I've been playing with a shiny new 2.4.17 kernel with CML2 2.0, under 
Redhat 7.1. The python binary is python 1.5, so i changed the top line 
of autoconfigure.py, cmlcompile.py, cmlconfigure.py, configtrans.py and
kxref.py to use python2. Unfortunatly it didn't run quite properly, so
here are some fixes that seem to make it run for me. I don't know if
they are really bugs in the code, they seem to be quite obvious and in 
the way, but these fixed some of my troubles.

autoconfigure.py

Line 170, python complains about a continue statement embedded in a try
block. I replaced the continue with 'raise DIENOW', and changed the
matching except statement to 'except IOError,DIENOW:'. I can't say if
this realy works, but it eliminates the error message.

Line 563, execfile seems to never be defined anywhere. Replaced with
'scripts/autoconfigure.rules'. Is this really what's supposed to be
executed here?

autoconfigure.rules, line 2780. I can't find any symbol UHCI, so i'm
guessing this is USB_UHCI.

Second to last line in autoconfigure.py, I changed this to sys.argv[1],
otherwise it get's mashed in with the debuging info. I assume this will
be changed later anyways.

cmlsystem.py, line 444, errors about converting a list to a string.
After a bit of fun with the python interactive interpreter, this seems
to work.

          if changes:
              errorl = []
              errorl += self.loadcommit(freeze)
              for error_item in error_list:
                  errors += str(error_item)

I believe there might be a few rulebase bugs (LKM support configureable 
as a module?), but ill look at them a bit more first.


_______________________________________________
kbuild-devel mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/kbuild-devel

Reply via email to