User Activity

  • Posted a comment on discussion GnuCOBOL on GnuCOBOL (formerly OpenCOBOL)

    Not quite there yet. The HMAC code is in, but not verified. Mostly it was a evening of code cleanup passes (removing the tons of debug code put in before figuring out the init vector reset). Lots of size_t vs int issues too. size_t kinda gets in the way with GnuCOBOL. It's 8 bytes on some platforms but not all and 64-bit isn't the safest of tests. For these trials (on this machine) binary-c-long support works well, but that will need to be made a compile time conditional. size_t also fights with...

  • Modified a comment on discussion Help getting started on GnuCOBOL (formerly OpenCOBOL)

    Well the objective it set padder to 16 and slide the memory size up by 16, so the snippets are equivalent. padder was zero by if, and add will set it to 16 while increasing the core-len calculated buffer size up by 16. And I'm taking the two votes for clarity as a good sign. :-) That was the form that ended up in the code. The post came up while doing a callgcry cleanup pass after riddling the source with debug lines while trying to figure out why the early test of decrypt had 16 bytes of trash at...

  • Posted a comment on discussion Help getting started on GnuCOBOL (formerly OpenCOBOL)

    Well the objective it set padder to 16 and slide the memory size up by 16, so the snippets are equivalent. padder was zero by if, and add will set it to 16 while increasing the core-len calculated buffer size up by 16. And I'm taking the two votes for clarity as a good sign. :-) That was the form that ended up in the code. The post came up while doing a callgcry cleanup pass after riddling the source with debug lines while trying to figure out why the early test of decrypt had 16 bytes of trash at...

  • Posted a comment on discussion Help getting started on GnuCOBOL (formerly OpenCOBOL)

    Simple question really, more opinion than technical. What you rather be faced with when reading code? *> padding byte will be the value of how many bytes of pad needed *> and there will always be padding bytes, if zero, add a block if padder equal zero then add AES256-BLOCK-SIZE to core-len move AES256-BLOCK-SIZE to padder end-if clarity *> padding byte will be the value of how many bytes of pad needed *> and there will always be padding bytes, if zero, add a block if padder equal zero then add AES256-BLOCK-SIZE...

  • Posted a comment on discussion Help getting started on GnuCOBOL (formerly OpenCOBOL)

    Sorry, gnana, what link? I looked through PyPi and found python-cobol, but that links to a GitHub repo by by Rodrigo Prado which is a fork of the original by Paulus Schoutsen. Is the Pypi link the bestest link? I'll be making a FAQ entry. https://pypi.python.org/pypi/python-cobol/0.1.2 ?? That might come in very handy with FUNCTION PYTHON(...) as a loader from right in COBOL. Cheers, Brian

  • Posted a comment on discussion The Lounge on GnuCOBOL (formerly OpenCOBOL)

    Canada, and the world, lost Gord Downie today. 1964-2017. Incurable brain cancer. Armed with will and determination, and grace too The secret rules of engagement are hard to endorse when the appearance of conflict meets the appearance of force Rest in peace, Gordie, and thanks for the music and the poetry, you were ahead by a century. I quit drinking on New Year's day 1997, after seeing the Hip play downtown in Ottawa. I woke up hungover and couldn't remember what song they played for the encore....

  • Posted a comment on discussion GnuCOBOL on GnuCOBOL (formerly OpenCOBOL)

    Today's lesson: The HMAC, Hashed Message Authentication Code is one of the sacrosanct pieces of the crypto puzzle. The HMAC used in Tyler's example includes a scan over the salt, the IV (Initialization Vectors) and the ciphertext. Adding the salt and IV as pieces in an encrypted message does leak information to crackers, but not overly usable information, once protected by the HMAC. There are countless attack vectors for deriving the key used in encryption. The rabbit holes are deep, and there are...

  • Posted a comment on discussion GnuCOBOL on GnuCOBOL (formerly OpenCOBOL)

    Yayy bug hunting. That wasn't overly fun. Finally got the fluke clue from one of the other rare places that discusses using libgcrypt. Thanks to Burt Wagner, https://cboard.cprogramming.com/c-programming/105743-how-decrypt-encrypt-using-libgcrypt-arc4.html#post937511 The trick was resetting the Init Vector field attached to the cipher-handle. The IV field is modified during the encryption as part of the encryption. (Tyler's orginal C borks up too, but in subtly different ways. Should have tested...

View All

Personal Data

Username:
btiffin
Joined:
2012-04-09 05:04:42
Location:
Ottawa / Canada / EDT
Gender:
Male

Projects

Skills

  • Programming Language

Personal Tools