Bug 219327

Builds fine in mock(FC-6).
* /usr/lib/kazehakase is not owned by the package
* the desktop file categories should be:
  Categories=Network;Application;X-Fedora;

(In reply to comment #1)
> Builds fine in mock(FC-6).
> * /usr/lib/kazehakase is not owned by the package
Oops... I will fix this.

> * the desktop file categories should be:
>   Categories=Network;Application;X-Fedora;
Both categories (Application, X-Fedora) are deprecated 
and warned or treated as error from 
desktop-fiele-utils >= 0.11 (FC-devel) and should not 
used any longer (this is applied to at FC5/6).

Uploaded.
http://www.ioa.s.u-tokyo.ac.jp/~mtasaka/dist/extras/development/SPECS/kazehakase.spec
http://www.ioa.s.u-tokyo.ac.jp/~mtasaka/dist/extras/development/SRPMS/kazehakase-0.4.3-2.src.rpm

* Wed Dec 13 2006 Mamoru Tasaka <mtasaka.u-tokyo.ac.jp> - 0.4.3-2
- Forgot to own %%{_libdir}/%%{name}, fixing.

I would appreciate it if someone would review this...

Package doesn't build in an ordinary user environment on FC6:

# rpmbuild -ba kazehakase.spec
...
libtool: compile: unable to infer tagged configuration
libtool: compile: specify a tag with `--tag'
make[4]: *** [GtkPromptService.lo] Error 1
make[4]: *** Waiting for unfinished jobs....
libtool: compile: unable to infer tagged configuration
libtool: compile: specify a tag with `--tag'

Also, I will not review nor approve any package which runs any of the autotools
during building, esp. if they are incorrecly applied, such as in this case
(Hint: aclocal, version of autoconf being used).

Well, would you attach the full build log, Ralf?
I cannot catch only from your comment.

I use rawhide, and I cannot check "ordinary" FC6 environ.
Note: mockbuild on FC6/devel i386 succeeds. also, for me
"ordinary" rpmbuild succeeds on FC-devel. 

And as this srpm "touch"es all files which are not 
needed to be regenerated, rebuilding this srpm should succeed.

Created attachment 144168 [details]
Mock build log of kazehakase 0.4.3-2 on FC6 i386

Well, it is true that mockbuild of 0.4.3-2 on FC6 i386
succeeds.

Maybe rebuilding failure is related with your settings
or something else?

Created attachment 144198 [details]
Log of build failure on FC6

(In reply to comment #7)
> Maybe rebuilding failure is related with your settings
> or something else?
Without having looked into details (I am scarce on time, ATM), my guess would be
the breakdown related to you running the autotools, so may-be something related to
* The package uses autoconf-2.60a, while FC6 has 2.59.
* My user environment also has autoconf/automake/libtool installed, a mock
environment doesn't.
* They are using autoconf-2.60a (a beta version), so they might also be using a
beta of libtool. rpm and/or you running autoconf can interfere.

(In reply to comment #9)
> (In reply to comment #7)
> > Maybe rebuilding failure is related with your settings
> > or something else?
Urgh, I have to apologize - This was a DUE (Dumb User Error) on my part. 
The machine I was building on didn't have g++ installed, sorry.

But ... after having corrected this ... the package still doesn't build:
...
kz-mozlauncher.cpp:69: error: prototype for 'nsresult
KzContentHandler::Show(nsIHelperAppLauncher*, nsISupports*, PRBool)' does not
match any in class 'KzContentHandler'
kz-mozlauncher.h:57: error: candidate is: virtual nsresult
KzContentHandler::Show(nsIHelperAppLauncher*, nsISupports*, PRUint32)
/usr/include/firefox-1.5.0.9/xpcom/nsCOMPtr.h: In instantiation of
'nsDerivedSafe<nsIHelperAppLauncher>':
kz-mozlauncher.cpp:81:   instantiated from here

(In reply to comment #10)
> But ... after having corrected this ... the package still doesn't build:

Well, again please attach your full build log and
config.h on failure rebuild?
I have only FC-devel i386 and I can check if this can be rebuild
on FC-6 only by mockbuild and only on i386.

My mockbuild log on FC-6 i386 is in comment #7.

