StrComp question

Mikael S · 2012-01-11 14:42

Hello!

I try to compare a string recieved from my gsm modem with a given DAT byte, but without luck. I think the problem is that the recieved string does not contain a terminating zero.
Can someone please give me a example how i can add that zero to my recieved string, or what can it be??

Thanks!

/Micke

JonnyMac · 2012-01-11 14:47

It's a pretty simple matter of replacing the terminator from your original string with a zero, then doing the comparison. If you provide a code snippet of what you're doing that doesn't work, it will be easier to show you a specific solution.

Mikael S · 2012-01-11 15:03

Thanks for your fast answer!
Here comes the code:

var

 byte message[15], phonenumber[16], mess1[16], mess2[16], mess3[16], mess5[16], mess6[16], mess7[16]

dat

   on byte "on", 0
   off byte "off", 0

pub main
   
debug.start (115200)
gsm.start (26, 27, 0, 9600)
waitcnt (clkfreq * 2 + cnt)

dira[2] :=1

readmessage

pub readmessage 


repeat  

  waitcnt (clkfreq * 5 + cnt)
  gsm.rxflush
  gsm.str (string("AT+CMGR=1")) 'Read sms in first memory
  gsm.tx(13)                    'CT
  
  debug.str (string("Get message in memory1..."))

  debug.newline
  gsm.RxStrTime(300,@mess1)
  gsm.RxStrTime(300,@mess2)
  gsm.RxStrTime(300,@mess3)
  gsm.RxStrTime(300,@phonenumber) 
  gsm.RxStrTime(300,@mess5)
  gsm.RxStrTime(300,@mess6)
  gsm.RxStrTime(300,@mess7)
  gsm.RxStrTime(200,@message)
  waitcnt (clkfreq * 1 + cnt)
  gsm.str(string("AT+CMGD=1"))      'Delete message in first memory
  gsm.tx(13)
  waitcnt (clkfreq * 1 + cnt)


     if strcomp (@on, @message)
        on2 
     elseif strcomp (@off, @message)
        off2

  debug.str(string("no message, repeat"))

Duane Degn · 2012-01-11 15:15

Something I've found useful is to do a bytefill prior to gathering data.

bytefill(@message, 0, 127)

I think I added up 127 bytes in your buffers.

JonnyMac · 2012-01-11 15:57

I've used Duane's trick, too. What you do is have a buffer that is at least one byte longer than any string you expect to receive. You use bytefill to fill that array with zeros and then receive your new string into it. So long as your string is in fact shorter than the buffer used, you will now have a zero terminator. This is probably the best route as you're using a canned routine and don't want to change it.

T Chap · 2012-01-11 17:42

If you have to receive the data each time before the compare, another option is to use rxtime, and specify a time that will allow the receive code block to time out after the last byte. After the timeout, the return value is -1, so you can create a test in the read loop that says that if it gets a -1 from a time out, the current and position is set to 0. Once a -1 is received. quit the loop.

Example using a 4port serial object:

PUB ReadData   | ii, val    'copy content of incoming message to an array
     ii := 1
     response := ser.rxtime(2, 1000)  serial port 2, waits here for first byte response for 1 second
     If response > -1   ' if something was received, get the rest of the data now
      Repeat 511                    'look for more, else return
        Val :=  ser.rxtime(2, 2)   'read more, wait only 2 ms if no byte found
        If val == -1                     'if the wait timed out with no data
          response[ii] := 0      'set this array position to a 0
          quit             'quit, return, or abort...  no more reading 
        response[ii] := Val   ' write to the array if data was received
        ii++

Mikael S · 2012-01-12 10:31

Thanks everyone for your answers!

It doesent work for me...
I have increased the "var byte message" to 127, and i have put the "bytefill(@message, 0, 127" just before the gsm.RxStr.
What is it im doing wrong, can someone please help me?!

/Micke

T Chap · 2012-01-12 10:51

Do you already know how many bytes are in the message reply? Is it possible it is more than 127?

Mikael S · 2012-01-12 11:01

no, the reply is just two characters.

T Chap · 2012-01-12 11:05

Have you tried to print to screen the content of message to be sure that you do have the correct characters in the string before doing the compare?

Mikael S · 2012-01-12 11:13

yes, i have. I deleted that part, just to shorten my post.

T Chap · 2012-01-12 11:17

Try getting rid of the elseif, just use another if

Mikael S · 2012-01-12 11:30

I really appreciate your help!

Now i have tried that, same result...

Here is my latest code, the message consists just of the two letters "on".

