rpmlint message '%ifarch-applied-patch'

[Richard W.M. Jones]

What does this rpmlint message mean?  I don't understand why it's 
harmful to have patches only applying to particular architectures.

-----

ical.src: W: %ifarch-applied-patch Patch3: ical-2.2-ia64.patch
A patch is applied inside an %ifarch block. Patches must be applied
on all architectures and may contain necessary configure and/or code
patch to be effective only on a given arch.

-----

The patch in question is:

--- ical-2.2/calendar/uid.C.ia64	Wed May 29 22:51:03 1996
+++ ical-2.2/calendar/uid.C	Mon Apr 30 00:13:58 2001
@@ -13,7 +13,7 @@
  #ifndef HAVE_GETHOSTNAME_PROTO

  #ifdef linux
-extern "C" int gethostname(char*, unsigned int);
+extern "C" int gethostname(char*, unsigned long);
  #else /* !linux */
  extern "C" int gethostname(char*, int);
  #endif /* linux */

Rich.

-- 
Emerging Technologies, Red Hat - http://et.redhat.com/~rjones/
Registered Address: Red Hat UK Ltd, Amberley Place, 107-111 Peascod
Street, Windsor, Berkshire, SL4 1TE, United Kingdom.  Registered in
England and Wales under Company Registration No. 03798903

[Jindrich Novy]

Hi,

On Wed, Sep 19, 2007 at 01:04:18PM +0100, Richard W.M. Jones wrote:
>  What does this rpmlint message mean?  I don't understand why
> it's harmful 
>  to have patches only applying to particular architectures.
>
>  -----
>
>  ical.src: W: %ifarch-applied-patch Patch3: ical-2.2-ia64.patch
>  A patch is applied inside an %ifarch block. Patches must be applied
>  on all architectures and may contain necessary configure and/or code
>  patch to be effective only on a given arch.
>
>  ----- 
It likely means that you have a construct like this in a spec file:

%ifarch x86_64
Patch1: some.patch
%endif

what means that the patch is included to srpm only on the x86_64 arch.
This is wrong since the patch has to be included in sprm for all arches
even though it's applied only on one arch.

I had build problems related to this recently ;)

-- 
Jindrich Novy <jnovy(a)redhat.com&gt;   http://people.redhat.com/jnovy/

[Ralf Corsepius]

On Wed, 2007-09-19 at 14:10 +0200, Jindrich Novy wrote:
>  Hi,
>  
>  On Wed, Sep 19, 2007 at 01:04:18PM +0100, Richard W.M. Jones wrote:
>  > What does this rpmlint message mean?  I don't understand why it's harmful 
>  > to have patches only applying to particular architectures.
>  >
>  > -----
>  >
>  > ical.src: W: %ifarch-applied-patch Patch3: ical-2.2-ia64.patch
>  > A patch is applied inside an %ifarch block. Patches must be applied
>  > on all architectures and may contain necessary configure and/or code
>  > patch to be effective only on a given arch.
>  >
>  > -----
>  
>  It likely means that you have a construct like this in a spec file:
>  
>  %ifarch x86_64
>  Patch1: some.patch
>  %endif 
Besides this, this patch seems rather silly to me.

Instead of trying to mess around with local defs of gethostname(),
you should probably better "#include <unistd.h>".

Ralf

PS.:  __linux__ is the correct (POSIX-compliant) define to identify
linux.

[Richard W.M. Jones]

Ralf Corsepius wrote:
>  On Wed, 2007-09-19 at 14:10 +0200, Jindrich Novy wrote:
> > Hi,
> >
> > On Wed, Sep 19, 2007 at 01:04:18PM +0100, Richard W.M. Jones wrote:
> >> What does this rpmlint message mean?  I don't understand why it's harmful
>
> >> to have patches only applying to particular architectures.
> >>
> >> -----
> >>
> >> ical.src: W: %ifarch-applied-patch Patch3: ical-2.2-ia64.patch
> >> A patch is applied inside an %ifarch block. Patches must be applied
> >> on all architectures and may contain necessary configure and/or code
> >> patch to be effective only on a given arch.
> >>
> >> -----
> > It likely means that you have a construct like this in a spec file:
> >
> > %ifarch x86_64
> > Patch1: some.patch
> > %endif
>  
>  Besides this, this patch seems rather silly to me.
>  
>  Instead of trying to mess around with local defs of gethostname(),
>  you should probably better "#include <unistd.h>".
>  
>  Ralf
>  
>  PS.:  __linux__ is the correct (POSIX-compliant) define to identify
>  linux. 
This code is ancient history anyway.  Development on ical was 
_abandoned_ around 1994.

https://bugzilla.redhat.com/show_bug.cgi?id=296161

Rich.

-- 
Emerging Technologies, Red Hat - http://et.redhat.com/~rjones/
Registered Address: Red Hat UK Ltd, Amberley Place, 107-111 Peascod
Street, Windsor, Berkshire, SL4 1TE, United Kingdom.  Registered in
England and Wales under Company Registration No. 03798903

[Jakub Jelinek]

On Wed, Sep 19, 2007 at 01:04:18PM +0100, Richard W.M. Jones wrote:
>  What does this rpmlint message mean?  I don't understand why
> it's 
>  harmful to have patches only applying to particular architectures. 
It is harmful, because the patch isn't tested on all the other arches.
Also eventually if further patches are applied which touch the same
locations you end up with a twisty unmaintanable maze of per-arch
patches.

>  The patch in question is: 
That is very bogus patch.
1) gethostbyname has the second argument size_t, so if anything,
   you should patch it everywhere for the right type
2) you really should find out why configure doesn't find
   gethostname prototype on Linux, it is present in unistd.h
   in glibc on all arches, unless you request a namespace which
   forbids it in unistd.h.

>  --- ical-2.2/calendar/uid.C.ia64	Wed May 29 22:51:03 1996
>  +++ ical-2.2/calendar/uid.C	Mon Apr 30 00:13:58 2001
>  @@ -13,7 +13,7 @@
>   #ifndef HAVE_GETHOSTNAME_PROTO
>  
>   #ifdef linux
>  -extern "C" int gethostname(char*, unsigned int);
>  +extern "C" int gethostname(char*, unsigned long);
>   #else /* !linux */
>   extern "C" int gethostname(char*, int);
>   #endif /* linux */
>   
	Jakub