[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: libdiskfs: build system inconsistencies; the Hurd `collecting box'
From: |
Marcus Brinkmann |
Subject: |
Re: libdiskfs: build system inconsistencies; the Hurd `collecting box' |
Date: |
Tue, 07 Feb 2006 18:36:14 +0100 |
User-agent: |
Wanderlust/2.14.0 (Africa) SEMI/1.14.6 (Maruoka) FLIM/1.14.7 (Sanjō) APEL/10.6 Emacs/21.4 (i486-pc-linux-gnu) MULE/5.0 (SAKAKI) |
At Tue, 7 Feb 2006 10:17:53 -0500,
Thomas Schwinge <tschwinge@gnu.org> wrote:
>
> Hi!
>
> libdiskfs #includes pager.h (libpager). If pager is not contained in
> HURDLIBS, then the system-wide file from /include will be used and not
> the one from the currently-being-built Hurd tree. That won't be an issue
> if both of them adhere to the same API, which they didn't for me, as I
> was building a pre-ogi ext2fs, but the system had the usual Debian
> GNU/Hurd ext2fs installed, containing Ognyan's API changes.
>
> I'd suggest to go for one of the two following possibilities: either we
> implement a reliable mechanism that makes the Hurd's build system not use
> the system-wide header files (how to do that; only for the Hurd's own
> header files, but not for glibc, etc.) or we split the whole Hurd package
> into small packages--libpager, libdiskfs, libstore, libihash, ...,
> ext2fs, ...--that _only_ use the system-wide header files (many packages;
> how to combine them reasonably).
The Hurd build tree should use the build-local header files when available.
Thanks,
Marcus