[code]
obj

gsm: "Extended_FDSerial"
debug: "Parallax Serial Terminal"
var

byte message[127], phonenumber[16], mess1[16], mess2[16], mess3[16], mess5[16], mess6[16], mess7[16]

dat

on byte "on", 0
off byte "off", 0

pub main

debug.start (115200)
gsm.start (26, 27, 0, 9600)
waitcnt (clkfreq * 2 + cnt)

dira[2] :=1

readmessage

pub readmessage

repeat

waitcnt (clkfreq * 5 + cnt)
gsm.rxflush
gsm.str (string("AT+CMGR=1")) 'Read sms in first memory
gsm.tx(13) 'CT

debug.str (string("Get message in memory1..."))
bytefill(@message, 0, 127)
debug.newline
gsm.RxStrTime(300,@mess1)
gsm.RxStrTime(300,@mess2)
gsm.RxStrTime(300,@mess3)
gsm.RxStrTime(300,@phonenumber)
gsm.RxStrTime(300,@mess5)
gsm.RxStrTime(300,@mess6)
gsm.RxStrTime(300,@mess7)
gsm.RxStrTime(200,@message)
waitcnt (clkfreq * 1 + cnt)
gsm.str(string("AT+CMGD=2")) 'Delete message in first memory
gsm.tx(13)
waitcnt (clkfreq * 1 + cnt)
debug.newline
debug.str(@mess1)
debug.newline
debug.str(@mess2)
debug.newline
debug.str(@mess3)
debug.newline
debug.str(@phonenumber)
debug.newline
debug.str(@mess5)
debug.newline
debug.str(@mess6)
debug.newline
debug.str(@mess7)
debug.newline
debug.str(string("H

T Chap · 2012-01-12 11:48

Try this code instead:

   if message[0] == "o"
      if message[1] == "n"
           outa[2] :=  1     'assumed to be an LED???
   if message[0] == "o"
      if message[1] == "f"
        if message[2] == "f"
           outa[2] := 0

Mikael S · 2012-01-12 12:13

No luck with this either....

T Chap · 2012-01-12 12:19

Ok since I am on a wild goose chase, please try this and post the result:

debug.str(string("H

Mikael S · 2012-01-12 12:57

Here comes the result...
(I cant copy and paste from Serial Terminal, i hope this will do)

T Chap · 2012-01-12 13:50

I have tested this and it works, not sure why your code is not working.

var 
   byte test1[3]

pub 
   test1[0]  := "o"
   test1[1]  := "n"
   test1[2]  := 0
   if strcomp(@test1, @test2)
     dosomething
dat
    test2 byte "on", 0

The only last thing I might try for some understanding is to look at each element in the var:
debug.str(@message[0])
debug.newline
debug.str(@message[1])
debug.newline
debug.str(@message[2])
debug.newline

Hopefully someone will enlighten us both.

John Abshier · 2012-01-12 14:20

This works for me.

CON
        _clkmode = xtal1 + pll16x                                               'Standard clock mode * crystal frequency = 80 MHz
        _xinfreq = 5_000_000

VAR
  byte Str2[30]
   
OBJ
  SIO      : "FullDuplexSerialPlus"
  
Pub Main
  waitcnt(clkfreq * 4 + cnt)
  SIO.StartPST(115200)
  repeat
    SIO.GetStr(@Str2)
    SIO.Str(string("Str2 is: "))
    SIO.Str(@Str2)
    SIO.tx(13)
    if strcomp(@Str1, @Str2)
      SIO.Str(string("Str1 and Str2 are equal",13,13))
    else
      SIO.Str(string("Str1 and Str2 are different",13,13))
          
DAT
Str1 byte "Hello World", 0

Results from Parallax Serial Terminal

Str2 is: Fred
Str1 and Str2 are different

Str2 is: Hello World
Str1 and Str2 are equal

John Abshier

kuroneko · 2012-01-12 18:11

The output from post #17 looks odd (date/time for example, why is it split when double quotes are used?). The default delimiter used with RxStrTime is a comma. So assuming date and time are delimited by a comma (not unreasonable) then they will be returned over two separate calls to RxStrTime. To check this theory please add the following call somewhere after gsm.start in your main method:

gsm.SetDelimiter(13)   ' do not use comma, CR is hardwired into RxStrTime so no harm re-using it

Also, you should increase the 15 character limit for RxStrTime in the Extended_FDSerial object. Not doing so will split longer messages after 15 characters which may upset your (hard-wired) receive sequence (e.g. date and time won't fit into a 15 character buffer assuming they come as one line).

Mikael S · 2012-01-12 22:22

The output is ok in my eyes. Im just need to recieve three different messages, on, off, and status. So the buffersize is ok. Im only intressted in the message and the phonenumber for the reply message.
My problem is why i cant parse the incomming message with the actual three given words.

kuroneko · 2012-01-12 22:40

OK, what does the uninterpreted message look like? After

debug.str (string("Get message in memory1..."))
  bytefill(@message, 0, 127)
  debug.newline

add the following loop

repeat
    if (c := gsm.rx) < 32
      debug.char("<")
      debug.hex(c, 2)
      debug.char(">")
    else
      debug.char(c)

This is simply to verify whether the command actually arrives.

