Is Requires:       automake for the -devel subpackage really needed? I guess not.

Does this really work? Normally --disable-static is used:
%configure --enable-static=no

Thanks

Fixed

May you look at them again.

SPEC: http://rakesh.fedorapeople.org/spec/gflags.spec
SRPM: http://rakesh.fedorapeople.org/srpm/gflags-0.9-2.fc9.src.rpm

[NOT OK] rpmlint output:
gflags-devel.i386: W: no-documentation
There is a html file in the doc/ dir of the tarball that contains
documenation, that would fit in the -devel subpackage imho.

[OK] Spec in %{name}.spec format

[OK] license allowed: BSD 3 clause
[OK] license matches shortname in License: BSD
[OK] license in tarball and included in %doc:
COPYING

[OK] package is code or permissive content:
[OK] Source0 is a working URL
[OK] Source0 matches Upstream:
bd4871398e9019b241d89cc21fb62def  gflags-0.9.tar.gz

[OK] Package builds on all platforms:
httsp://koji.fedoraproject.org/koji/taskinfo?taskID=776073

[OK] BuildRequires are complete (koji builds)
(OK) No file dependencies outside of /etc /bin /sbin /usr/bin /usr/sbin 

[OK] Every (sub)package containing libraries runs ldconfig
[OK] .h (header) files are in -devel subpackage
[OK] .a (static libraries) are in -static subpackage:
no static libraries are packaged
[OK] contains .so.X(.Y) files and .so is in -devel
[OK] -devel subpackage has Requires: %{name} = %{version}-%{release}
[OK] .la files (libtool) are not included

[OK] Prefix: /usr not used (not relocatable)
[OK] Owns all created directories
[OK] no duplicates in %files
[OK] %defattr(-,root,root,-) is in every %files section
[OK] Does not own files or dirs from other packages
[OK] included filenames are in UTF-8

[OK] %clean is rm -rf %{buildroot} or $RPM_BUILD_ROOT 
[OK] %install starts with rm -rf %{buildroot} or $RPM_BUILD_ROOT 
[OK] Consistent macro usage
[OK] %doc does not affect runtime

{OK} no pre-built binaries (.a, .so*, executable)
{OK} well known BuildRoot
%{_tmppath}/%{name}-%{version}-%{release}-root-%(%{__id_u} -n)
{OK} PreReq not used
{OK} no duplication of system libraries
{OK} no rpath
{NOT OK} Timestamps preserved with cp and install
https://fedoraproject.org/wiki/Packaging/Guidelines#Timestamps
You should change the make install to:
make install DESTDIR=$RPM_BUILD_ROOT INSTALL="install -p"

{OK} Uses parallel make (%{?_smp_mflags})
{OK} Requires(pre,post) style notation not used
{OK} only writes to tmp /var/tmp $TMPDIR %{_tmppath} %{_builddir} (and %{buildroot} on %install and %clean)
{OK} no Conflicts
{OK} nothing installed in /srv
{OK} Changelog in allowed format
<OK> Sane Provides: and Requires:

{OK} Follows Naming Guidelines

Other stuff:
- this line should be removed, because no static libraries are build:
find $RPM_BUILD_ROOT -type f -name "*.a" -delete

- Can you maybe submit an patch to upstream that changes the installation of the
  header files within the automake stuff or file a bug? Then you could remove
  these lines:
mkdir -p $RPM_BUILD_ROOT%{_includedir}/%{name}
mv $RPM_BUILD_ROOT%{_includedir}/google/*.h $RPM_BUILD_ROOT%{_includedir}/%{name}/

- On my system no rpath is set in the libraries, why did you use chrpath?
- Why don't you install the python stuff?
- not the NOT OK for rpmlint and timestamps above

There is a unittest running script in the python directory, maybe you can integrate it into the spec somehow (you can use a section named %check). I tried to just run it in %check and it failed, but without using rpmbuild it did not. Not sure, what can be done here.

I have reported upstream, about header files not going to proper place and an extra folder being created which doesn't make sense. I assume it as a cosmetic issue and non-blocking one.

Other all issues are taken care.

