wrong md5 checksum

[kkwweett]

Hi,

the download page (http://www.python.org/download/releases/3.0/) for
Python 3000 shows :

9119625244b17aa35ed94b655a2b2135 13491200 python-3.0b1.msi

but I got

9119625244a57aa35ed94b655a2b2135 13491200 python-3.0b1.msi

(44a57 rather than 44b17 )

with several md5 checkers ( MdChecker (http://getmd5checker.com/) or a C
prog I've written)

what's wrong ?

[kkwweett]

kkwweett a écrit :

Hi,

the download page (http://www.python.org/download/releases/3.0/) for
Python 3000 shows :

9119625244b17aa35ed94b655a2b2135 13491200 python-3.0b1.msi

but I got

9119625244a57aa35ed94b655a2b2135 13491200 python-3.0b1.msi

(44a57 rather than 44b17 )

with several md5 checkers ( MdChecker (http://getmd5checker.com/) or a C
prog I've written)

Wrong md5 checksum even with Python2.5 :
import md5
m=md5.new()
message=open("python-3.0b1.msi","rb").read()
m.update(message)
m.hexdigest()

Answer : '9119625244a57aa35ed94b655a2b2135'
Does anyone know what can be done to decide wether the file is corrupted
or the checksum given is wrong ?

[Peter Pearson]

On Tue, 01 Jul 2008 17:54:05 +0200, kkwweett <kk******@hotmail.frwrote:

kkwweett a écrit :

>Hi,

the download page (http://www.python.org/download/releases/3.0/) for
Python 3000 shows :

9119625244b17aa35ed94b655a2b2135 13491200 python-3.0b1.msi

but I got

9119625244a57aa35ed94b655a2b2135 13491200 python-3.0b1.msi

(44a57 rather than 44b17 )

[snip]

Does anyone know what can be done to decide wether the file is corrupted
or the checksum given is wrong ?

For what it's worth, md5sum (GNU coreutils) 5.93 agrees with
your result.

This interesting almost-agreement could result from

A. an error in the software that computed the value for the web page,
B. a transcription error in putting the checksum on the web page,
C. accidental corruption of the file, or
D. a deliberate attempt to substitute a file with a similar MD5 sum.

Possibility C is unlikely, but would be the most exciting
development in cryptology this year. If C turns out to be the
explanation, I hope someone will make sure sci.crypt gets told.
Possibility D would be awfully interesting, too.

--
To email me, substitute nowhere->spamcop, invalid->net.

[kkwweett]

>
A. an error in the software that computed the value for the web page,
B. a transcription error in putting the checksum on the web page,
C. accidental corruption of the file, or
D. a deliberate attempt to substitute a file with a similar MD5 sum.

It seems that the answer is probably B. But, according to the release
website manager, a mystery remains :

http://mail.python.org/pipermail/pyt...ly/014281.html

[Terry Reedy]

kkwweett wrote:

>

>>
A. an error in the software that computed the value for the web page,
B. a transcription error in putting the checksum on the web page,
C. accidental corruption of the file, or
D. a deliberate attempt to substitute a file with a similar MD5 sum.

It seems that the answer is probably B. But, according to the release
website manager, a mystery remains :

http://mail.python.org/pipermail/pyt...ly/014281.html

Marcin Kowalczyk pointed out in a follow-up to the above that the
replacement of 'a5' with 'b1' could have been part of a global
search/replace that would have been correct everywhere else on the page.
Thus case C. Barry remembers doing some such.

Lesson. Take care with blind global replacements.