"Hamish Guthrie" <[EMAIL PROTECTED]> writes:
> Eric,
>
> I have just got back from a short absence and found that you had committed
> your code updates.
>
> Thanks for sorting out the start32.inc code - should be much better now.
Welcome.
> A few comments:
>
> Using a config file from the previous tree, I ran into some problems
> immediately, some of which I have tracked down, others of which I am still
> working on:
>
> 1. Makefile issue.
>
> This is a small issue, but may be a gotcha for some others, the config file
> name is now trated as an absolute pathname in the python command line for
> creating the Makefile etc, so whereas previously the following command:
>
> python ~/freebios/util/config/NLBConfig.py <configfile> ~/freebios
>
> created the Makefile correctly, this command now requires
>
> python ~/freebios/util/config/NLBConfig.py Absolutepath/<configfile>
> ~/freebios
>
> 2. Have you tested the code without SMP support? The tree seems to be broken
> where there is no SMP support required.
Not really. I've been a little busy. But I'll try and get the fixes
checked in a timely manner. Of what I did there shouldn't be much
that is more than compile deep.
> For sure in include/smp/atomic.h, the define for atomic_dec is wrong - well
> atomic_inc is defined twice.
Ugh. Catching this stuff is really appreciated.
I am sort of looking for a good way to conditionally compile code without
having #ifdef XXX all over the source files.. I think I can do this in the
Makefile so that is my target for start_stop.c More than anything
that is why I haven't checked in a fix for it yet.
My real choice with the code was to check it in when I did or let it
diverge even farther from the main tree, and cause even worse havoc
then I have :(
> I have corrected this, but then there are a whole lot of other compile
> issues now - I will keep you posted.
Updates are good. A patch file produced with diff -u is also easy.
Eric