Mikael S · 2012-01-13 00:58

Now i have tried that, i get a strange result!
But in my code i have the debug.str(@message) and that give me the right result from the string (on).

Another question, is there any way to copy the result from the Serial Terminal, without using the print screen?

kuroneko · 2012-01-13 01:10

Mikael S wrote: »

Now i have tried that, i get a strange result! Another question, is there any way to copy the result from the Serial Terminal, without using the print screen?

What's your _clkmode setup like? If you don't have any please add some (115k2 is not reliable at RCFAST). If you don't have a crystal try a lower baudrate.

CON
  _clkmode = XTAL1|PLL16X
  _xinfreq = 5_000_000

As for the rest, it looks like you get CRLF line endings from the gsm device (<0D><0A>). Unfortunately the string reader only uses CR as a delimiter. Which means your command will be preceded by the remaining LF. This needs filtering out otherwise strcomp will never match.

copy/paste is possible. Just select the area with the mouse then press ctrl+C. Once you have your clock setup sorted can you generate another data dump?

Apologies for using the wrong method names. PST uses char instead of tx (I modified the code fragment posted earlier).

Mikael S · 2012-01-13 08:49

I have that clockmode that you refer to.

Mikael S · 2012-01-13 13:42

OK, i think i have a <LF> in the begining of my string, how can i get rid of that?
Or is there any other solution to read a "part" of a string without knowing the start bit??

Mike G · 2012-01-13 13:58

Let's see if this example helps

CON
   _clkmode = xtal1 + pll16x
   _xinfreq = 5_000_000

DAT
  message     byte  " Some stuff", $0A, "My received string", $00
  compareTo   byte  "My received string", $00 
  
VAR
  Byte buffer[3]


OBJ
  pst : "Parallax Serial Terminal"
 

PUB Main | i, isMatch
  pst.start(115200)
  Pause(500)

  i := 0
  isMatch := false

  ' Find the string that starts after an <lf>
  'Then compare the string to "compareTo" 
  repeat strsize(@message)
    if(message[i++] == $0A)
      isMatch := strcomp(@message[i], @compareTo)

  'Print the results
  pst.str(string("Have a match: "))
  if(isMatch)
    pst.str(string("True", $0D))
  else
    pst.str(string("False", $0D))
         

PRI Pause(Duration)  
  waitcnt(((clkfreq / 1_000 * Duration - 3932) #> 381) + cnt)
  return

T Chap · 2012-01-13 14:55

Would searching within the string (instring) find the "on" if there is a LF at the front? I can't test it right now. I included Jon's string methods.

pub instr(str1, str2) | len1, len2, pos, idx

'' Returns position of str2 in str1
'' -- if str2 not in str1 returns -1

  len1 := strsize(str1)
  len2 := strsize(str2)
  pos  := -1                                           
  idx  := 0

  if (len1 => len2)                                    
    repeat (len1 - len2 + 1)                           
      if (byte[str1] == 0)                             
        quit
      else
        if (strncmp(str1++, str2, len2) == 0)          
          pos := idx                                   
          quit                                         
        else
          ++idx

  return pos

Or, try this test:

PUB parsemessage  |  i
 i := 0
 repeat 4   '10 should handle 'on',  or repeat strsize(message) for crude method of repeats
   if message[i] == "o"
      if message[i+1] == "n"
           outa[2] :=  1     'assumed to be an LED???
   if message[i] == "o"
      if message[i+1] == "f"
        if message[i+2] == "f"
           outa[2] := 0
   i++

This would normally search within the string and find on or off, not sure how it reacts with LF as part of the string.

Mikael S · 2012-01-14 01:43

Thanks!
How can i add the <LF> in my DAT byte?
I have a "on byte "on", 0" in my DAT section, is there any way to add the <LF> command before the "on" ?
Just to test my theory...

kuroneko · 2012-01-14 01:48

Mikael S wrote: »

How can i add the <LF> in my DAT byte?

label   byte    10, "on", 0

That said, this feels wrong. But maybe that's just me

Mikael S · 2012-01-17 11:35

When i add the <LF> in my compare string everything works, thanks everyone!

Now to the next issue....
I use the code below to read the message from my modem, and it seems to work fine, for a while...
It seems like the prop hangs if the modem recieves a message exactly at the same time this code is executed.
If there is no message to read, i just use the time-out part in the RxStrTime, and its then this seems to happen.

[code]
gsm.rxflush
gsm.str (string("AT+CMGR=1")) 'Read sms in first memory
gsm.tx(13) 'CT

debug.str (string("Get message in memory 1..."))
bytefill(@message, 0, 127)
gsm.RxStrTime(300,@mess1) '

StrComp question

Comments