250 bytes - what to do with it?

If you could get SL to send 250 bytes of text of your choice directly to your computer and run a program of your choice to do something with the 250 bytes, what use could you make of this ability?

I bet retailers would like a Point of Sale Journal of sorts. One that can be compiled into a database so sales/marketing data can be extracted.

I'd do nothing at all. llHTTPRequest does this now, with (theoretically) no limit on the size sent to your compueter, and up to roughly 2000 bytes in reply.

Why would I want to downgrade to 250 bytes?


- Newfie

I was under the impression that llHTTPRequest was another of the commands that is intended for those with access to a server to use it. How does one get it to speak to a machine behind a cable modem and a router, like I and many others have?

Product update verification info, bug reporting, networked vendor sales reports.

Hmmm - web based shopping would get a whole lot easier to set up..

suez, i haven't really looked at the specifics on httprequest, but it's been possible to set up your home pc like a webserver, and free. check out apache server. also i remember using something called DNS2Go in the past to get around the changing ip address issue

That in itself is a universal problem, regardless of if it is llHTTPRequest or not. To have LL communicate with a home machine, it either has to have an open access connection (personal web server/ftp server/etc), or a method that can request the info from within the firewall (email, XML-RPC, etc).

Even an IM service could do this, but in all cases they boil down to a program running on your system, with some type of open incoming port, or a third-party server that acts as a forwarding agent.

In other words...we have a form of connection already that fits that bill, called llHTTPRequest.


- Newfie

This unfortunately requires human intervention to respond to a dialog box, but here it is anyway.

The llLoadUrl command sends an URL to your default browser.

The registy key HKEY_CLASSES_ROOT\HTTP\shell\open\command stores the default browser that is called by llLoadUrl.

I created a batch file to test this and called it FakeBrowser.bat .

I set the above key to "C:\Program Files\Fake-Browser\Fake-Browser.bat" %1

In the FakeBrowser.bat I put


I created an object with a script that calls llLoadUrl, set to call "http:qqqq"

This worked, the batch file got called, echoed "qqqq", and ran Opera, which failed to find a page called qqqq.


Instead of a batch file use a program you write that checks the url for a specific string that you add at the beginning of the strings that you want to process yourself. If your custom program doesn't find the flag string at the beginning of the string it gets passed, it calls your real browser and gives it the %1 parameter and the url gets opened in your browser as normal.

If your program does find the flag string, it process the rest of the message and doesn't call the browser.


Like I said, this requires you to click on the dialog that LSL produces that asks if you want to open the browser.

I have looked at both apache and something like DNS2Go, I don't recall which one it was. It looked like getting apache to work would make my brain explode so I haven't pursued that.

I just like to fiddle around sometimes.

I use no-ip.com myself, as a dynamic-IP DSL subscriber. Ive found it extremely reliable, but I still wouldnt want to run anything of a commercial nature off of my home computers.

Apache isnt really that bad - Im just getting my feet wet in the server arena, and it was just point and click to setup a simple webpage. (It even gives you a little taskbar widget that lets you turn it off and on )

I modified a few settings to move the default website folder to a different harddrive, but if you dont need that, you dont even need to touch the config file.

I still havent fiddled with any PHP though, Ill reply once I figure that out.
==Chris

Ya know i've done stuff like that with other programs but never thought to use the LoadUrl as a quickie method of sending data to someones PC. Brilliant Suz

Thanks, Seronis, for your overstatement.

I think I got the memory of looking at the Apache instructions confused with looking at the vBulletin installation instructions.

I have now installed Apache. Running the installation program worked. Folllowing installation one is supposed to configure it, and configuring it means editing a text file. I guess one is supposed to just ignore the contents of the text file, except for defining the root folder, but just looking at the configuration file is disturbing. The Apache feather did not show up in the system tray following installation as it should have, but the 127.0.0.1 page shows up like it should.

The real challenge here is "Can I get this installed and move on to the next step in being able to get llHttpRequest to work before I lose interest or just plain forget what I was doing?"

But how do you get the client's IP address to send the HTTP request to the right place?

I'd culture some cells from a frog brain, and use telemetry information from SL to "train" the cell culture to navigate a box around SL. It could wander the land and help people like Kane did in "Kung Fu".

The apache version of the apache server didn't seem to be able to install the apache monitor, so I tried the xampp version. This did appear to install completely and correctly but that doesn't mean you have a a functioning web server. You have to get your ports working correctly, by getting them forwarded or mapped in your router. This appears to be a bit of a pain, you may have already filled up the slots in the router's port forwarding screen, and your ISP may be blocking the commonly used port 80 and other common ports.

D-mn! i LIKE the way you think.....

I just wrote my own program to get and recieve data directly from SL/llHTTPRequest. It does nothing usefull, just parsed some of the basic data fields. I tried apache, but it was easier for me to make my own program then learn apache, lol..

As far as I know, you can send a body up to 2048 bytes, I never tested the limit.
The wiki says HTTP_BODY_MAXLENGTH = 2048, but also says its not yet supported.

I wouldn't. That's all I need is for some bug to screw up my puter.

trojan!

This thread is a bit obsolete at this point, but - the information would be information of your choice, and the program that intercepts the call to your default web browser is a program that you select.

If you are unwilling to run programs of your own choice - possibly one you wrote yourself - to process data of your own choice, such as, for example, using a web browser to send the keystrokes that your produce and cause them to be displayed in a forum website, then your computer should be sold or donated to a school or sent for recycling because you won't be able to do anything at all with it.

Bah! There would be even MORE bugs if I wrote it myself.

I get where you were going with this, but how about a different question.

If you could assign the DEBUG_CHANNEL to be a .txt file on your pc would you find it helpful.

Answer - bl***y hell yes.

More bugs than Second Life?

I think not!

It takes a team of software engineers to do that.

its 2048, i hit it all the time

Earlier I used a page scraper with one of these free "Your IP address is" pages

I also know one person that used a screen reader to automate the dialog boxes that you have to click no to keep getting payed at my site. So much for randomizing the button order.
Now I am working on random misspellings.

Also rember the older http request system allows your PC to contact an LL object.