Re: md5 version 2.0.4 to 2.0.5

Clint wrote:
On Jan 30, 3:59 pm, Eric Hassold <hass...@xxxxxxxxxxx> wrote:
Hi,

Clint wrote :



All,
I have noticed that since md5 has went to version 2.0.5 that I have
seen different results from previous versions of md5. I have older
apps that use md5 version 2.0.4 and if I upgrade md5 to 2.0.5., the
values return from the ::md5::md5 -hex command are different. For
example:
md5 v. 2.0.4:
-> ::md5::md5 -hex whatupwitdat
39054E44B13795EBC148CF17FE5ED22D
md5 v. 2.0.5
-> ::md5::md5 -hex whatupwitdat
2C9DEC74AE2AA10452D8820C117C6A4D
Anyone know if there is some new switch in version 2.0.5 for backward
compatible, or do I have to go back to all my old apps and rework if I
upgrade to version 2.0.5?
thanks,
md5sum is strictly defined, and must not change, neither from one
version of md5 package to another, nor from tcllib's md5 implementation
to other implementation. Correct digest for "whatupwitdat" is "3905...",
so what you get with 2.0.5 is broken. However, using md5 2.0.5 from
tcllib 1.10, I can't reproduce the problem you report, and get correct
answer. So it looks like your md5 distrib is broken. Which platform are
you running? Also, check if your md5 package falls back to one of the
accelerator (critcl, cryptkit or trf) instead of the pure Tcl
implementation.

Eric

The 2.0.4 version I am using is on a Linux box using tcl 8.4.12. The
2.0.5 version I am using is on a Mac (Tiger) using tcl 8.4.16.

Looks like have a bad copy of the 2.0.5. I get the 3905... answer with 2.0.5 in both 8.4.7 and 8.5.0 on Mac (Intel 10.5.1).

--
+--------------------------------+---------------------------------------+
| Gerald W. Lester |
|"The man who fights for his ideals is the man who is alive." - Cervantes|
+------------------------------------------------------------------------+
.

Relevant Pages

  • Re: md5 version 2.0.4 to 2.0.5
    ... Clint wrote: ... I have noticed that since md5 has went to version 2.0.5 that I have ... Correct digest for "whatupwitdat" is "3905...", so what you get with 2.0.5 is broken. ... Also, check if your md5 package falls back to one of the accelerator instead of the pure Tcl implementation. ...
    (comp.lang.tcl)
  • Re: Native md5 and base64 functions ?
    ... If you looked a bit closer at the tcllib code you would have found the ... md5 and some other time critical packages in tcllib. ... critcl installed and build/install the tcllib on that machine you get cC ... So the existence of a pure Tcl version of those algorithms isn't insane, ...
    (comp.lang.tcl)
  • Re: Native md5 and base64 functions ?
    ... If you looked a bit closer at the tcllib code you would have found the ... md5 and some other time critical packages in tcllib. ... So the existence of a pure Tcl version of those algorithms isn't insane, ...
    (comp.lang.tcl)
  • Re: Message digest MD5 implementation
    ... I use an MD5 implementation from http://www.sawatzki.de ... Eyal Post ... "Ivo Bauer" wrote in message ...
    (borland.public.delphi.thirdpartytools.general)
  • Re: Humble Contribution
    ... "Brian Ross" writes: ... do you know how md5 works? ... modifying an md5 implementation to set the magic constants from your ...
    (sci.crypt)