I made a diff for the following issues:

- use simpler {__make} DESTDIR=%{buildroot} install instead of %makeinstall
- don't package .la files
- add %post/postun -p /sbin/ldconfig
- cleaner deffatr
- add %doc, especially licence
- don't own %{includedir}
- add version-release 
- in devel Requires:       %{name} = %{version}-%{release}

One thing remains to be corrected: the summary for the devel package is too long.

Created attachment 120994 [details]
patch for the spec file with cleanings

OK, updated.

Spec Name or Url: http://paula.comtv.ru/chmlib.spec
SRPM Name or Url: http://paula.comtv.ru/chmlib-0.37.4-1.src.rpm

Please consider adding -q to %setup.

(In reply to comment #4)
> Please consider adding -q to %setup.

Done.

Spec Name or Url: http://paula.comtv.ru/chmlib.spec
SRPM Name or Url: http://paula.comtv.ru/chmlib-0.37.4-1.src.rpm

A couple warnings you want to tell upstream about, like "warning: "memcpy"
redefined"


- rpmlint's happy except for some no-documentation from the devel package
- package meets naming guidelines
- package meets packaging guidelines
- license (LGPL) OK, text in %doc, matches source
- spec file legible, in am. english
- source matches upstream
- package compiles on FC4 i386
- no missing BR
- no unnecessary BR
- no locales
- not relocatable
- owns all directories that it creates
- no duplicate files
- permissions ok
- %clean ok
- macro use consistent
- code, not content
- no need for -docs
- nothing in %doc affects runtime
- no need for .desktop file
- devel package ok
- no .la files
- post/postun ldconfig ok
- devel requires base package n-v-r 


One change, the BuildRoot needs to be cleaned at the beginning of %install, like
%clean. Do this and this is APPROVED.

Failed build for PPC arch.

I haven't access to the PPC-computer, so I can't find out what's wrong exactly.

Log:

http://buildsys.fedoraproject.org/logs/fedora-development-extras/1868-chmlib-0.37.4-1.fc5/ppc/build.log

Error was generated by #error-directive at that string "src/chm_lib.c:186". If
we look there, we'll find this piece of code:

=========================================================

/* i386, 32-bit, Windows */
#ifdef WIN32
typedef unsigned char           UChar;
typedef __int16                 Int16;
typedef unsigned __int16        UInt16;
typedef __int32                 Int32;
typedef unsigned __int32        UInt32;
typedef __int64                 Int64;
typedef unsigned __int64        UInt64;

/* I386, 32-bit, non-Windows */
/* Sparc        */
/* MIPS         */
/* PPC          */
#elif __i386__ || __sun || __sgi || __ppc__
typedef unsigned char           UChar;
typedef short                   Int16;
typedef unsigned short          UInt16;
typedef long                    Int32;
typedef unsigned long           UInt32;
typedef long long               Int64;
typedef unsigned long long      UInt64;

/* x86-64 */
/* Note that these may be appropriate for other 64-bit machines. */
#elif __x86_64__
typedef unsigned char           UChar;
typedef short                   Int16;
typedef unsigned short          UInt16;
typedef int                     Int32;
typedef unsigned int            UInt32;
typedef long                    Int64;
typedef unsigned long           UInt64;

#else

/* yielding an error is preferable to yielding incorrect behavior */
#error "Please define the sized types for your platform in chm_lib.c"
#endif

=========================================================

Looks like wrong arch-definition at that place:

> #elif __i386__ || __sun || __sgi || __ppc__

What arch-macro defined by gcc at PPC-box? __ppc32__? __ppc64__?

Instead of using architecture defines, why don't you just #include <stdint.h>  
and then typedef uint32_t UInt32; ? 

Peter, I did answer your question on the mailing list, although the broken list
Reply-To: settings mean that you didn't receive a direct copy of my mail -- it's
'__powerpc__'. 

However, Thomas's suggestion is the correct answer. Please make it use standard
C99 types, instead of the mess of hackish ifdefs.

> Peter, I did answer your question on the mailing list, although the broken list
> Reply-To: settings mean that you didn't receive a direct copy of my mail -- it's
> '__powerpc__'. 

Thanks. Builds fine now.

> However, Thomas's suggestion is the correct answer. Please make it use standard
> C99 types, instead of the mess of hackish ifdefs.

I'll send e-mail to the developer of chmlib ASAP.

Peter,

Could you build chmlib for FC-4 and FC-3?
The FC-3 and FC-4 CVS branches can be created via the Wiki page: 
http://fedoraproject.org/wiki/Extras/CVSSyncNeeded

Thanks,
jpo

Peter,

As I didn't get any answer to my previous post I took the liberty of building
chmlib for FC-4.

Best regards,
jpo

Hi,

Is someone planning to submit and maintain one of these CHM viewers?

* kchmviewer
  http://www.kchmviewer.net/

* xchm
  http://xchm.sourceforge.net/

* kchm
  http://kchm.sourceforge.net/

Regards,
jpo

John, you should assign that bug to yourself.

Thanks, taking.

Package Change Request
======================
Package Name: chmlib
New Branches: EL-4 EL-5
Updated Fedora Owners: lemenkov,pertusus
Updated EPEL Owners: pertusus

CVS done.

Package Change Request
======================
Package Name: chmlib
Updated EPEL Owners: lemenkov,pertusus

cvs done.