[Opensim-dev] HG Versions and Changes

[Diva Canto]

Also, there has been generalized confusion about HG in/out to/from 
OSGrid. It looks like HG in OSGrid recently stopped working. This is not 
because of OpenSim/HG itself, it's because the OSGrid setup is a lot 
more complicated than a normal grid, and something seems to have been 
misconfigured recently -- or is pushing the boundaries of OpenSim again.

OSGrid uses several proxies -- these are HTTP servers that are placed 
between client requests and the actual OpenSim servers. Proxies are 
common place on the Web in general, and I'm happy to see that OSGrid is 
chartering territory in figuring out how to use them for OpenSim-based 
worlds. But, as expected of new territories, it's a bumpy road. The 
OSGrid admins haven't had time to revise their settings for the HG proxy 
yet, which is clearly at fault in the failures.

On 12/12/2010 8:13 AM, Diva Canto wrote:
> On 12/12/2010 2:26 AM, Ai Austin wrote:
>> The "interface" numbers I was referring to was something in a commit 
>> a while ago which meant HG1.5 grids that previously could 
>> interoperate stopped being able to do so.  kidd peko of "The 
>> Hypergates" started referring to HG1.5 i6 and the newer HG1.5 i7 
>> grids to help testers and those interoperating know which sub-set 
>> they were in.
>
> OK, thanks for the clarification about where these numbers come from. 
> I understand the intention, but this numbering is misleading and, 
> essentially, incorrect when it comes to HG. Let me clarify.
>
> We have been using "interface version numbers" to refer to 
> communications between simulators and their central services *within* 
> grids. We introduced these numbers mainly because of OSGrid, where the 
> central services are run by one group of people and the simulators are 
> run by others, and therefore the software upgrades are not coordinated.
>
> The HG uses some of the communications that are used within grids, 
> and, as such, when these interface numbers change, it usually means 
> that HG between grids in different interface numbers may stop working. 
> I say "usually" but not "always"; it really depends on what changed, 
> because the HG uses only a subset of the grid communications. In the 
> case of the bump from interface 6 to 7, it affected the HG, because of 
> the way appearance is handled. (appearance handling is done in exactly 
> the same way intra- and inter- grids)
>
> However, HG introduces a whole new set of communications that doesn't 
> exist within grids. And, so far, there hasn't been any number for 
> denoting this new set of comms, other than HG 1.0 vs. HG 1.5. As far 
> as I know, no changes within the HG 1.5 series have introduced 
> backwards incompatibilities. So if things don't work between grids in 
> interface number 7, they are bugs that can be fixed, if the situations 
> are properly reported.
>
> HG version numbering doesn't exist yet, just like archive version 
> numbers haven't been put to use yet. Now it's about time that we start 
> introducing these version numbers. Justin is doing some of that for 
> archives, I'll try to do it for HG too.
>
>
> _______________________________________________
> Opensim-dev mailing list
> Opensim-dev at lists.berlios.de
> https://lists.berlios.de/mailman/listinfo/opensim-dev
>