On 3/27/21 7:41 AM, P Falth wrote:
> When a char
> is read from the input buffer it is immediately converted to lower case
> ( a simple $20 OR instead of a UTF8 conversion )

For ascii 91 thru 95 I get something that looks wrong using $20 OR.
What am I missing?

Btw, it made no difference in the unique words count in the
kjvbible.txt file. Maybe that's why you used it.

Thanks.
-Doug

: go ( c --) dup emit $20 or emit ;
91 go [{
92 go \|
93 go ]}
94 go ^~
95 go _

On Saturday, 7 August 2021 at 13:28:09 UTC+1, Doug Hoffman wrote:
> On 3/27/21 7:41 AM, P Falth wrote:
> > When a char
> > is read from the input buffer it is immediately converted to lower case
> > ( a simple $20 OR instead of a UTF8 conversion )
>
> For ascii 91 thru 95 I get something that looks wrong using $20 OR.
> What am I missing?
>
> Btw, it made no difference in the unique words count in the
> kjvbible.txt file. Maybe that's why you used it.
>
> Thanks.
> -Doug
>
> : go ( c --) dup emit $20 or emit ;
> 91 go [{
> 92 go \|
> 93 go ]}
> 94 go ^~
> 95 go _

'$20 OR' always sets the 5th bit which is why he chose lower case.

I use :-

: ?upper ( c -- f ) 'A' 'Z' 1+ within ;
: ?lower ( c -- f ) 'a' 'z' 1+ within ;

: toupper ( c -- c ) dup ?lower if 32 xor then ;
: tolower ( c -- c ) dup ?upper if 32 xor then ;

: test 127 32 do cr i . space i emit space i toupper emit loop ; test
: test 127 32 do cr i . space i emit space i tolower emit loop ; test

On Saturday, 7 August 2021 at 14:28:09 UTC+2, Doug Hoffman wrote:
> On 3/27/21 7:41 AM, P Falth wrote:
> > When a char
> > is read from the input buffer it is immediately converted to lower case
> > ( a simple $20 OR instead of a UTF8 conversion )
>
> For ascii 91 thru 95 I get something that looks wrong using $20 OR.
> What am I missing?
>
> Btw, it made no difference in the unique words count in the
> kjvbible.txt file. Maybe that's why you used it.

Yes I checked that and found that it worked

Peter

>
> Thanks.
> -Doug
>
> : go ( c --) dup emit $20 or emit ;
> 91 go [{
> 92 go \|
> 93 go ]}
> 94 go ^~
> 95 go _

On 07/08/2021 13:58, NN wrote:
> On Saturday, 7 August 2021 at 13:28:09 UTC+1, Doug Hoffman wrote:
>> On 3/27/21 7:41 AM, P Falth wrote:
>>> When a char
>>> is read from the input buffer it is immediately converted to lower case
>>> ( a simple $20 OR instead of a UTF8 conversion )
>>
>> For ascii 91 thru 95 I get something that looks wrong using $20 OR.
>> What am I missing?
>>
>> Btw, it made no difference in the unique words count in the
>> kjvbible.txt file. Maybe that's why you used it.'A
>>
>> Thanks.
>> -Doug
>>
>> : go ( c --) dup emit $20 or emit ;
>> 91 go [{
>> 92 go \|
>> 93 go ]}
>> 94 go ^~
>> 95 go _
>
> '$20 OR' always sets the 5th bit which is why he chose lower case.
>
> I use :-
>
> : ?upper ( c -- f ) 'A' 'Z' 1+ within ;
> : ?lower ( c -- f ) 'a' 'z' 1+ within ;
>
> : toupper ( c -- c ) dup ?lower if 32 xor then ;
> : tolower ( c -- c ) dup ?upper if 32 xor then ;
>
> : test 127 32 do cr i . space i emit space i toupper emit loop ; test
> : test 127 32 do cr i . space i emit space i tolower emit loop ; test
>

Simmilar but a bit more efficient is

: up/down ( ch1 ch1 ch3 -- ch1|ch2 ) over + #26 u< bl and xor ;
: ch>lower ( ch1 -- ch1|ch2 ) [ 'A' negate ] literal up/down ;
: ch>upper ( ch1 -- ch1|ch2 ) [ 'a' negate ] literal up/down ;

: bounds ( ca u -- ca+u ca ) over + swap ;
: set-case ( ca u xt -- ) \ xt for ch>upper or ch>lower
-rot bounds ?do i c@ over execute i c! loop drop
;

--
Gerry