Day: May 20, 2025

2025 week #21: SL SUG meeting -SLua

Posted on by Inara Pey
Chaos and Calm – Chaos Theory, March 2025 – blog post

The following notes were taken from the Tuesday, May 20th, 2025 Simulator User Group (SUG) meeting. They form a summary of the items discussed, and are not intended to be a full transcript. The notes were taken from my chat log of the meeting. Pantera also recorded the meeting, and that recording is embedded at the end of this piece – my thanks to Pantera, as always, for providing it.

Meeting Overview

  • The Simulator User Group (also referred to by its older name of Server User Group) exists to provide an opportunity for discussion about simulator technology, bugs, and feature ideas.
  • These meetings are conducted (as a rule):
  • Meetings are open to anyone with a concern / interest in the above topics, and form one of a series of regular / semi-regular User Group meetings conducted by Linden Lab.
  • Dates and times of all current meetings can be found on the Second Life Public Calendar, and descriptions of meetings are defined on the SL wiki.

Simulator Deployments

  • There are no planned deployments to any channels this week, only restarts.

SL Viewer Updates

SLua Updates

I’m going to write up an implementation of better event handling in pure SLua so that people can get a feel for the API and make suggestions before we commit to actually implementing a specific implementation in the server.
Got sidetracked improving setlinkprimitiveparamsfast perf with many PRIM_LINK_TARGETs since I saw that bogging down the SLua alpha sims but I’m back on track now 🙂

– Harold Linden (SLua development contractor)

  • It was reiterated that the plan remains to move SLua to a more open beta on Agni (the Main grid), once further updates have been made by LL. (Such as the event handling and memory management mentioned). When this happens, it will initially be to a “handful of regions” prior to gaining broader deployment.
  • Harold Linden indicated that most of the current Canny Feature Requests for SLua will be implemented during the beta phase of the project – although some, such as “require” will take time to “do properly”, and so there may be stopgap solutions “similar to the ones people already use for LSL that uses local pre-processing.”
Not ideal but there’s a lot of yak shaving that needs to be done to get proper “require” working with new “library” type assets and don’t want to block the rest of SLua on it. 
  • In responding to a comment about an possible inter-script dependency system, he further noted:
There would be no inter-script dependencies per-se, this is more akin to pure modules that can just export types and functions etc with no shared state. More like a statically linked library that only gets evaluated when the script is compiled. It’s essentially just enabling something that people already tend to do with pre-processors, the only difference is the assets live with the rest of your stuff.
  • This sparked a debate on the use of libraries, plug-ins ,etc., a lot of which is “still TBD”, other with matters of object update handing, etc. This discussion ran through must of the remaining 40 minutes of the meeting. Please refer to the video for specifics (as I am not a coder, and far from qualified to interpret all of the discussion!).
  • One important additional point was raised by Harold Linden:
Oh right, I feel like I should mention, we’re trying to track down a couple potential memory corruption bugs in SLua that we’ve had trouble reproducing on our dev machines. We’re going to look at pushing a version of the server with debugging instrumentation (AddressSanitizer) to the SLua sims on beta grid. It might make things slightly slower, but it’s only temporary.

In Brief

“Missing Content on Login”

We had a repro recipe for “missing content on login” and I was able to kinda fix it. It was semi-reliable (would happen maybe 20% of the time) for one our viewer dev contractors, but I couldn’t reproduce it myself. It is “kinda” fixed in that I have eliminated some unnecessary resets during login that would help trigger the bug, with a subsequent savings of effort and bandwidth of object update data on arrival. However the root race condition that is being avoided… I didn’t actually figure that out. 
So… I would expect improvements (reduction of missing content on login) but I don’t think this fixed everything. In particular, I thought maybe it would help reduce the rate of missing attachment items on avatars, but upon more thought I realized this fix is not relevant to that problem. Anyway, heads up: this fix is expected to be in Fig Pudding [the next simulator update].

– Leviathan Linden

TP Failure and Log-in Issues

  • Reports suggest that that has been a “big increase” in TP failures and login failures over the last week (I can attest to the former; having had around a 40% TP failure rate on first attempts to get somewhere on several days). Viewer logs and times incidents occurred, submitted via Canny, would be helpful in debugging issues.
  • It was also noted that LL is engaged in making some network infrastructure changes, so again, a request was made for those “experiencing connectivity problems not of the usual sort or frequency” to file a Canny report would be useful. Times, places, names, and log files.
  • The game_control capability remains on-hold. Leviathan Linden is currently engaged in other work, and unfortunately, game_control is is “near the bottom of” the stack of work he has on-hold as a result.

† 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 rooftop of people every week. They are taken from my list of region visits, with a link to the post for those interested.