[parted-devel] Parted on GNU Hurd based systems

Debarshi 'Rishi' Ray debarshi.ray at gmail.com
Wed Mar 14 17:46:37 CET 2007


> > But 'struct fdasd_anchor' is a struct defined in include/parted/fdasd.h.

> Then I wonder why there was such an #if directive in the first place.
> Maybe some of the types used in that struct definition are s390-specific?

I will look into it.

> BTW, what is your goal in excluding them?
> If you insist on excluding them, be sure that you're
> not also excluding them from the distribution tarball (the one
> created by "make dist" run on an x86 Linux system) -- unless
> that is the intent, in which case I'd say you should just remove
> them altogether.

I just do not want to install the headers on a system which does not
have __s390__ or __s390x__ defined, since the associated functions
would not be compiled into the library. This can cause some confusion
if the headers are installed. However I surely want them to be on the
distribution tarball. Am I doing something which you think is contrary
to the goal?

> Makefiles must have TABs in some places, but not in definitions
> like that one, so you could write it like this (which I prefer,
> e.g., since changing the variable name length doesn't require changing
> all of the following lines):
>
>  partedinclude_HEADERS = \
>    gnu.h           \
>    linux.h         \
>    constraint.h    \
>    natmath.h       \
>    timer.h         \
>    unit.h          \
>    parted.h        \
>    vtoc.h          \
>    fdasd.h         \
>    parted.h        \
>    $(S390_HEADERS)
>
> The problem with inconsistent indentation like the above is that
> it makes diffs look bad.  Sometimes very bad.

I get the point, but making this change will require changing
everything here. :-) I will just use tabs and let it be reformatted
later.

Happy hacking,
Debarshi
-- 
GPG key ID: 63D4A5A7
Key server: pgp.mit.edu



More information about the parted-devel mailing list