[Opensim-dev] Baked Texture persistence?
Justin Clark-Casey
jjustincc at googlemail.com
Mon Oct 3 21:49:13 UTC 2011
At the moment OpenSim just ignores it and sends back "nothing cached" every time.
On 03/10/11 20:52, Latif Khalifa wrote:
> Forgot to add. So the sim would have associate CacheID from the AgentSetAppearance's WearableData block with texutre IDs
> provided at that specific bake indexes of the TextureEntry in the same packet and respond accordingly
> in AgentCachedTexture/AgentCachedTextureResponse response excahnge.
>
> What does opensim do currently when it gets AgentCachedTexture message?
>
> On Mon, Oct 3, 2011 at 9:46 PM, Latif Khalifa <latifer at streamgrid.net <mailto:latifer at streamgrid.net>> wrote:
>
> Yes, block ID in the WearableDataBlock AgentCachedTexture packet is a hash of all wearables that are part of certain
> bake + 1 magic uuid per bake. The way it works is:
>
> * After login, client gets the list of wearables from the sim.
> * Computes the hashes for each bake and sends it to sim with AgentCachedTexture packet.
> * Sim sends back AgentCachedTextureResponse with UUIDs of cached bakes it has. UUID.Zero if doesn't have it
> * Client creates uploads only bakes that are a mismatch
> * Client sends set appearance packet and this completes the exchange
>
> Latif
>
>
> On Sat, Oct 1, 2011 at 11:49 PM, Mic Bowman <cmickeyb at gmail.com <mailto:cmickeyb at gmail.com>> wrote:
>
> this is the one where the wearables uuids are xor'ed with some "pink flamingo" uuid to get a hash of the
> wearables combination? i looked at the code in libomv to get an idea for how it might work. the opensim side
> will be a bit challenging to support the map to the baked textures.
>
> --mic
>
>
> On Sat, Oct 1, 2011 at 1:51 PM, Latif Khalifa <latifer at streamgrid.net <mailto:latifer at streamgrid.net>> wrote:
>
> We have client side of the cached bakes check implemented in libomv. In case you are going to implement this
> server side you can ping me for details.
>
>
> On Mon, Sep 26, 2011 at 12:04 AM, Justin Clark-Casey <jjustincc at googlemail.com
> <mailto:jjustincc at googlemail.com>> wrote:
>
> That's great info, Mic - I wasn't aware of this message exchange.
>
> It sounds like in the future it would be a good idea to implement these messages and remove temporary
> assets from the asset service periodically.
>
> Justin
>
>
> On 25/09/11 19:38, Mic Bowman wrote:
>
> I will carefully disagree with Justin on this one. The viewer doesn't upload textures by default. It
> only uploads the
> textures if it believes something has changed. OpenSim doesn't currently respond correctly to the
> v2/v3 packets for
> cached appearance and earlier versions of the viewer can be told that that appearance parameters
> haven't changed so that
> textures are never uploaded (this saves a lot of bw... remember that if you upload textures they
> have to be sent out to
> everyone in the scene so it isn't the upload bw... its like putting new textures on objects every
> time the sim starts).
> In practice, people change appearance so infrequently that the baked textures do not fill up the
> asset database. And
> even when the do, so long as you don't get stuck in the "assets can never be deleted" way of
> thinking... (baked textures
> are marked as temporary anyway)... its pretty easy to clean up the asset db based on a reasonable
> set of policies.
>
>
>
> On Sun, Sep 25, 2011 at 1:25 AM, Neil Canham <neil at knowsense.co.uk <mailto:neil at knowsense.co.uk>
> <mailto:neil at knowsense.co.uk <mailto:neil at knowsense.co.uk>>> wrote:
>
> Ah - I hadn't appreciated that NPCs need to be cloned from a logged in AV. That makes sense.
>
> As for the persistence, like Mic I think there is potentially an argument for it, as an option,
> for closed private
> sims for example, or sims where you expect a lot of people to log in at roughly the same time.
> However, I
> appreciate that would be a big change so thanks for clarifying how it operates now.
>
>
> On Sat, Sep 24, 2011 at 11:18 PM, Justin Clark-Casey <jjustincc at googlemail.com
> <mailto:jjustincc at googlemail.com> <mailto:jjustincc at googlemail.__com
> <mailto:jjustincc at googlemail.com>>> wrote:
>
> For ordinary avatars, baked textures don't need permanent persistence since the avatar
> uploads them on every
> login or when any body part/wearable is changed. This will have no effect on slow rezzing
> or cloudiness - if a
> viewer is in the region it should always have uploaded the textures that it bakes.
