Cool VL Viewer forum

As the Cool VL Viewer's first public release 8th birthday happened a few days ago, here is an update to the roadmap:

• Today saw the promotion of the experimental branch (1.26.15) to the stable status (branch renumbered 1.26.16 to reflect this); see this message about the CEF3 plugin current status. Note that I will keep the QtWebkit plugin support in the viewer for a (probably long) while...
• The former stable branch (v1.26.14), which only differed from the new one by its lack of CEF3 plugin support, is now deprecated and has been retired.
• A new experimental branch (1.26.17) will soon be opened and will see the introduction of the new HTTP code (coroutine-based) as implemented in LL's (pre-RC-and-very-much-experimental) viewer-azumarill branch.

For any question or comment about this roadmap, please post here.

Since the "coroutinized" LLCore HTTP code I debugged (a lot !) adapted and optimized/cleaned up/simplified, i.e., in one (much fuzzier) verb, "backported" to the Cool VL Viewer experimental branch proved rock stable over the past several weeks, tomorrow's releases will see the opening of two new branches (v1.26.18 and v1.26.19) that will replace the v1.26.16 and v1.26.17 branches (the latter branches being immediately deprecated as a result).

The v1.26.18 branch will be the new stable branch, equivalent to v1.26.16 with the new LLCore HTTP code backported to it from v1.26.17.

The v1.26.19 branch will be the new experimental branch (merely a renumbering of v1.26.17), which, for now, only differs from v1.26.18 by the support of the Bento skeleton/avatar and the removal of the (deprecated since never adopted/used in either SL or OpenSim) mesh deformer code.

In the coming weeks, I'll make the 64 bits move for the viewer builds. Linux first (this week will see the move of the experimental branch to 64 bits), then Windows (when Linden Lab will have a full set of 64 bits pre-built libraries ready, meaning also that we must wait for them to replace the QuickTime plugin with another, a libvlc based one, most probably).

As for MacOS-X, I'd need to first finish the "de-Carbonization" of the sources, but not being a Mac developer myself, and not having any Mac at hand, this might take longer and cause quite a few bugs to get introduced in the Mac builds...
If a seasoned MacOS-X developer would be ready to contribute, I'd gladly take the offer...

The Cool VL Viewer is now 9 years old, so, to celebrate, here is an update of the roadmap:

When LL will finally release Bento (still no set date...), the current experimental branch (v1.26.19) will become the new release branch (v1.26.20) and the current release branch (v1.26.18) will be immediately deprecated (my advice is therefore that you start and use v1.26.19, so to see if anything is incompatible in it with your system).

I then intend to try and adapt the sources so to allow the building of the viewer on recent MacOS-X systems (including in 64 bits, when LL will make a full set of 64 bits MacOS-X pre-built libraries available). However I'll have to do it in "blind mode" (I don't have a MacOS-X system myself) by backporting LL's MacOS-X code, and let Catten sort out the compilation issues... Unless a seasoned MacOS-X programmer comes up and volunteers to do the work.

The (public release of the) Cool VL Viewer will soon be 10 years old (time flies !). In anticipation for this celebration, here is an update to the roadmap:

Today saw the first release of the new v1.26.21 experimental branch, with Animesh support and modifications paving the way to future libraries updates by LL, as well as to 64 bits MAC builds, whenever LL will feel like releasing their viewer64 branch... Note that v1.26.21 does not support VS2010 builds any more, and old plugins (QtWebkit, winmmshim) have also been removed.

After Animesh will go into release, the v1.26.21 branch will be renumbered and promoted to the next stable release, v1.26.22 branch, and v1.26.20 will be immediately deprecated. This will also mean the end of VS2010 Windows XP 32 bits builds for private builders (sorry, but I cannot afford rebuilding LL's future libraries for Windows XP).

In just a couple of weeks, the (public release of the) Cool VL Viewer will be 12 years old.

By then, the current experimental branch (v1.26.23) will likely have been promoted to a new v1.26.24 stable branch and, consequently, v1.26.22 will have been deprecated and retired (so, if you did not yet test the experimental branch, it is about time to do it and report any issue you would encounter, compared with the current stable branch).

The next milestone will be the EEP renderer implementation... or not !

Linden Lab's Extended Environment Project is bringing two things:

1. A new way to exchange and propagate environment parameters (and this is already implemented in the v1.26.23 branch, and will therefore as well be present in v1.26.24).
2. A new set of rendering shaders and a modified rendering pipeline (at least for the sky and water rendering parts), allowing to render more/extended effects and parameters.

Sadly, LL's current implementation of the EEP renderer is subpar (and especially taxing on performances), and since I am not personally interested in any "better" environment rendering (as a role-player, visual realism is far less important than what is left to my imagination), I would hate (read: refuse) to loose performances for a feature I didn't even ask for in the first place.
LL hired new developers to improve EEP (and I do hope their first task and goal will be to make its renderer as fast as Windlight's). So we will see what the future brings us. I personally do not expect to see EEP reaching release status in LL's viewer before 2020.

I will likely create an experimental viewer around the EEP renderer, as soon as significant work will be resumed on LL's EEP viewer. Should I make public releases of this viewer, I will likely bump its minor version (i.e. the experimental viewer would be v1.27.0 and, should it reach stable status, it would become v1.28.0), to reflect the large changes in the rendering pipeline itself. As usual, that viewer would be kept fully compatible with OpenSim (*).

For any question or comment about this roadmap, please post here.

---

(*) People telling you it is impossible to keep OpenSim compatibility in a viewer that also implements all of SL newest features is either not competent enough a programmer, or a plain liar (the reason for that lie being most probably laziness).

13 years and counting... And time for a small update to the roadmap !

The v1.28.1 experimental branch is about ready to be promoted to the next v1.28.2 stable branch (this will certainly happen before this year ends): if you did not test the v1.28.1 viewer just yet, it is about time to do it (and to report any issue you would encounter) !

The current v1.28.0 stable branch will be deprecated as soon as v1.28.2 is released. Keep in mind that v1.28.0 is the last 32 bits compatible branch and that v1.28.2 will be 64 bits only (given the result of the poll, it should not pose any problem to anyone).

For any question or comment about this roadmap, please post here.

This week sees the opening of the new experimental v1.29.0 viewer.

This is a version minor number increment and not just a branch number increment since it involves a lot of changes to the renderer (with the removal of Windlight renderer/shaders); it will see the backport (and the implementation, for Linux OpenGL shared contexts, since LL did not implement them in their viewer code) of LL's "performance viewer" code changes, which will bring very significative frame rates gains.

Please, note that this new version involves the compatibility breakage with old systems (in particular the ones unable to properly use vertex buffers); do not even think of running it on an old GPU (or in a VM) !

For this reason, I will keep v1.28.2 around for a while as a "legacy branch" viewer, even after v1.29.0 will have been deemed good enough to be granted the "stable" status (at which point it will see its version number bumped to v1.30.0).