The HURD hacking guide

[Christophe A. Gevaudan]

Hello,
I just read the HURD hacking guide, which is for me a very nice
introduction into HURD programming.

I found a paragraph which was quite confusing to me: it's the one
which describes the function 'io_read()' from the sample program
'dump.c' in the section 6.1 "Writing a file to standard output".
Confusing was, that you refer to the definition in the file
$(HURD)/hurd/io_defs.h and state that there's one argument less
defined as you use (and not the last one, BTW): this wouldn't
work in (ANSI) C, as a mismatch between a function's prototype
and usage would be reported by the compiler.
In fact, the arguments are all in the definition file.
Maybe the io.h file shows another definition - but I couldn't
check this as I don't yet have a HURD system in place.
Anyway, I still doubt this works, when the additional argument
is placed somewhere else than at last.
Anyway, I'd like to suggest to expand this paragraph to include
some description about the definition file 'io_defs.h', the
generation of the header file 'io.h' and th ebehaviour of MiG
(which I suppose is used for that), to demonstrate how these
parts of HURD work together - this would be very helpfull for
HURD newbies like me - or at least you should refer to another
document which describes this already (with the same or a very
similar example).
I inserted some comments in the 'hhg.texi' file to help you
find the place I mean.

BTW: I also corrected two typos in the 'hhg.texi' file.

Please find attached the output of diff -u hhg.texi hhg.texi.orig.

Thanks for your great work

Best regards

Christophe
-- 
+-----------------------------------------------------+
| Christophe A. Gévaudan  --> mailto:gevaudan@acm.org |
+-----------------------------------------------------+

diff_-u_hhg.texi_hhg.texi.orig.output
Description: Binary data