Created attachment 144252 [details]
config.h of mockbuild 0.4.3-2 on FC6 i386 

I attach config.h created during mockbuild of 
kazehakase 0.4.3-2 on FC6 i386.

Created attachment 144253 [details]
config.log of mockbuild 0.4.3-2 on FC6 i386 

Samely, config.log

Umm, I checked firefox-devel-1.5.0.9-1.fc6, however,
it seems that this should succeed in rebuilding.

/usr/include/firefox-1.5.0.9/exthandler/nsIHelperAppLauncherDialog.h
expects that the last argument of KzContentHandler::Show should
be PRUint32
---------------------------------------------------
    65    NS_IMETHOD Show(nsIHelperAppLauncher *aLauncher, nsISupports
*aContext, PRUint32 aReason) = 0;
---------------------------------------------------
Then kazehakase-0.4.3/src/mozilla/kz-mozlauncher.cpp says:
---------------------------------------------------
    66  NS_IMETHODIMP KzContentHandler::Show(nsIHelperAppLauncher *aLauncher,
    67                                       nsISupports *aContext,
    68  #ifdef MOZ_NSIHELPERAPPLAUNCHERDIALOG_NSPRBOOL_
    69                                       PRBool aForced)
    70  #else
    71                                       PRUint32 aForced)
    72  #endif
---------------------------------------------------
  this means that MOZ_NSIHELPERAPPLAUNCHERDIALOG_NSPRBOOL_ should be
  undefined.
  When rebuilding my srpm, this value is correctly undefined (as
  written in spec file), so at least no error should occur at this
  point....

A bit clean up.

http://www.ioa.s.u-tokyo.ac.jp/~mtasaka/dist/extras/development/SPECS/kazehakase.spec
http://www.ioa.s.u-tokyo.ac.jp/~mtasaka/dist/extras/development/SRPMS/kazehakase-0.4.3-4.src.rpm

-------------------------------------------------

* Thu Jan 18 2007 Mamoru Tasaka <mtasaka.u-tokyo.ac.jp> - 0.4.3-4
- Do not call autoconf, just fix configure.

* Sat Dec 23 2006 Mamoru Tasaka <mtasaka.u-tokyo.ac.jp> - 0.4.3-3
- Add firefox version dependency for gecko engine
--------------------------------------------------

Hello.

I'll be happy to review this for you. Full review to come tomorrow...

According to its website, 0.4.4 is released. Would you please update your
spec/SRPM accordingly? Thanks.

(In reply to comment #17)
> According to its website, 0.4.4 is released.

Oh.. 0.4.4 seems to be released yesterday. Thanks.

http://www.ioa.s.u-tokyo.ac.jp/~mtasaka/dist/extras/development/SPECS/kazehakase.spec
http://www.ioa.s.u-tokyo.ac.jp/~mtasaka/dist/extras/development/SRPMS/kazehakase-0.4.4-1.src.rpm

-------------------------------------------------
* Tue Jan 30 2007 Mamoru Tasaka <mtasaka.u-tokyo.ac.jp> - 0.4.4-1
- 0.4.4

Sorry to take so long on this one. I've been quite a bit backlogged at work. :(
I really appreciate your patience. 

Soo.... here we go!

