Number: 1672

Date: 18-Jul-84 17':15':42

Submitter: Sannella.PA


Subject: Make SPELLCASEARRAY look at KEYACTION table to set up equiv. classes; flush RPARKEY

Assigned To:


Status: Open


Problem Type: Design - UI

Impact: Annoying


Frequency: Everytime

Priority: Perhaps

System: Programming Environment

Subsystem: DWIM



Lisp Version: 

Source Files: 

Microcode Version: 

Memory Size: 

File Server: 

Server Software Version: 

Disposition: '
["Sannella.PA" "21-Aug-84 18':19':31" Assigned% To': Status':(Fixed->Closed) In/By': Description':]'
["Sannella.PA" "20-Sep-84 15':47':58" Subject': Attn': Status':(Closed->Open) In/By':]

Description: '
Date':  3-Jun-84 17':58':14 PDT'
Subject': new SPELL'
To': lispcore↑'
I did some work on the spelling corrector algorithm. One of the tings I did was to make the ''closekey'' algorithm use a CASEARRAY to determine whether two characters were supposedly on the same key -- this replaced a lot of special case ad-hocery inside the spelling corrector. (I also got rid of MODEL33FLG heh heh). '
What I haven''t done yet is to modify the case array on system startup. I also should and haven''t flushed LPARKEY and RPARKEY to look at the spelling case array. The idea is that at startup the system can look at the KEYACTION table and tell what characters are on the same keys and use those to set up the equivalence classes inside the SPELLCASEARRAY.'
I also made a step back to the old behavior of lowercase to uppercase coercion even with FIXSPELL.UPPERCASE.QUIET. Basically, there is a special case for finding "functions" which will recognize the U-CASE of  a function as a respelling of the L-CASE if the U-CASE is actually GETD, even if it isn''t on the spelling list. However, it requires FIXSPELL.UPPERCASE.QUIET = T for the respelling to be ''quiet'' -- without warning. '
It should be now that if you say'
car((a)) that you will get an = CAR, so that you know it is spelling correcting. I had removed the old behavior because it was not consistent.'
I have yet to recompile the files, much less get them into a loadup...'


Test Case: 

Edit-By: Sannella.PA

Edit-Date: 20-Sep-84 15':47':58