Server Deployments: week 4
Because Monday January 20th is a holiday in the USA, Maestro Linden has posted details of the deployments which will take place in week 4. Despite the US holiday, the deployments will still take place on Tuesday and Wednesday as usual.
As always, please refer to the week’s forum deployment thread for the latest news and updates.
Main Channel: Tuesday January 21st, 2014
The Main channel should received the server maintenance project which has been on the RC channels for the last couple of weeks. As notes in previously SL projects updates in this blog, this comprises:
- The forced delay in llLoadURL has been reduced to 0.1 seconds
- New LSL functions for uniformly scaling linksets (integer llScaleByFactor(float scaling_factor), float llGetMinScaleFactor(),float llGetMaxScaleFactor() )
- Preliminary support for custom default permissions on newly created objects – currently awaiting viewer-side support via STORM-68, which has yet to be released
- Preliminary support for the viewer fetching the LSL syntax rules from the simulator, for up-to-date syntax highlighting in the script editor – currently awaiting viewer-side support via STORM-1831, which has yet to be released.
Release Candidate Channels, Wednesday January 22nd, 2014
All three RC channels should receive the same server maintenance project, which contains a single fix for a crash mode.
SL Viewer Updates
The following section contains notes taken during the TPV Developer meeting held on Friday January 17th, a video of which is included below. My thanks, as ever, to North for the latter.
As reported here, the Fitted Mesh viewer reached release candidate status on Thursday January 16th, with the release of version 126.96.36.1995178, which also includes a number of bug fixes and STORM contributions from third-party developers.
Issues Affecting ATI/AMD Users
There are reports of additional issues arising within the Fitted Mesh viewer which are not apparent in other version of the SL viewer. These issues appear to impact those using ATI/AMD graphics cards and have Advanced Lighting Model disabled.
In the first issue, non-rigged mesh is invisible when viewed on the Fitted Mesh viewer (see FITMESH-20), which was reported for the new release candidate version of the viewer. In the second (FITMESH-6), worn mesh which is not rigged to the collision bones is seen to stretch away towards the 0,0,0 point of the region the avatar is standing in (which is oddly similar to how viewers without the Fitted Mesh updates render garments rigged to the collision bones in an avatar).
Again, these problems only appear to impact those using ATI/AMD graphics cards and who are running with ALM disabled. Initially, the problem had been thought by the Lab to be a driver issue, but testing reveals the problem to occur across a range of ATI/AMD drivers, including the most recent Catalyst updates.
New Sliders and Asset
Redpoly Inventor has also raised a feature request (see FITMESH-19) for what he calls a “Mesh Clothing Adjuster” which can be used to scale specific parts of the collision skeleton in order to achieve better fitting for the likes of knee boots, tank top straps or any clothing that doesn’t look good with alphas. He also includes code with the JIRA which people can add to their viewer in order to create slider to the shape menu which controls the clavicle volumes.
The HTTP project viewer was also updated to a release candidate status on Thursday January 16th, with the release of version 188.8.131.525253, which includes an experimental fix for long-standing DNS problems, improved mesh downloads with better TCP connection behaviour and general reliability improvements for mesh uploads / downloads.
There is a request for anyone who has experienced DNS problems in the past, and who may have been using the Google DNS workaround to download and try this viewer, as it should largely or completely fix those issues.
Those who have suffered network issues in general, particularly if they are using older modems / routers and have had frequent disconnects, should see an overall improvement in their connectivity to SL, if not any actually noticeable performance improvement per se (the work isn’t about viewer performance as such, it’s aimed at making SL much more tolerant of heavily used and lossy networks.
Monty Linden has written an extensive, but relatively easy-to-follow blog post on the HTTP work, and it makes recommended reading.
Interest List Viewer
Also a release candidate, the interest list viewer is reported to be “over the worst” of its crash issues, and is exhibiting a crash rate which is better than the release viewer, although it needs to accumulate more hours of user use before it is likely to be promoted to the de facto release viewer.
Merchant Outbox project viewer (currently version 184.108.40.2064731): It is anticipated that there will be further iterations of this viewer, aimed at resolving issues some users have experienced when trying to use the Merchant Outbox.
Google Breakpad viewer: there will be a further release of this viewer, which is aimed at improving the gathering and reporting of viewer crash data, although it has emerged that some of the problems with the viewer crash reporting mechanism are actually at the server end of the equation, and these are currently being looked at, and so a further update is unlikely to appear until after these have been corrected.
Group Ban List viewer: it is currently unclear as to the precise status of this work, although elements of it, in Oz’s words, speaking at the TPV Developer meeting, appear “to be a lot more complicated than we thought”. Doubtless, Baker is still working his way through the code and getting things sorted.
Oculus Rift viewer: the code for supporting Oculus Rift is “in the pipeline” for a release, although it is not clear whether this will be a project viewer or a release candidate when it appears.
Sunshine / AIS v3
The AIS v3 inventory updates, seen as a part of project sunshine are currently awaiting further testing prior to the viewer-side code being taken any further forward.
As reported in these pages, the Lab is keen to progress this project, but the code was only made available to TPVs to be included in test versions of their viewers just before the end-of-year holidays in 2013. As a result, time is required for TPVs to integrate the code for testing. The Lab has therefore decided to postpone a pile-on test which had been provisionally scheduled for week 3 until such time as more TPVs have test / experimental versions of their viewers ready to assist with the testing.
Concerns over Court ruling on Net Neutrality
On Tuesday January 14th, The US Court of Appeals for the District of Columbia Circuit struck down rules adopted by the Federal Communications Commission in 2010 meant to protect the openness of the Internet. As a result, this has led to some concern as to the possible future impact for SL users on the basis of the bandwidth required by the platform in server / viewer communications.
While there is no guarantee people will not be affected, it is thought that people should not find their use of SL adversely impacted in the future – although obviously, no guarantees can be given. Much is being done to reduce the amount of bandwidth SL uses (the ongoing HTTP work), and more will be done in future to further streamline things (HTTP pipelining), so hopefully by the time any changes resulting from the court’s ruling start to come into effect in the future (this won’t happen overnight), any risk of potential impact may be further mitigated.
A Racing Sun
Maestro Linden offered-up a curious tidbit for people attending the Server Beta meeting on Thursday January 16th.
“Somebody recently noticed that the default day cycle doesn’t have a consistent Sun speed,” he told attendees at the meeting. “It turns out that the default day cycle’s ‘6AM’ and ‘6PM’ sky settings are not actually set to 6AM or 6PM, they were more like 6:22AM and 5:42PM! So the sun would move more slowly between 3pm-6pm in the region’s day cycle, then pick up a bit between 6pm-9pm, for example.” So now you know why the evenings fly by when in-world 😉 .