Number: 2036

Date:  1-Sep-84 16':13':45

Submitter: Masinter

Source: Masinter

Subject: want new forms list to supplant AFTERSYSOUT/MAKESYS FORMS

Assigned To: 

Attn: Masinter, Lisp, Kaplan

Status: Open

In/By: 

Problem Type: Design - UI

Impact: Moderate

Difficulty: Moderate

Frequency: Everytime

Priority: Perhaps

System: Programming Environment

Subsystem: History

Machine: 1132

Disk: 

Lisp Version: 31-Aug-84 16':17':57

Source Files: 

Microcode Version: 5124

Memory Size: 4096

File Server: 

Server Software Version: 

Disposition: '
["Sannella.PA" "10-Sep-84 14':32':22" Description':]

Description: Date': 31 Aug 84 18':37 PDT'
From': JonL.pa'
Subject': RESTOREFILES'
To': KAPLAN'
cc': LISPSUPPORT'
'
This function from MACHINEINDEPENDENT is probably just "excess bagage" for Interlisp-D.  Two problems I recently encountered': (1) It isn''t being called anyway, due to a soon-to-be-reported bug in AFTERSYSOUTFORMS, and (2) the global variable it depends upon, RESTOREFILELST, is unbound in a sysout implying that the initialization didn''t get called, so the subsequen call form AFTERSYSOUTFORMS wouldn''t do anything anyway.'
'
Your initials, dating from 1979.  Care to comment?'
'
-- JonL --'
'
-----'
'
Date': 31 Aug 84 19':20 PDT'
From': JonL.pa'
Subject': Poll on AFTERSYSOUTFORMS'
To': LispCore↑'
Reply-to': JonL.pa'
'
Glaaag, I''ve finally tracked down an inscrutable bug whereby MAKESYS produces a sysout that, when sysin''d and started up, *doesn''t* run the AFTERSYSOUTFORMS. '
'
Since the IRM, Page 14.4 describes MAKESYS as "... almost the same as SYSOUT, except that it does some cleaning-up operations (such as clearing the screen).", then I propose not adding yet another set of variables for MAKESYS -- e.g., BEFOREMAKESYSFORMS and AFTERMAKESYSFORMS -- but simply have MAKESYS do the same "forms" that SYSOUT would do.'
'
Can anyone ever think of a situation wherein he would *not* want the AFTERSYSOUTFORMS run when starting up from a sysout that had been dumped via MAKESYS?'
'
-- JonL --'
'
'
-----'
'
Date': 31 Aug 84 21':29 PDT'
From': Kaplan.pa'
Subject': Re': Poll on AFTERSYSOUTFORMS'
In-reply-to': JonL.pa''s message of 31 Aug 84 19':20 PDT'
To': JonL.pa'
cc': Lispcore↑'
'
I think that MAKESYS automatically does greeting when the sysout starts up, whereas SYSOUT fixes it so that its behavior is conditioned on the value of SYSOUTGAG (e.g. prints a warning if you''re not the same user, but never re-greets automatically).'
'
I would imagine that lots of carefully tuned loadup procedures would behave differently if SYSOUT and MAKESYS were merged.'
'
My belief is that MAKESYS is NOT SUPPOSED TO run AFTERSYSOUTFORMS.'
'
--Ron'
'
-----'
'
From': MASINTER.pa'
Date': 31-Aug-84 23':04':35 PDT'
Subject': Re': Poll on AFTERSYSOUTFORMS'
In-reply-to': Kaplan''s message of 31 Aug 84 21':29 PDT'
To': Kaplan'
cc': JonL, LispSupport'
'
The problem is that a lot of separate things got confounded in AFTERSYSOUTFORMS that belonged more as user-specified things to happen in the same frame as the device-event functions. There''s the'
'
forms-to-be-run-when-you-might-be-on-a-new-machine'
'
forms-to-be-run-if-it-might-be-days-later'
'
forms-to-be-run-if-some-files-got-closed-underneath-me'
'
and a few others.'
'
There is some random cruft on AFTERLOGOUT forms too.'
'
Users use these ''cause they aint got no better.'
'
Oh, there is also PREGREETFORMS and POSTGREETFORMS.'
'
Sorting this out means looking at all of the various form-lists and gathering what the real semantics are, and making a definite proposal.'
'
There *is* a difference we''ve made between MAKESYS and SYSOUT which is exactly that one runs GREET and the other AFTERSYSOUTFORMs.'
'
I think the solution  is to take anything that you want on *both* and put it somewhere else.'
'
Date': 31 Aug 84 23':30 PDT'
From': JonL.pa'
Subject': Re': Poll on AFTERSYSOUTFORMS'
In-reply-to': MASINTER.pa''s message of 31-Aug-84 23':04':35 PDT'
To': MASINTER.pa'
cc': Kaplan.pa, JonL.pa, LispSupport.pa'
'
PREGREETFORMS and POSTGREETFORMS aren''t documented.  Anyway, the problem, as I see it, is that the discrimination if *far far* too fine for most user needs (the system''s needs seem to be served by \DEVICEEVENT now).  I see a lot of <LispUsers> type code that puts the same forms on just about every AFTER<anything>FORMS.'
'
I currently don''t see any documented way to do some thing at every "start up" time.'
'
-- JonL --'
From': MASINTER.pa'
Date': 31-Aug-84 23':33':58 PDT'
Subject': Re': Poll on AFTERSYSOUTFORMS'
In-reply-to': JonL''s message of 31 Aug 84 23':30 PDT'
To': JonL'
cc': MASINTER, Kaplan, LispSupport'
'
How bout this':'
'
we add AFTERSTARTUPFORMS.'
'
We make MAKESYS and SYSOUT and LOGOUT execute AFTERSTARTUPFORMS and urge maintainers of LispUsers packages to use AFTERSTARTUPFORMS if that is what they want to happen.'
'
Date':  1 Sep 84 02':35 PDT'
From': JonL.pa'
Subject': Re': AFTERSTARTUPFORM'
In-reply-to': JonL.pa''s message of 1 Sep 84 02':30 PDT'
To': JonL.pa'
cc': MASINTER.pa, Kaplan.pa, LispSupport.pa'
'
Uh, there''s more.  AFTERSAVEM is a condition now recognized. should it be in AFTERSTARTUPFORMS too?'
'
Worse yet, there may be a need for BEFORESHUTDOWNFORMS -- namely the similar union of BEFORELOGOUTFORMS, BEFORESYSOUTFORMS, and BEFOREMAKESYSFORMS (the latter doesn''t really exist).'
'
-- JonL --'
'
'
'


Workaround: 

Test Case: 

Edit-By: Sannella.PA

Edit-Date: 10-Sep-84 14':32':26