On su, 2010-10-10 at 13:09 +0200, Peter Steinlechner wrote: > thanks for your reply, I had it run for about a half hour without any > results. Also when I try to load Albertos sample scene, it doesn't > show up local. But if I use my test scenes which have smaller mesh > size, they show up local. I'm affraid it must be something related to > the Win7 installation I have, as Albertos scene seems to run perfectly > on his XP installation. About the zip file you mentioned, where should > I look for it? I checked in the same folder where the scene files are.
Yes it seems to make the zip in the same folder. I think that it gets stuck for you at 14% and you don't see .zip files there mean that the zip creation fails somehow. Unfortunately I think the errors from there are not going to the log nor f1 console now, only to the console window which is disabled in releases .. we have to fix that too. That worked for me with both of your test scenes, they both show locally and Naali makes the zips, but so far have managed to publish only the other to the server 'cause am encountering some other prob with my server setup perhaps. Alberto's EDIFICIO 6 gives me a parser error when reading the file locally. I suspect a bug / incompatibility in the simple dotscene loader from the python-ogre project. Are you saying that worked for you? Or did I find a different scene zip than the one you meant. Matti already did some fixes and little ui improvements today, we'll be testing more. > Pedro ~Toni > On Sun, Oct 10, 2010 at 8:47 AM, Toni Alatalo <[email protected]> > wrote: > On su, 2010-10-10 at 09:42 +0300, Toni Alatalo wrote: > > On la, 2010-10-09 at 15:29 +0200, Peter Steinlechner wrote: > > > Have you tried also to export the scene without the camera > and the > > > lamp ? I just selected the objects and it worked well. > > That won't affect the zip making + uploading part. If you > can see the > > scene locally after the load, it could be read ok. The > publish button > > just blindly makes a zip of the files and uploads it using > http. > > > Oops sorry apparently I was wrong there -- the publisher > actually does > read the scene structure to find material references etc., I > guess to > know what textures etc. it needs to put in the zip, so it is > indeed > possible there is some bug there if it doesn't take into > regard entities > without mats like cameras. So this may be worth a try too. > > ~Toni > > > > The people who had problems: did you wait for long? > > > > The fact that the GUI status bar now gets stuck does not > mean that he > > tool got stuck. It just is not yet implemented so that it > would update > > incrementally -- it jumps in big steps when everything is > done, so the > > GUI is misleading. > > > > I read the publish code now, and the zip making part > apparently copies > > all the textures to a temporary directory etc. which can > take a while. > > > > This is probably made worse by that zip making being run in > a thread in > > the same process. So if you have Naali running at > not-ultra-high fps, it > > can be slow (Naali eats all cpu of that processor, and if > you are on > > multicore, the other processors are not utilized for the zip > making now > > because cpython has a global lock). > > > > So if the people who had probs can please test again, let it > run for a > > long time, and see if a .zip file finally appears etc. that > would be > > helpful. > > > > On the dev side I think we should port that to use a > subprocess -- then > > it will not be slowed down by Naali mainloop with the > rendering etc., > > especially on multicore machines should be normal speed for > zip making > > then. Matti R. can see an example in pychessview if wants to > give this a > > shot. Fortunately it should be trivial, perhaps only > 1-2hours of work to > > adapt. I may try it next week if have time, can give an > update then if > > it works out. > > > > Thanks for the feedback! > > > > ~Toni > > > > > On Sat, Oct 9, 2010 at 3:01 PM, MasterJ > <[email protected]> wrote: > > > here i have same trouble too (and more becaue i > have a python > > > error > > > with blender) > > > the progress bar stop at 14%. > > > but i use Naali 0.3.1 and Taiga 0.1.4 > > > > > > i have changed nothing on the opensim.ini here. > > > and i never try uplaod scene from console because > i can't > > > found the > > > command to type. > > > > > > > > > > > > for the ones know python my python error is : > > > http://www.masterj.name/blendersceneerror.jpg > > > > > > MasterJ > > > > > > On 9 oct, 13:00, pedro <[email protected]> > wrote: > > > > > > > > > > Hi all > > > > > > > > Despite scene upload via console workes fine, > scene upload > > > via naali > > > > resists to work on my setup. On the server I > have the > > > following > > > > settings: > > > > > > > > serverside_object_permissions=true > > > > > > > > allow_grid_gods=false > > > > > > > > ; This allows somne control over permissions > > > > ; please note that this still doesn't duplicate > SL, and is > > > not > > > > intended to > > > > region_owner_is_god = true > > > > parcel_owner_is_god = true > > > > > > > > On the client side I run naali 0.3.2 on Win 7pro > with Admin > > > rights and > > > > the most actual updates. > > > > > > > > So when I load a local scene it looks fine, > everything is > > > there. When > > > > I try to publish the scene to the server > nothing happens, > > > except the > > > > progressbar jumps to 14% and gets stuck there. I > can't see > > > any > > > > activity on the consoles, so my wild guess is > that it must > > > get stuck > > > > at creating the zip file. > > > > > > > > Installed is 7zip, but I also tried after > uninstalling it > > > and also > > > > tried with switched off virus scanner without > any success. > > > > > > > > What else could I try to make this strong > feature working ? > > > > > > -- > > > http://groups.google.com/group/realxtend > > > http://www.realxtend.org > > > > > > > > > > > > -- > > > http://groups.google.com/group/realxtend > > > http://www.realxtend.org > > > > > > > -- > http://groups.google.com/group/realxtend > http://www.realxtend.org > > > > -- > http://groups.google.com/group/realxtend > http://www.realxtend.org -- http://groups.google.com/group/realxtend http://www.realxtend.org
