I have a semi-working spec file for the Qt 3 based one (1.1.x). 
I did not check the Qt4 source package but I expect the build procedure to be
the same ( i.e. plain broken :) ) so if you still have not started to write the
spec maybe you can benefit from my previous work.

Do you need/want me to submit a formal review request?

I have start written the qsa spec file for Qt4.
But this is very difficult at this time without seen the final one for Qt4.
So I write it so, that it all way's work. But one feature of qsa must be disabled.
When you have an working spec file for qsa then you can post it here and do the
job. And I close this review. I think to modify it for Qt4 is not very hard when
you disable the editor feature then you can use your spec file for QSA 1.2.1
with Qt 4.1.2.
compiling works with configure -release -no-editor -prefix ...

FYI, qt4 review is bug 188180
I'd say base your qsa work off that, for now, at least.

As for qsa/qt3, qsa/qt4, I'd recommend submitting each as a separate review 
request (unless it's more convenient to build package both from a single 
specfile).  Once a spec/srpm is mentioned here, I can start a review.

Ok I do the work for qsa/Qt4 and Gianluca Sforna can do this for qsa/Qt3.
I have found possible a solution for the qt3support part problem.

So the first version of the spec file is ready.
http://prdownloads.sourceforge.net/qsmartcard/qsa.spec?download
As far the Qt4 package is ready I can remove the hard coded Qt path and the rest
of the work.

There's no need to hard-code the qt path, at least not in the qt4 case, since it
uses pkg-config.  Use this instead to define qtdir:

BuildRequires: qt4-devel
%global qtdir %(pkg-config --variable=prefix QtCore)

and %files should probably be like:
%files
...
%qtdir/lib/libqsa.so.*

%files devel
...
%qtdir/lib/libqsa.so

qsa-devel certainly will need 
Requires: %{name} = %{version}-%{release}
Requires: qt4-devel

Well, are there any particular reasons qt3 package cannot make its way into fc5
extras? At least two programs would benefit from it (LabPlot and museek). Greets.

Yes. For Qt3 you will need another version of QSA.

So this BR is for qt4 qsa. Will have to wait till fc6 then.

You can build an package for QSA on Qt3 when you will. But you must call it then
qsa3 or something like.

So after the new version of the Qt package is out here the first public version
of  the package.
http://prdownloads.sourceforge.net/qsmartcard/qsa-1.2.1-3FC4.src.rpm?download

so the next version is ready for review.
here the URL:
http://prdownloads.sourceforge.net/qsmartcard/qsa-1.2.1-4FC4.src.rpm?download
Changelog:
- enable qt3support
- Prevent for compiling using Qt >=4.2 witch will need another version of QSA.
- add missing .h files

NEEDSWORK:
* qsa.spec (and probably) qsa's configure script assumes
QTLIB=%qtdir/lib
QTINC=%qtdir/include
which isn't necessarily true.  In particular, on x86_64, QTLIB=%qtdir/lib64. 
Further, future iterations of qt4 may move QTLIB=%_libdir and/or QTINC=%_includedir.

specfile patch forthcoming.

Created attachment 130748 [details]
fix QTLIB,QTINC

fix QTLIB,QTINC
use install -p
only libqsa.so -> -devel, others in main pkg

NEEDSWORK
* Source's should be full URL:
Source:       
ftp://ftp.trolltech.com/qsa/source/qsa-x11-opensource-%{version}.tar.gz

Thanks but I have found an Bug in the Qt packae the variable  for the include
path of the QtTest is wrong it show to %{prefix}/include but it must show
%{prefix}/include/QtTest.
So the line qtinc %(pkg-config QtTest --variable=includedir) must be
qtinc %(pkg-config QtCore --variable=prefix)/include

Thanks.  I only chose QtTest because the --includedir output *was* prefix/include.

I suppose we could do something not-so-clean-but-still-should-work like:
%define qtinc %(pkg-config QtCore --variable=includedir)/..

Here the next version:
http://prdownloads.sourceforge.net/qsmartcard/qsa-1.2.1-5FC5.src.rpm?download

It'll speed the review process a bit too if you could also post links to the 
specfile as well.  That way, I won't have to redownload the entire source over 
and over again.

