Number: 1433

Date: 14-Jun-84 15':37':44

Submitter: Sannella.PA

Source: MASINTER.PA

Subject: Tedit, Chat processes should grab tty immediately, if at all

Assigned To: 

Attn: Sybalsky, vanMelle

Status: Open

In/By: 

Problem Type: Design - UI

Impact: Moderate

Difficulty: Easy

Frequency: 

Priority: Perhaps

System: Operating System

Subsystem: Processes

Machine: 

Disk: 

Lisp Version: 

Source Files: 

Microcode Version: 

Memory Size: 

File Server: 

Server Software Version: 

Disposition: 

Description: '
Date':  5 MAY 84 01':10 PDT'
From': MASINTER.PA'
Subject': grabbing the TTY away'
To':   vanMelle'
cc':   LispCore↑'
'
I agree with your sentiment about ''grabbing the TTY away'', and in fact, would like to see it extended':'
'
I think that CHAT and TEDIT should also not ''grab the tty'' when they are started. In fact, I might make the rule that NO process in the system automatically grab the TTY.'
'
I''ve had it happen tht I''ve started a TEDIT, which was taking a long time, so I went to another process to do something and TEDIT came back and took away the TTY.'
'
I know Mesa works that way, and I like its consistency. It takes a little getting used to switching back and forth between Lisp and Mesa. There are some times when I know that the TTY is going to be in chat later, but the problem is that often the TTY switch is quite a while after the user action tht initiated the interaction.'
'
'
'
     ----- Next Message -----'
'
Date':  6 May 84 21':49 PDT'
From': vanMelle.pa'
Subject': Re': grabbing the TTY away'
In-reply-to': MASINTER.PA''s message of 5 MAY 84 01':10 PDT'
To': MASINTER.PA'
cc': LispCore↑.PA'
'
It sounds like the main thing you are complaining about is that TEDIT and CHAT take the tty away only (potentially) long after they are started.  I agree; however, I believe the correct fix for this is for them to take the tty away immediately, rather than never take it.'
'
I think it is perfectly okay to immediately grab the tty when your process is started by mouse action.  This is not startling to the user; in fact, I would think it confusing if the new process did NOT have the mouse automatically.'
'
I personally don''t like the way Mesa does it.  If I type "Chat" to the exec, it''s annoying to have to go click in the chat window before I can do anything.  But I wouldn''t swear that isn''t because I am used to Lisp''s Chat.'
'
	Bill'
'
'
     ----- Next Message -----'
'
Date':  6 May 84 23':27 PDT'
From': JonL.pa'
Subject': Re': grabbing the TTY away (and AR 891 -- PROMPTFORWORD locks . . .)'
In-reply-to': vanMelle.pa''s message of 6 May 84 21':49 PDT'
To': vanMelle.pa'
cc': MASINTER.PA, LispCore↑.PA'
'
I too tend to think that a direct call on TEDIT/CHAT should switch the TTY into the tedit/chat process asap.  Your comments on the TEDIT/CHAT grabbing the tty, if applied to PROMPTFORWORD, would say that the caller of promptforword should grab the tty even as it makes the call.  Assuming that the caller wants input from the keyboad (probaby 99.44% do).'
 '
The problem for PROMPTFORWORD seems to be that it is being used in many places in a way not quite consistent with its original motivation (the original model for PROMPTFORWORD was SmallTalk''s "unscheduled prompter").  I dislike adding args to an already-overloaded interface, but there ought to be *some* combination of entry into PROMPTFORWORD that maintains that original scenario.'
'
'
'
'
     ----- Next Message -----'
'
Date':  7 May 84 09':56 PDT'
From': sybalsky.pa'
Subject': Re': grabbing the TTY away'
In-reply-to': vanMelle.pa''s message of 6 May 84 21':49 PDT'
To': vanMelle.pa'
cc': MASINTER.PA, LispCore↑.PA'
'
I''m with Bill--I''d prefer to have the new process grab the TTY at once, instead of at some indeterminate time in the future.'
'
     ----- End Forwarded Messages -----'


Workaround: 

Test Case: 

Edit-By: vanMelle

Edit-Date: 21-Jun-84 12':53':24