offset broken?

[J. Landman Gay]

On 2/22/14, 1:35 PM, Richmond wrote:
> Sorry: Neanderthal question time:
>
> I cannot see anything about charsToSkip in the inbuilt (i.e. in the IDE)
> documentation (6.6. dp1).

It's the third parameter in the "offset" function. See "offset" in the 
dictionary.

> I set up a boring little stack with a fld "fORIGIN" conatinong "The
> quick brown fox jumps over the lazy dog"
>
> and that was jolly nice,
>
> then I set up another fld called "fREZ"
>
> and a button with this script:
>
> on mouseUp
>     put empty into fld "fREZ"
>     put fld "fORIGIN" into ORIGIN
>     put offset("o",ORIGIN) & ", " after fld "fREZ"
>     put offset("o",ORIGIN) & ", " after fld "fREZ"
>     put offset("o",ORIGIN) & ", " after fld "fREZ"
> end mouseUp
>
>
> in the hope I would get something like "13, 18, 27," in fld "fREZ"
>
> but all I got was "13, 13, 13"
>
> so I really cannot see what the utility of 'offset' is at all.

Oh ye of little faith. That's what we've been talking about; every 
instance of "offset" starts at the beginning of the string unless you 
include the number of characters to skip as the third parameter.

This is a much faster way to get all instances of an offset in a block 
of text, exactly parallel to the differences between "repeat with x" and 
"repeat for each". Like "for each," offset will read from the current 
marked location rather than counting from the top during each iteration, 
provided you include the third parameter. Your example handler doesn't 
use the parameter, so of course you're getting the same offset repeatedly.

If there is a third parameter, then offset will start counting from that 
point, and return the number of characters from that location in the 
text to the next instance of the character. It's up to you to add that 
number to the current "skip" location in order to keep the marker moving 
along.

This should do what you want:

function getAllOffsets pData,pChar -- find all instances of a char
   put 0 into tSkip -- start at beginning
   repeat
     put offset(pChar,pData,tSkip) into tNextOffset -- chars since skip 
point
     if tNextOffset = 0 then exit repeat -- no more found
     add tNextOffset to tSkip -- move the marker
     put tSkip & comma after tRes -- store it
   end repeat
   return tRes
end getAllOffsets

This should be much faster than your "repeat" example (which counts 
characters from the top during each iteration) or a version that deletes 
characters from the original data. Try some comparison timings with a 
very large block of text.

-- 
Jacqueline Landman Gay         |     jacque at hyperactivesw.com
HyperActive Software           |     http://www.hyperactivesw.com