[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#695152: t-p-u pre-approval lftp/4.3.6-1+deb7u1



On Mon, Dec 10, 2012 at 10:44:43PM +0100, Julien Cristau wrote:
> On Tue, Dec  4, 2012 at 18:21:25 +0100, Ivo De Decker wrote:
> 
> > diff -Nru lftp-4.3.6/debian/patches/lftp_sys-stdint-kfreebsd.patch lftp-4.3.6/debian/patches/lftp_sys-stdint-kfreebsd.patch
> > --- lftp-4.3.6/debian/patches/lftp_sys-stdint-kfreebsd.patch	1970-01-01 01:00:00.000000000 +0100
> > +++ lftp-4.3.6/debian/patches/lftp_sys-stdint-kfreebsd.patch	2012-12-01 17:42:44.000000000 +0100
> > @@ -0,0 +1,18 @@
> > +Description: prevent sys/_stdint.h from redefining intptr_t et.al.
> > +Author: Andreas Henriksson <andreas@fatal.se>
> > +Bug-Debian: http://bugs.debian.org/677861
> > +Forwarded: no
> > +
> > +--- lftp-4.3.6.orig/lib/stdint.in.h
> > ++++ lftp-4.3.6/lib/stdint.in.h
> > +@@ -75,6 +75,10 @@
> > +      _@GUARD_PREFIX@_STDINT_H is defined.
> > +      The include_next requires a split double-inclusion guard.  */
> > + # @INCLUDE_NEXT@ @NEXT_STDINT_H@
> > ++#ifdef __FreeBSD_kernel__
> > ++// prevent sys/_stdint.h from being included and redefine intptr_t et.al.
> > ++#define _SYS__STDINT_H_
> > ++#endif
> > + #endif
> > + 
> > + #if ! defined _@GUARD_PREFIX@_STDINT_H && ! defined _GL_JUST_INCLUDE_SYSTEM_STDINT_H
> 
> That is rather horrible.  I probably don't want to know why this is
> necessary.  Reluctant ack, go ahead with the upload.

Sorry for sharing some insight despite you not wanting to know.... ;P

I fully agree with it being horrible, that's why I didn't want to upload
it myself. On the other hand, since the maintainer acked Ivos NMU it's
not my place to disagree either.

A much better solution would have been to actually fix the kFreeBSD
system headers! (The details are available in the original lftp bug report
referenced in the patch headers.) Unfortunately noone seems interested in
doing that. On the other hand, I guess it could also be considered a
bug in portability layers like "gnulib" to not correctly handle the
brokenness (even after kFreeBSD headers potentially gets fixed) so
someone should work with upstream gnulib to find a suitable longterm-
maintainable fix! (Completely blocking a system header isn't very nice.
Even though that header today just contain redundant crap, that could
potentially change in the future.)

This patch is a from my point of view a double workaround. Work around
kFreeBSD brokenness and work around the release teams need to punish
all (Linux) users of lftp with a removal because of kFreeBSD brokenness.
I would not want to carry a patch like this in a package I maintain.
I hope noone will blame me later for this patch, because please remember
that I did not upload it and take no responsability for it.

-- 
Andreas Henriksson


Reply to: