I'm new to the list & to python. I did look, but to no avail, in the
source distribution for information on the following, so I hope this
isn't something that's in a FAQ I haven't found.
I am trying to compile 0.9.9 on several platforms (rs_aix32,
sun4m_412, & next_mach30). I've been able to build the rs_aix32
preliminary interpreter using gcc-2.4.5 without much difficulty (some
extra preprocessor defines & some "#ifdef _AIX"). I did, however,
have problems compiling the md5 module. I have RFC-1321 & have
successfully built & tested mddriver.
Have I completely missed something while configuring or are the
following actually errors/oversights/typos?
- the Makefile tells the user that md5 code is available in "md5.doc"
rsa.com. At least as of this writing, the file is
"//rsa.com/pub/md5.txt". This file is actually just RFC-1321.
Perhaps reference should also be made to RFC-1321 as RFCs are
readily available (and I assume this one is as well).
- the Makefile states md5c.c is needed. The RFC contains md5c.c.
This is also true of the object file (which the Makefile gives
as md5.o).
- global.h from the md5 code is not #included (necessary for UINT4)
- md5_digest calls MD5Final with only one argument (MD5_CTX
mdContext)
- also in md5_digest, there is a reference to the "digest" member of
mdContext (declared as above). MD5_CTX in my versions of RFC-1321,
md5.txt, and RSAREF does not contain a digest member.
- the "/* Really static, forward */" declarations in md5module.c,
mpzmodule.c, & Fontobject.c (and probably other files which I
didn't use) do not work. Gcc creates the .o but the AIX ld
says the symbols are not defined. AIX cc complains about MD5type
(and the other types for other modules) being redeclared. I
had to '#ifndef _AIX' the 'static' modifier.
I did manage to fixup md5module.c so the rsademo.py script
apparently succeeded (only output was "Generating small primes
list...") on the rs_aix32 machine.
comments welcomed,
michael