Hi friends,
there is a script which generates encrypted passwords for Starship users. There is a series of marshal, zlib and base64 calls, which is reversed by the script.
Is there a known bug in Marshal, or should I start the debugger now? The passwphrase for the attached script is "hey".
cheers - chris
Christian Tismer wrote:
Hi friends,
there is a script which generates encrypted passwords for Starship users. There is a series of marshal, zlib and base64 calls, which is reversed by the script.
Is there a known bug in Marshal, or should I start the debugger now? The passwphrase for the attached script is "hey".
Aehmmm... can it be that code objects are no longer compatible between Python 2.0 and 2.1?
sigh - ciao - chris
Christian Tismer wrote:
Christian Tismer wrote:
Hi friends,
there is a script which generates encrypted passwords for Starship users. There is a series of marshal, zlib and base64 calls, which is reversed by the script.
Is there a known bug in Marshal, or should I start the debugger now? The passwphrase for the attached script is "hey".
Aehmmm... can it be that code objects are no longer compatible between Python 2.0 and 2.1?
Yes, not suprisingly though... AFAIK the pyc format changed in every single version between 1.5.2 and 2.1.
On Fri, Jun 01, 2001 at 09:01:38PM +0200, M.-A. Lemburg wrote:
Yes, not suprisingly though... AFAIK the pyc format changed in every single version between 1.5.2 and 2.1.
Worse, it's changed several times between each release :)
Thomas Wouters wrote:
On Fri, Jun 01, 2001 at 09:01:38PM +0200, M.-A. Lemburg wrote:
Yes, not suprisingly though... AFAIK the pyc format changed in every single version between 1.5.2 and 2.1.
Worse, it's changed several times between each release :)
But I didn't use .pyc at all, just a marshalled code object. There are no version headers or such. The same object worked in fact for Py 1.5.2 and 2.0, but no longer with 2.1 . I debugged the unmarshalling and saw what happened: The new code objects with their new scoping features were the problem. The new structures were simply added, and there is no way to skip these for older code objects, since there isn't any info. Some option for marshal to umarshal old-style code objects would ave helped. But then, I'm not sure if the opcodes are still assigned the same way in 2.1, or if there was some movement? This would kill it anyway.
ciao - chris
(now looking for another cheap way to do something invisible in Python without installing *anything* )
Christian Tismer wrote:
Thomas Wouters wrote:
On Fri, Jun 01, 2001 at 09:01:38PM +0200, M.-A. Lemburg wrote:
Yes, not suprisingly though... AFAIK the pyc format changed in every single version between 1.5.2 and 2.1.
Worse, it's changed several times between each release :)
But I didn't use .pyc at all, just a marshalled code object.
That's the point: the header in pyc files is meant to signal the incompatibility of the following code object. Perhaps we should moev this version information into the marshal format of code objects themselves...
There are no version headers or such. The same object worked in fact for Py 1.5.2 and 2.0, but no longer with 2.1 . I debugged the unmarshalling and saw what happened: The new code objects with their new scoping features were the problem. The new structures were simply added, and there is no way to skip these for older code objects, since there isn't any info. Some option for marshal to umarshal old-style code objects would ave helped. But then, I'm not sure if the opcodes are still assigned the same way in 2.1, or if there was some movement? This would kill it anyway.
AFAIK, the assignments did not change, but several opcodes were added in 2.1, so code compiled in 2.1 will no run in 2.0.
ciao - chris
(now looking for another cheap way to do something invisible in Python without installing *anything* )
Why don't you use freeze or py2exe or Gordon's installer for these one file executables ? Alternatively, you should check the Python version and make sure that it matches the one used for compiling the byte code.
participants (3)
-
Christian Tismer -
M.-A. Lemburg -
Thomas Wouters