Newbie question - string buffers

ramman345

Hey guys, I'm making a transition from the Javelin and I'm curious about the best way to do string buffers on the Propeller. The manual says to use the DAT block to define buffers but I'm just not sure on the exact syntax to declare a string buffer, then to append to the buffer. Also, anybody got a good "indexOf" string search?

Josh

Mike Green

A string is just a byte array on the Propeller where the string normally has a zero termination byte (which you have to supply).

To append to an existing string in the array "X", you'd first find the current length of the string using STRSIZE(@X), then copy the new string to the end of the existing one like:

L := STRSIZE(@X) ' Get length of existing piece
N := STRSIZE(@Y) ' Get length of new piece
BYTEMOVE(@X+L,@Y,N+1) ' Copy new piece plus its terminator

Note that I've included no checking for too long a result string.

ramman345

Thanks a lot Mike, this clears things up.

One more question, how do you append a constant string to that? Suppose you want to append "World" to string X.

Josh

Mike Green

Since you would know the length of the new piece, you could write

BYTEMOVE(@X+STRSIZE(@X),STRING("World"),6)

ramman345

Mike, thanks so much for your help so far. I've decided to implement these functions in a StringBuffer object in order to easily replicate them throughout my program. Can you tell me if the following code looks correct to you:

VAR

byte buffer[noparse][[/noparse]buffer_size]
byte base_buffer[noparse][[/noparse]buffer_size]
byte cmp_buffer[noparse][[/noparse]buffer_size]

CON

buffer_size = 32

PUB clear | index
repeat index from 0 to buffer_size-1
buffer[noparse][[/noparse]index] := 0

PUB append(strptr) : Success
if strsize(strptr) > buffer_size 'Supplied string is too big for buffer
return -1

if strsize(strptr) > (buffer_size - strsize(@buffer)) 'Not enough room in the buffer for string
return -1

bytemove(@buffer+strsize(@buffer), @strptr, strsize(@strptr) + 1)

return 1

PUB toUpper(strptr) | index 'Convert each character to uppercase ASCII
repeat index from 0 to strsize(@strptr) - 1
if strptr[noparse][[/noparse]index] != 0
strptr[noparse][[/noparse]index] -= 32

PUB indexOf(strptr) : strindex | index
bytemove(@base_buffer, @buffer, strsize(@buffer) + 1)
bytemove(@cmp_buffer, @strptr, strsize(strptr) + 1)

toUpper(base_buffer)
toUpper(cmp_buffer)

strindex := -1 'Init strindex to -1 for failed search

repeat index from 0 to strsize(@base_buffer) 'Slide cmp_buffer through base_buffer
if strcomp(@basebuffer + index, @cmp_buffer)
strindex := index 'Store the current index
return strindex

return strindex

Post Edited (ramman345) : 5/29/2008 7:48:27 PM GMT

Mike Green

For the "clear" method, you only need to set the first byte to zero. You don't need to clear the whole array.

For the "append" method, you don't need to test separately for the size of the string to be added. Just a test for the total length of the result is enough. Also, it would help to compute the current length of the buffer string only once and save it in a local variable. Similarly, compute the length of the new piece only once and save it in a local variable. It might be useful to return the total length of the result (or -1 if an error occurs). Note that strptr is already a pointer. You'd have:

PUB clear
   buffer[noparse][[/noparse]0]~

PUB append(strptr) | old, new
   old := strsize(@buffer)
   new := strsize(strptr)
   if (result := old + new) => buffer_size
      return -1
   bytemove(@buffer+old, strptr, new+1)

PUB address
   result := @buffer

Post Edited (Mike Green) : 5/29/2008 7:44:15 PM GMT

ramman345

Thanks Mike, guess I was just covering my bases by clearing every element. Care to advise on these two (and too):

PUB toUpper(strptr) | index 'Convert each character to uppercase ASCII
repeat index from 0 to strsize(@strptr) - 1
if strptr[noparse][[/noparse]index] != 0
strptr[noparse][[/noparse]index] -= 32

PUB indexOf(strptr) : strindex | index
bytemove(@base_buffer, @buffer, strsize(@buffer) + 1)
bytemove(@cmp_buffer, @strptr, strsize(strptr) + 1)

toUpper(base_buffer)
toUpper(cmp_buffer)

strindex := -1 'Init strindex to -1 for failed search

repeat index from 0 to strsize(@base_buffer) 'Slide cmp_buffer through base_buffer
if strcomp(@basebuffer + index, @cmp_buffer)
strindex := index 'Store the current index
return strindex

return strindex

Mike Green

First of all, it looks like you're using C operators and this is Spin and they're different. Try

PUB toUpper(strptr) | t
repeat while t := byte[noparse][[/noparse]strptr++]
   if t =< "a" and t => "z"
      byte[noparse][[/noparse]strptr-1] -= 32

For the "indexOf" method, you want to check first to see if a match is possible based on length.

ramman345

Ok but I'm not sure which operators you are referring to as C, and I have a question about your function. Going through your version:

PUB toUpper(strptr) | t
repeat while t := byte[noparse][[/noparse]strptr++] <- cast whatever's at strptr++ as a byte and assign to t
if t =< "a" and t => "z" <- make sure t is a character between a-z ** I was missing this condition
byte[noparse][[/noparse]strptr-1] -= 32 <- strptr was already incremented above so we need the previous element

How does it get out of the loop?

ramman345

Ok, so I see that != is not equal in C and it should be <> in Spin. But I'm still not sure how the loop exits, wouldn't it keep going through memory?

Josh

Jimmy W.

ramman345 said...


How does it get out of the loop?

ramman345 said...

repeat while t := byte[noparse][[/noparse]strptr++]

as long as the condition evaluates to not zero it will be true so:
strings end with a zero terminator so when it hits the end it stops

ramman345

And the light slowly flickers on...

Thanks so much Jimmy!

Josh