sukumar@ut-emx.UUCP (sukumar) (08/06/90)
some time back i had posted a request to SMALLTALK v/286's problem of not returning control when windows are created with a scheduleWindow. here is a summary of responses. Thanks to all those who replied. Sukumar rathnam sukumar@emx.utexas.edu --------------------------------------------------------------------------- From: Doug Heins - - - HP Boise Division <dheins@hpbsl88.boi.hp.com> I have experienced the same problem several times. When one calls "toppane dispatcher openWidnow scheduleWindow" the Scheduler is sent a message called "dropProcessSenderChain" (I might be slightly off) which drops all pending messages from the stack, and consequently the returning point back into foobar to continue on with the method foo. So here is what you can do to achieve the same result, but regain control after opening the window: "Scheduler add: toppane dispatcher open. toppane displayWindow" Atleast, this is what I have been using to get around the problem; however, I am no expert by any means and there might be a more elegant approach...but it does work. ============================================================================= The message scheduleWindow does not return. The mechanism Digitalk Smalltalk uses to handle multiple windows with multiple processes, is to DROP THE PROCESS STACK when a new window is created. In other words, the entire context of execution is thrown away. As a result, you won't ever see the end of your routine. I believe Digitalk wrote something up in their last issue of SCOOP on this problem, and gave an alternative. Ken Cooper USENET: cooper%netcom@uunet.uu.net Acumen Software Compu$erve: 71571,407 FAX: (415) 324-4383 BIX: kcooper ============================================================================= I have some code at home that accomplishes what you want. I essentially had to reverse engineer what the scheduler methods were doing, then pull that up into my code. the problem is that the scheduler sets up and schedules the window, then throws the method stack away! Brain-dead. If you're interested, I can pull up a fragment and send it to you. Send me mail at yon@world.std.com (the account I read from home) if you want me to hunt that code down. David Yon CASE Consultant ============================================================================= -- Sukumar Rathnam Multi Participant Task Support Systems Research Group, CMIS UT Austin. Internet : sukumar@emx.utexas.edu Home : #3011 Whitis Avenue, Apt #106, Austin TX 78705. Tel: 512-474-2771 Office : MSIS Dept., CBA 5.202, The University Of Texas, Austin TX 78712