%files list is still wrong.  In particular, only libqsa.so should be 
in -devel, easiest template to use is something like:

%files 
%{qtlib}/lib*.so.*

%files devel
%{qtlib}/lib*.so

ok here the URL of the spec file.
http://prdownloads.sourceforge.net/qsmartcard/qsa.spec?download
it is every try the same.

And
URL:     ftp://ftp.trolltech.com/qsa/source/qsa-x11-opensource-1.2.1.tar.gz
Source:  qsa-x11-opensource-%{version}.tar.gz

should be:
URL:    http://www.trolltech.com/products/qt/
Source: 
ftp://ftp.trolltech.com/qsa/source/qsa-x11-opensource-%{version}.tar.gz

the the next please::)
http://prdownloads.sourceforge.net/qsmartcard/qsa-1.2.1-6FC5.src.rpm?download
http://prdownloads.sourceforge.net/qsmartcard/qsa.spec?download

Personally, I'd feel more comfortable if you used (something like):
%define qtinc %(dirname $(pkg-config QtCore --variable=includedir))
instead of
%define qtinc %(pkg-config QtCore --variable=prefix)/include

In case we ever do decide to move it out of the qt-prefix (to say, %_includedir).

But that's your call as maintainer.  

Yes this can be useful.
So here the new one:
http://prdownloads.sourceforge.net/qsmartcard/qsa-1.2.1-7FC5.src.rpm?download
http://prdownloads.sourceforge.net/qsmartcard/qsa.spec?download
rpmlint say that the doc part of the devel package are not nice.:(
Or can we ignore this? 

Learned something cool today about qmake, you may want to consider this
cleaner/simpler method of determining QTDIR, QTINC, QTLIB:

%define qtdir %(qmake4 -query QT_INSTALL_PREFIX)
%define qtinc %(qmake4 -query QT_INSTALL_HEADERS)
%define qtlib %(qmake4 -query QT_INSTALL_LIBS)

re: rpmlint
A couple of things:
1.  make sure %defattr is *first*, so %files should be:
%files
%defattr(-,root,root)
...

2.  why in -devel %defattr(755,root,root) instead of %defattr(-,root,root)?

3.  Re: wrong-file-end-of-line-encoding, --hidden-file-or-dir warnings.  IMO,
these are small-potatoes.  You can fix these if you want, but it's not a blocker.

Ok now the package with the qmake use.:)
http://prdownloads.sourceforge.net/qsmartcard/qsa-1.2.1-8FC5.src.rpm?download
http://prdownloads.sourceforge.net/qsmartcard/qsa.spec?download

Change
%files devel
%defattr(0755,root,root)
%attr(0644,root,root)%{qtinc}/*
...

To
%files devel
%defattr(0644,root,root)
%{qtinc}/*
...

And we're *real* close.  (unless you have a good reason for doing it your way).

I'll recheck all the blocker guidelines tomorrow morning...

Looks like the ACCEPT got lost in the bugzilla crash, reclosing.

(In reply to comment #29)
> Change
> %files devel
> %defattr(0755,root,root)
> %attr(0644,root,root)%{qtinc}/*
> ...
> 
> To
> %files devel
> %defattr(0644,root,root)
> %{qtinc}/*
> ...

Rex, just for the record, it should have been "%defattr(0644,root,root,755)". 
See bug #214577. I imagine that it was a typo.

Heck, I *should* have simply suggested
%defattr(-,root,root)

Package Change Request
======================
Package Name: qt4-qsa
New Branches: F-7

Because the F-7 directory is missing in the CVS repo.

found it:)

Package Change Request
======================
Package Name: qt4-qsa
Remove Branches: devel

It can be removed for F-9 and higher, because it have reach it's EOL.
And the script function is available since Qt 4.3.
So it will be only need until F-8, to get old code working and give the
developers  some time to adapt her code to Qt 4.3.

Please follow the package end of life procedure for the devel branch:
http://fedoraproject.org/wiki/PackageMaintainers/PackageEndOfLife

done.
dead.package file add and send the mail to rel-eng