Re: [Closed REQ 5523]: webOOGL

["Tamara Munzner" <munzner>]

Hi Rich,
 Thanks for checking out the package and the explicit bug list! 
 
>x weboogl.sgi.tar.gz, 984328 bytes, 1923 blocks
>I pretty sure the last of these files isn't needed and it 
>seems to take up half the distribution. 

Uh, yeah, I was being an idiot. I just fixed it, thanks for pointing
it out.

>1) Sometime when I first click on a vrml url geomview, weboogl and
>gvgeturl all start up and then they all die with out reporting any
>error messages. Clicking on the same url after I've got geomview to
>start ok doesn't cause any problems.

Hmm. We'll have to look into this.

>2) The programs don't allways seem to clear up their tempory files,
>this caused me some disk space problems as they have been creating
>files in /tmp part of my machines very small root partition.

While the program is running it doesn't clear away the temporary
files, since it's trying to cache them. But it should clean everything
up when you quit the program. Are there still files lying around after
you've killed off Geomview? 

Right now the only way to clear the cache is to kill and restart
Geomview. We might come up with something a little more elegant in the
future...

>3) I've also ran out of swap space when looking at some of the
>large architeture models. 

No great fix coming for that soon from the WebOOGL side, some of those
files are just huge. But on the systems side, Irix 5 deals with swap
space more rigorously than Irix 4 did, so processes run out of swap
space a lot more than they used to. You can reconfigure things to run
the old lax way, which will give you more breathing room but isn't
guaranteed to be correct and will occasionally lock up the machine as
it used to. I don't know the details of how to do this, but I think
Stuart does.


>4) Sometimes I get a Netscape: subprocess diagnoistic
>Can't load command_language's from /tmp/21482.808149797.weboogl.tmp
>when I look at the tmp file it says 
>vrml2oogl: Couldn't parse VRML.
>followed by some geomview commands.

>I get this error looking at any
>of the files in http://coney.gsfc.nasa.gov/Mathews/Objects/.

All of his files are not legitmate VRML, they have this illegal
"Callback" node in them that kills the parser. There's a real problem
with people putting up files that aren't legal VRML.

> (Shouln't the error message go to stderr rather than stout?) 

Maybe we'll change this. 


>p.s. You might like to have a look at some of my webOOGL files at
>http://thom.amtp.liv.ac.uk/~rjm/parade/

Nice! I take it you're one of the people who'd like to use the
oogl2vrml converter? We'll try to get that packed up real soon now. 

Tamara

PS We might get another release out in the next day or two that fixes
some bugs. Keep an eye on
http://www.geom.umn.edu/software/weboogl/setup.html