[Opensim-dev] HG Versions and Changes

[Diva Canto]

First, let me just state upfront that I have no idea what interface 
numbers mean. I haven't used them, and since there have been no 
releases, I'm not sure what these numbers mean. Is this the grid-wide 
interface number? If so, HG compatibility is correlated with it, but but 
HG involves a lot more than intra-grid comms and needs its own 
numbering. So far, there have been no HG interface numbering yet. Likely 
to change soon.

All the work that you are talking about is happening in developers code. 
Developers code often has bugs. That was probably the case here.

otakup0pe did a serious cleanup of service addressing in on his own 
repo, which wasn't that much exposed to testing. I then took his changes 
and merged them to master. Additionally, before I merged his code, I did 
a few more things to HG related to the preservation of creator 
information -- that was the reason why the configs changed.

As is often the case, otakup0pe's code wasn't perfect, there were a few 
bugs that were exposed as I and others started testing, so they were 
corrected. I'm sure my own changes may still have a few buglets. If you 
happen to be running a revision that has those bugs -- too bad, things 
will likely fail.

In general, when ppl test developers code, you *must* expect that things 
will *not* work. Many times things break by accident, and we fix them a 
few revisions later; other times they break on purpose, and that's when 
we warn people.

When things are supposed to work, and don't, that's a bug. Please report 
bugs as usual! -- but using revision numbers, not these interface 
numbers.  Thanks!

On 12/11/2010 3:24 PM, Ai Austin wrote:
> We understand that there is incompatibility between the old HG 1.0 and 
> the more secure HG1.5.
>
> As I understand it there was a change in some data base entries from 
> "interface 6 to interface 7" which also meant that HG 1.5 i6 versions 
> and before were incompatible with HG i7 and later.  At least it was 
> observed that HG 1.5 i7 girds could interwork, and those could not 
> work with HG 1.5 i6 ones. Is that correct Diva, or is that just the 
> assumption that was made by some seeking to diagnose HG jump issues?
>
> But now in testing recent versions with recent config changes for HG, 
> there may be further incompatibility with previous HG 1.5 i7 grids, or 
> it could be that they are just not yet properly configured.  It would 
> be helpful if Diva could let us know if an incompatib8klity to earlier 
> HG1.5 i7 grids is known, =and what release version in 0.7.1 dev master 
> is the cut off point to have as a minimum upgrade level to be 
> compatible again.  If there is an incompatibility, it cold be useful 
> to have a way to refer to this new partition of the hypergrid 
> compatible grids.
>
> I am happy to continue testing if anyone wants to check out what is 
> happening, or to try to identify the issues in recent versions of 
> OSGrid with lack of HG jump capability. Some of the testing being 
> done, and the grids known to work under the new versions and that can 
> be used to test are listed at 
> http://opensimulator.org/mantis/view.php?id=5259
>
> Grids that are known to work (via a chatted link) two way when on 
> latest 0.7.1 dev master are on our Openvue grid, and a test region 
> that Diva set up at UCI...
>
> secondlife://virtual.aiai.ed.ac.uk:8002:Openvue/ (at 1000,1000)
> secondlife://virtual.aiai.ed.ac.uk:8002:Vue-5000/ (at 5000,5000)
> secondlife://virtual.aiai.ed.ac.uk:8002:Vue-9000/ (at 9017,9017)
> secondlife://ucigrid03.nacs.uci.edu:9000/ (at 9000,9000) region name 
> "UCI Grid 03"
>
> _______________________________________________
> Opensim-dev mailing list
> Opensim-dev at lists.berlios.de
> https://lists.berlios.de/mailman/listinfo/opensim-dev
>