There is also the XML socket-based remote debug protocol that is in there now which is used by the Eclipse IDE. This is not suitable for in-page app-to-app communication though, since it needs to connect via TCP socket.
The LocalConnection stuff is quite fast, though there is a limit of about 46 kbytes or something on the amount of data that can be sent in a call, and it silently fails if you go over that. We had plans to implement a simple peek-poke command set, just haven't gotten the cycles or demand to do that yet. Sounds like it might be worth pushing this stuff to the next level.
On 11/18/05, Jim Grandy <[EMAIL PROTECTED]> wrote:
On Nov 18, 2005, at 2:13 PM, Henry Minsky wrote:I thought I heard you and Tucker talk about splitting the console out so that it ran separately from the target canvas. This would make it into a standalone lzx app that could persist its own state without worrying about all the issues that come from running in-process.Side note, from a failed DHTML developer: All this stuff would be simpler if we could use components in the debugger UI, so you could pop up a prefs window like we used to, and click a checkbox for this and other goodies.
Also: could you also do the same with the debugger? I guess that would require a remote debugging api...I tried to make a pop up dialog window in the dev console, but it was painful because the console dev app is just embedded in a thin strip in the page, and it needs to be larger to see the dialog box, which in turn means painful _javascript_ hacking in the browser to resize the containing HTML DIV element, and at that point I gave up.
--
Henry Minsky
Software Architect
[EMAIL PROTECTED]_______________________________________________Laszlo-user mailing list
--
Henry Minsky
Software Architect
[EMAIL PROTECTED]
_______________________________________________ Laszlo-user mailing list [email protected] http://www.openlaszlo.org/mailman/listinfo/laszlo-user
