Hi there.

I'm working on a research project, and have been asked to get Second Life and processing to communicate.

I have successfully gotten processing to write a file with a php script, and then access it from SL (with slhttprequest), but I am having trouble doing the reverse.

This part of this very simple experiment works as such :
There are three coloured toruses in Second Life, that when I click any of them, accesses a php script which writes either a 1,2 or a 3 to a text file, depending on which torus clicked. (it writes this file to the /data/ directory of my sketch) The file is called sl2proc.txt

Then, I have this simultaneous Processing sketch running, where I attempt to access the file with the loadStrings function.

The thing is : It works the first time, ie. When I load my browser (firefox) up, and load the sketch url, it accesses the file and pulls the proper data.

Then, when I change the data in the textfile (either through clicking a torus, or manually editting the text file), suddenly Processing wont recognize the change (It tries to update every 3 seconds).

It continues to tell me that the data it has pulled from the file is the same (by not changing the colour indicator, and via a println). If I close firefox at that point, and reopen it, the sketch now recognizes the change and updates accordingly. Just closing the sketch window and/or reloading the Processing url doesn't fix it. I need to completely close firefox to have it access the file anew.

It seems like, when I first use Loadstrings, the Java runtime environment puts the textfile in memory, and refuses to shake it, despite the actual change in the file. Is there some way to address this? To get processing to recognize that the file has been modified? to forcibly close the filestream?

http://caferisko.ca/TEST/test3/sl2proc/applet/index.html
// this is the applet

http://slurl.com/secondlife/Tkachiev/91/211/601
// this is where the modifying torus are
// they are on a platform 601 ft in the air.

Any help would be greatly appreciated
thanks!
max

That had originally been my problem,

When that was the case, the loadstrings() always read the same data, until I re-exported the processing file. Which had me very confused. Exactly like you say.

I did figure that out, and now I remove all of the text files before I recompile.

but - since the sketch reads the proper info when I reboot the browser (firefox), this suggest that it actually is pointed at the proper file (and not accessing some constant file within its jar), does it not?

thanks for
the response though.
I may try using loadstrings with a full path, instead of just a relative one, to see if that helps.

Well, I tried the absolute path string, with no success.

.. but, I've found that if I open the Java Console Control Panel, and go into the 'cache' section, and browse the 'current files' in the cache, I can see the SL2proc.txt there.

If I delete that entry from the cache,
the sketch will work properly, and load proper data from the file, but of course then it puts this new datafile back in the cache, and refuses to budge from it unless I delete the file from the cache again.

So apparently there is no problem with pathnames, or filenames, or the loadstrings function at all, but rather a problem with the Java console saving a cached version of the file I access, and then refusing to re-access it on request - relying instead on reading the saved version in memory.

That is why when I shut Firefox (or IE) right down, and reload it, it disengages from the cache, and reaccesses the file on the server anew.

Is there some function I can use to reset my Java cache from within processing?

thanks
max

if the data is being pulled from a server, try adding a ? and a random number to the end of the url, to fool the cache. or disable the java/browser cache. but we're basically at the mercy of how the browser chaches things here.
loadStrings("http://somesite.com/yourthing.php?1432425");

to run locally, use a full path name to the file:
loadStrings("/home/yourname/Desktop/somefile.txt");
which should not cache.

hey maxtuckett!

I am also doing reasearching work about Second Life and Real Word for my Master degree, i would like to share some ideias with you, and discuss some topics as well. Fell free to e-mail me fm_84@clix.pt or send a IM to Filipe Roff, inworld.

Thanks