1. rpmlint & timestamp
2. python module & unittest
3. removed useless stuff

SPEC: http://rakesh.fedorapeople.org/spec/gflags.spec
SRPM: http://rakesh.fedorapeople.org/srpm/gflags-0.9-3.fc9.src.rpm

> Can you maybe submit an patch to upstream that changes the installation of
the
>  header files within the automake stuff or file a bug? Then you could remove
>  these lines:
>mkdir -p $RPM_BUILD_ROOT%{_includedir}/%{name}
>mv $RPM_BUILD_ROOT%{_includedir}/google/*.h
>$RPM_BUILD_ROOT%{_includedir}/%{name}/

I had a discussion with upstream and had a look at some other packages. It seems to me that it is wrong to move header files to some other place as it may break some other useful stuff.
  Few other packages have similar structure and they don't move header files.

I will like to keep them in place.

Will update ASAP.

Thanks,

SPEC: http://rakesh.fedorapeople.org/spec/gflags.spec
SRPM: http://rakesh.fedorapeople.org/srpm/gflags-0.9-4.fc9.src.rpm

Updated.

Thanks,

@Till

ping
Waiting impatiently ;)

ping Again?

Thanks Till

I have fixed the issues. I have disabled test suite because 2 out of 17 where failing on x86_64 on rawhide.

Package build on F9 and rawhide on most archs -
http://koji.fedoraproject.org/koji/taskinfo?taskID=803522

SPEC: http://rakesh.fedorapeople.org/spec/gflags.spec
SRPM: http://rakesh.fedorapeople.org/srpm/gflags-0.9-5.fc10.src.rpm

It failed on F-8 only - because the Python eggs are not created in Fedora 8 by default.

Will update shortly.

Thanks,

Updated:

SPEC: http://rakesh.fedorapeople.org/spec/gflags.spec
SRPM: http://rakesh.fedorapeople.org/srpm/gflags-0.9-6.fc10.src.rpm

Build on F-8 test-box also.
Thanks.

[OK] rpmlint output: silent
{OK} Timestamps preserved with cp and install

Other issues from comment 3: OK

(In reply to comment #10)

> I have fixed the issues. I have disabled test suite because 2 out of 17 where
> failing on x86_64 on rawhide.

Did you report this to upstream? Are you sure that the test suite is buggy and not the package itself, i.e. does it really work as intended on x86_64? There is also a binary testsuite that can be run with "make test", but it uses unsecure tempfiles: It always uses the same directories in /tmp to perform the test, which can maybe exploited.

I consider the testsuite issues not to be a blocker, therefore this package is now  APPROVED. Nevertheless, please try to get upstream to make the testsuite workable within Fedora builds.

I have reported it upstream and will be working on it upstream.

New Package CVS Request
=======================
Package Name: gflags
Short Description: Library for commandline flag processing
Owners: rakesh
Branches: F-8 F-9
InitialCC: rakesh

Did you mean to include yourself as both the owner and initialcc?  Because that's pointless.  I've just ignored that line; if there's some other user who wants to be CC'd, they can add themselves in pkgdb.

CVS done.

The build issue is fixed upstream and will include test suite in 1.0 version to be released soon.

Thanks

gflags-0.9-6.fc8 has been submitted as an update for Fedora 8.
http://admin.fedoraproject.org/updates/gflags-0.9-6.fc8

gflags-0.9-6.fc9 has been submitted as an update for Fedora 9.
http://admin.fedoraproject.org/updates/gflags-0.9-6.fc9

build and imported
Will update to latest 1.0 after its release in few days.

Thanks

gflags-0.9-6.fc8 has been pushed to the Fedora 8 stable repository.  If problems still persist, please make note of it in this bug report.

gflags-0.9-6.fc9 has been pushed to the Fedora 9 stable repository.  If problems still persist, please make note of it in this bug report.

Package Change Request
======================
Package Name: gflags
New Branches: el6
Owners: peter
InitialCC:

Git done (by process-git-requests).

Package Change Request
======================
Package Name: gflags
New Branches: epel7
Owners: peter ivaxer
InitialCC:

Git done (by process-git-requests).