> > Jakub wants to have a single routine do it all, that is why. [...]
> I guess we can use the same technique used for exception handling, ie. have
> one more exception section (or put it into the same??) and mark in there
> routines which want to do exception handling. [...]
yep ... this was one of the original ideas: to extend the 'exception prone
code section' with 'extra information'.
This 'extra information' is equivalent to a compile-time constant passed
to a single instance of function.
Why i didnt implement it in csum_partial_copy_fromuser(): IMHO, we are
trapped by our own concept: we are playing tricks with the compiler to
implement a nonexistant compiler feature, and there seems to be no return
:*( It's getting very ugly IMHO. I dont know where the right point is to
wait for compiler support to emerge. If you think it's all cool enough,
then fine with me ;)
-- mingo