md5 object problem

Bobb Fwed

I can't get the MD5 object to work, there are no samples or notes on how to use it, so this is what I am trying (looks correct anyway):

CON
  MAX_STR_LEN = 25

OBJ
  MD5 : "MD5"

VAR
  byte nd[noparse][[/noparse]MAX_STR_LEN]
  byte nstr[noparse][[/noparse]MAX_STR_LEN]

PUB do_hash
  bytemove(@nd, string("abcdefghijklmn 1234567890"), MAX_STR_LEN)
  MD5.hash(@nd, MAX_STR_LEN, @nstr)

and what I get in nstr is: one line down, and 11 or so in: 8vA_ Ұ
not, the 16 byte string I was expecting.
I've gotten other weird results with other input strings. What am I doing wrong?

Mike Green

VAR
   byte nd[noparse][[/noparse]MAX_STR_LEN]
   byte nstr[noparse][[/noparse]MD5#HASH_LENGTH]

PUB do_hash | str
   str := string("abcdefghijklmn 1234567890") ' Save the address of the constant string
   bytemove(@nd,str,strsize(str)) ' Use the actual length of the string
   MD5.hash(@nd,strsize(str),@nstr) ' Again use the actual length

The do_hash call will create a 16 byte hash in nstr. Note that this hash is not in displayable form.
It consists of 16 bytes with values from 0 to 255. You have to display the bytes in hexadecimal or
decimal form to make any sense of them. There are hexadecimal display routines in most of the
display objects (serial.hex(<value>,<#digits>) for example).

Bobb Fwed

My problem wasn't the coding (so much) but the return, I have always just seen printable versions of an MD5 checksum, now it makes more sense, thanks.

Bobb Fwed

so I display it with

REPEAT MD5#HASH_LENGTH
    DEBUG.hex(nstr[noparse][[/noparse]i++], 2)

It seems to return something legit, but it doesn't match the MD5 I know works (on PSPad and PHP).
Prop MD5: B1963AC8BCB9905FE513D45A98633210
PSPad MD5: 63d7f5da7e0c01fec14ca1cf9016541f

any ideas?

Post Edited (Bobb Fwed) : 10/6/2008 5:49:46 PM GMT

Bobb Fwed

Also, the last bytes from the Prop are always $32, $10, and the one before those two seems to usually end in 3
Something doesn't seem right.

Post Edited (Bobb Fwed) : 10/6/2008 5:51:26 PM GMT

hippy

Did you change your code to how Mike suggested ? If not it would be calculating MD5 using data other than it should be.

Bobb Fwed

Yes I did.

Farmer Scott

I'm not sure I understand why, in Mike's comment, he changed the routine to assign a string value to str, and then used bytemove to copy str to nd.

I've done it both ways, and see the bytemove is required to get the correct hash, just don't understand why.

Can anyone elaborte?

Thanks,

Scott

Bobb Fwed

This works without the move.

CON

  _CLKMODE      = XTAL1
  _XINFREQ      = 5_000_000                          ' 5MHz Crystal
  
OBJ

  DEBUG  : "FullDuplexSerial"
  MD5    : "MD5"

VAR

  byte nstr[noparse][[/noparse]MD5#HASH_LENGTH]

PUB Main

  DEBUG.start(31, 30, 0, 57600)
  waitcnt(clkfreq + cnt)  
  DEBUG.tx($0D)

  do_hash

PUB do_hash | str, i
  str := string("abcdefghijklmn 1234567890")
  MD5.hash(str, strsize(str), @nstr)

  i~
  REPEAT MD5#HASH_LENGTH
    DEBUG.hex(nstr[noparse][[/noparse]i++], 2)
  DEBUG.tx($D)

But this is probably what should be used.

CON

  _CLKMODE      = XTAL1
  _XINFREQ      = 5_000_000                          ' 5MHz Crystal

  MAX_STR_LEN = 64
  
OBJ

  DEBUG  : "FullDuplexSerial"
  MD5    : "MD5"

VAR

  byte nd[noparse][[/noparse]MAX_STR_LEN]
  byte nstr[noparse][[/noparse]MD5#HASH_LENGTH]

PUB Main

  DEBUG.start(31, 30, 0, 57600)
  waitcnt(clkfreq + cnt)  
  DEBUG.tx($0D)

  do_hash

PUB do_hash | str, i
  str := string("abcdefghijklmn 1234567890")
  bytemove(@nd, str, strsize(str))
  MD5.hash(@nd, strsize(str), @nstr)

  i~
  REPEAT MD5#HASH_LENGTH
    DEBUG.hex(nstr[noparse][[/noparse]i++], 2)
  DEBUG.tx($D)

Both output the correct value: "63D7F5DA7E0C01FEC14CA1CF9016541F"

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
April, 2008: when I discovered the answers to all my micro-computational-botherations!