loss. The quoted comment wasn’t intended to give the impression the Lab consider this to be “the” issue and “the” solution (although on reflection, perhaps I should have used “one” rather than “a”, in difference to Kyle Linden’s comment). As such, expect to see further attempts to identify other causes and fixes.

      Like

      1. Good to hear that. I hope they are aware of the ‘non-scripted-multi-prim-linked-object’ cause I described, haha. I have a totally unscripted chain with monocle that can be *guaranteed* to ghost whenever I TP with it worn.

        Like

Comments are closed.