>
> Indeed, there's little value in persisting these since the client will upload them on every
> login/region
> entrance. Persisting baked textures would bloat the asset service more than is already the case.
>
> An NPC, on the other hand, has no viewer from which these baked textures are uploaded. So
> they have to be disk
> persisted at the point from which it is cloned from a 'live' avatar.
>
>
> On 24/09/11 19:28, Neil Canham wrote:
>
> I confess that I am puzzled. I found this fantastic description from Mic of how 0.7
> appearance works
> http://opensim-dev.2196679.n2.____nabble.com/What-I-ve-__Learned-__About-__AvatarAppearance-__td5655933.__html
> <http://nabble.com/What-I-ve-Learned-__About-AvatarAppearance-__td5655933.html>
>
> <http://opensim-dev.2196679.__n2.nabble.com/What-I-ve-__Learned-About-__AvatarAppearance-td5655933.__html
> <http://opensim-dev.2196679.n2.nabble.com/What-I-ve-Learned-About-AvatarAppearance-td5655933.html>>
> and it
> fits with what I
> see analysing the packets against 0.7.1. So that seems quite clearly to say that baked
> textures are never
> persisted,
> and that would explain the slow rezzing and 'cloudiness' of avatars that I see
> regularly. However, the new
> npc code
> seems to require that a baked texture is available to copy from. So are baked textures
> now persisted in the
> latest code?
>
>
>
> ___________________________________________________
> Opensim-dev mailing list
> Opensim-dev at lists.berlios.de <mailto:Opensim-dev at lists.berlios.de>
> <mailto:Opensim-dev at lists.__berlios.de <mailto:Opensim-dev at lists.berlios.de>>
> https://lists.berlios.de/____mailman/listinfo/opensim-dev
> <https://lists.berlios.de/__mailman/listinfo/opensim-dev>
> <https://lists.berlios.de/__mailman/listinfo/opensim-dev
> <https://lists.berlios.de/mailman/listinfo/opensim-dev>>
>
>
>
>
> --
> Justin Clark-Casey (justincc)
> http://justincc.org/blog
> http://twitter.com/justincc
> ___________________________________________________
> Opensim-dev mailing list
> Opensim-dev at lists.berlios.de <mailto:Opensim-dev at lists.berlios.de>
> <mailto:Opensim-dev at lists.__berlios.de <mailto:Opensim-dev at lists.berlios.de>>
> https://lists.berlios.de/____mailman/listinfo/opensim-dev
> <https://lists.berlios.de/__mailman/listinfo/opensim-dev>
> <https://lists.berlios.de/__mailman/listinfo/opensim-dev
> <https://lists.berlios.de/mailman/listinfo/opensim-dev>>
>
>
>
> _________________________________________________
> Opensim-dev mailing list
> Opensim-dev at lists.berlios.de <mailto:Opensim-dev at lists.berlios.de>
> <mailto:Opensim-dev at lists.__berlios.de <mailto:Opensim-dev at lists.berlios.de>>
>
> https://lists.berlios.de/__mailman/listinfo/opensim-dev
> <https://lists.berlios.de/mailman/listinfo/opensim-dev>
>
>
>
>
> _________________________________________________
> Opensim-dev mailing list
> Opensim-dev at lists.berlios.de <mailto:Opensim-dev at lists.berlios.de>
> https://lists.berlios.de/__mailman/listinfo/opensim-dev
> <https://lists.berlios.de/mailman/listinfo/opensim-dev>
>
>
>
> --
> Justin Clark-Casey (justincc)
> http://justincc.org/blog
> http://twitter.com/justincc
> _________________________________________________
> Opensim-dev mailing list
> Opensim-dev at lists.berlios.de <mailto:Opensim-dev at lists.berlios.de>
> https://lists.berlios.de/__mailman/listinfo/opensim-dev
> <https://lists.berlios.de/mailman/listinfo/opensim-dev>
>
>
>
> _______________________________________________
> Opensim-dev mailing list
> Opensim-dev at lists.berlios.de <mailto:Opensim-dev at lists.berlios.de>
> https://lists.berlios.de/mailman/listinfo/opensim-dev
>
>
>
> _______________________________________________
> Opensim-dev mailing list
> Opensim-dev at lists.berlios.de <mailto:Opensim-dev at lists.berlios.de>
> https://lists.berlios.de/mailman/listinfo/opensim-dev
>
>
>
>
>
> _______________________________________________
> Opensim-dev mailing list
> Opensim-dev at lists.berlios.de
> https://lists.berlios.de/mailman/listinfo/opensim-dev
--
Justin Clark-Casey (justincc)
http://justincc.org/blog
http://twitter.com/justincc
More information about the Opensim-dev
mailing list