SL Viewer Beta and Viewer Development and CHUI
There have been updates to both the beta and development viewers. The beta viewer moved to release 126.96.36.1991345, and development to 188.8.131.521386. Both releases were focused on CHUI, and according to Oz Linden, “Right now they are almost identical; different only in the viewer number I believe.” He went on to say, “There is significant uncertainty as to when CHUI will come out of the beta channels, and it’s very likely that will remain true for a while yet – that is, that the two will remain pretty similar.”
With one CHUI update having just gone to the beta viewer, the Lab anticipates there will be at least one more update for the project while in beta before it makes any more to the viewer release channel.
Server-side Baking (SSB)
Server-side Baking saw the release of a further viewer update – 184.108.40.2061419 – on March 7th, which was partially in response from information gained from the first public pile-on / load test for SSB, held on Thursday February 21st.
Commenting on the ongoing work to integrate SSB into TPVs at the TPV Developer meeting on Friday March 8th, Oz linden thanked the TPVs for their work to date on trying to integrate the viewer-side SSB code into their offerings in order to keep the project on track. “I’m very pleased to say that right now it does not look like support for it [SSB] in third-party viewers is going to end up being the limiting factor on when it can be deployed,” he said, although he did caution against showing any complacency in getting the code to a deployable condition.
Merge with CHUI
Also at the TPV Developer meeting, Nyx indicated that overall, the project viewer for SSB is becoming more and more stable, and that his team now has and initial merge with the CHUI code, which they are going to be “hammering on”, prior to pushing it to the internal Sunshine branch. As such, the most recent release of the SSB project viewer is possibly the last push prior to CHUI appearing in the viewer.
Given there are concerns over the merger and its possible impact on TPVs, Nyx is going to see if it is possible to maintain a side branch of the SSB viewer code which does not include the CHUI merge, which could continue to receive fixes and make it easier for TPVs to obtain them, however, until this has been looked into in more detail, he is unwilling to commit to how easy it would be to achieve and maintain – or how long for.
Second Pile-on / Load Test
Attending the Server Beta meeting in week 10, Nyx Linden announced that the updated viewer will be required for a further pile-on / load test, which has been scheduled for Thursday March 14th. The test is liable to be in much the same format as the first test, and those wishing to participate are advised to attend the Server Beta meeting on Aditi ahead of time (the meeting commences at 15:00 SLT on Thursdays).
Those wishing to take part in the test should also ensure they are using the latest version of the SSB project viewer linked-to above, as this as this has been specifically set-up to enable correct logging of data, etc., for the test. Nyx hopes that the latest updates to the logging parameters, coupled with the recent inventory fixes applied to Aditi should do much to both lessen the impact of Aditi issues should they occur and make them easier to identify when analysing feedback. However, Nyx suggested that those who have experienced Aditi inventory issues and who wish to join-in the test on March 14th should log-into the Sunshine test regions ahead of time and confirm that they can access their inventory using the project viewer.
The Z-offset solution seen in the current SSB project viewer came up at the TPV Developer meeting, with Nyx commenting, “We’re still talking about it internally. We realise the new solution will work well where people have things like shoes that put them above or into the ground, but should not be very effective for the use-case of needing to adjust your offset for animations.” He went on to say that nothing has yet been decided, and so he has nothing further to report on the matter for the time being. Oz Linden added, “We’re discussing it. We’ll see what comes out of that. We’ve got some of the use cases covered, and we may or may not end up covering them all. We’ll just have to play it by ear at this point.”
There are still no definitive dates for any SSB deployment to the main grid. Several viewers now have the SBB code merged into experimental / alpha versions which are publicly available, with Firestorm having a private merge of the code which is still being worked on. However, LL are still very involved in working on things at their end on both the viewer and server sides of the equation.
One of the concerns with the project is that the server-side of the new service will have sufficient “ovens” (servers to handle the baking process for thousands of avatars on the main grid), and Nyx confirmed this is still on LL’s mind as well. “It’s going to be a pretty slow roll-out process,” he said, “And we’re doing a lot of load testing to make sure we are acquiring enough ovens to do all of this baking.” As a part of this work, Nyx and his team will also be working with the Firestorm Beta Testers in a further load test utilising just the Firestorm viewer, in order for the Firestorm team to better assess their readiness for an SSB release of Firestorm to a wider audience once SSB roll-out approaches.
While there is still no release date for this project and work is progressing on fixing the current viewer-side problems with materials processing, Oz Linden has confirmed that the viewer code is unlikely to be released until after CHUI has been merged with the release viewer (as noted above, when CHUI is to be merged with the release viewer still has yet to be determined) and “almost certainly” after SSB has been deployed to the main grid.
The server-side HTTP changes are now confirmed as being out of QA. Currently, the plan is to set-up a number of project channels on Aditi to allow for further testing of the updates, which I outlined in part 2 of my week 9 update. “There are a couple of special regions we agreed to create for the testing of meshes and testing of textures, as well as a few other ones,” Monty Linden informed those attending the TPV Developer meeting on Friday March 8th.
In all, Monty anticipates three channels being set-up on Aditi for testing purposes. One of this will be the current channel the Lab has been using for tests – DRTSIM-203, which is liable to be the channel which is carried over for future testing on Agni, together with two other channels which have yet to gain official designations, but which Monty has temporarily named “Happy” and “Angry”. One of these will have “more generous” connection capabilities associated with it and the other more controlled capabilities, being stricter in areas of resource-sharing, etc. One set of regions on these channels will be very texture-heavy – Monty estimates around 6,500 textures per region, and another set of regions will be mesh heavy with low textures and there may be a third set of regions with a mix of textures and mesh. There is also likely to be some generic sandbox regions as well to allow scripters to test the capabilities with their external HTTP services. In addition, the project channels will be available for developer to request regions be added to them (e.g. by contacting Maestro Linden).
As soon as these channels are set-up, which will most likely be in week 11 (week commencing Monday 11th March), Monty plans to announce their availability through the relevant LL mailing lists, to allow people to start testing the new capabilities. As mentioned previously, he’s particularly keen to see the TPVs and scripters join-in the testing on Aditi, although as he again noted during the TPV Developer meeting, the “real test” will come when the capabilities move to a Release Candidate channel on Agni, where he expects to see a “true load” placed on the new services. In the meantime, the Aditi test environment will give people the opportunity to try-out the new capabilities and see if anything breaks, as well as seeing how everything works with various flavours of viewer.
Password Changes and Aditi
While there have been fixes implemented for the Aditi inventory issues, it is still likely that any password change will result in creating Aditi inventory issues. Linden Lab is still looking at this, and a further fix is in progress, but has yet to be rolled out. So the current advice is that if the most recent fixes to Aditi have repaired your inventory – do not change your password in order to try to force a further Aditi inventory update.