== Formal Review of kazehakase 0.4.4-1 == 

GOOD: rpmlint is silent on both the source and binary RPMs.
GOOD: The package follows the naming guidelines, and the spec file is named
accordingly ("%{name}.spec"). 
GOOD: BuildRoot is "%{_tmppath}/%{name}-%{version}-%{release}-root-%(%{__id_u}
-n)", following the packaging guidelines. 
GOOD: No duplicate BuildRequires are included; and all necessary BuildRequires
are listed.
GOOD: Included documentation (%doc) is OK.
GOOD: Package builds and runs against system copies of installed tools and
libraries; and does not include its own local copies thereof.
GOOD: Package includes an appropriate .desktop file since it is a graphical
application; and desktop-file-install is properly used to install it. A
BuildRequires: desktop-file-utils is also present.
GOOD: Macros are used instead of harcoded file names, and usage of these macros
(including $RPM_BUILD_ROOT) is consistent throughout the spec file.
GOOD: Locale files are handled correctly, using %find_lang.
GOOD: Package is not relocatable.
GOOD: Package includes appropriate code and content; and final directory and
file ownership is OK. Configuration files are marked appropriately (%config).
GOOD: Package does not own any system files/directories or any files/directories
that conflict with another package.   
GOOD: Package license (GPL) is OK; and a copy of it is included in the package
as documentation (%doc COPYING). The License field in the spec file properly
reflects this.
GOOD: Spec file is legible; and written in American English.
GOOD: The source tarball matches that of upstream:
  $ md5sum SOURCES/kazehakase-0.4.4*
  049fd40c238e6838bcdbe14c37cc9051  SOURCES/kazehakase-0.4.4-srpm.tar.gz
  049fd40c238e6838bcdbe14c37cc9051  SOURCES/kazehakase-0.4.4-upstream.tar.gz
GOOD: The package successfully builds in mock into x86 binary RPMs on both FC6
and devel/FC7.
GOOD: No duplicates are listed in the %files section; and its %defattr line is good.
GOOD: Package has an appropriate %clean section, which contains simply "rm -rf
$RPM_BUILD_ROOT"
GOOD: When installed, the application runs well with no apparent segfaults or
major bugs.
GOOD: Files are converted to UTF-8 encoding properly.
GOOD: Package uses %{?_smp_mflags} and honors %optflags compiler flags properly.
GOOD: Package contains no libtool archives (*.la files) 

N/A: No static libraries or rPath exclusions are needed.
N/A: Package is not a web application.
N/A: No ExcludeArch/ExclusiveArch tweaking is required.
N/A: Package installs no shared libraries into the standard $LIBDIR; thus
%post/%postun scriplets of /sbin/ldconfig are not needed.
N/A: No large (neither in size nor in quantity) documentation is included, thus
no -doc subpackage is needed.
N/A: No headers, no pkgconfig files, and no static or unsuffixed shared
libraries are included. Thus, no -devel subpackage is needed.
N/A: Package contains no %description or Summary translations.
N/A: Scriplets are not required for this package.


** FIXME: The only potential issue I see I see with this is the following in the
%configure output:
-------------------
checking for IceConnectionNumber in -lICE... no
checking for SmcSaveYourselfDone in -lSM... no
checking X11/SM/SMlib.h usability... no
checking X11/SM/SMlib.h presence... no
checking for X11/SM/SMlib.h... no
-------------------

Perusing through the source (kazahakse-0.4.4/src/kz-app.c), it seems that it can
optionally build with XSM (X session manager) support, which is probably a
desired feature. :) Adding libSM-devel to the BuildRequires enables this. If you
fix this, I'll approve the package for importing. 

Woo!

Thank you for first reviewing!!

(In reply to comment #19)
> ** FIXME: The only potential issue I see I see with this is the following in the
> %configure output:
> -------------------
> checking for IceConnectionNumber in -lICE... no
> checking for SmcSaveYourselfDone in -lSM... no
> checking X11/SM/SMlib.h usability... no
> checking X11/SM/SMlib.h presence... no
> checking for X11/SM/SMlib.h... no
> -------------------
Umm.. Why didn't I notice this? Bad... fixed.
And it seems that 0.4.4.1 is released two days ago.

http://www.ioa.s.u-tokyo.ac.jp/~mtasaka/dist/extras/development/SPECS/kazehakase.spec
http://www.ioa.s.u-tokyo.ac.jp/~mtasaka/dist/extras/development/SRPMS/kazehakase-0.4.4.1-1.src.rpm

---------------------------------------------------
* Fri Feb  2 2007 Mamoru Tasaka <mtasaka.u-tokyo.ac.jp> - 0.4.4.1-1
- 0.4.4.1

* Fri Feb  2 2007 Mamoru Tasaka <mtasaka.u-tokyo.ac.jp> - 0.4.4-2
- Add more BuildRequires: anthy-devel, libSM-devel

I'm probably overlooking something, but why is anthy-devel needed at build-time?
Thanks.

(In reply to comment #21)
> I'm probably overlooking something, but why is anthy-devel needed at build-time?
> Thanks.

Well, I rechecked the build log and I found:
--------------------------------
checking for ANTHY... no
--------------------------------
This can be turned to yes in current Fedora packages by
installing anthy-devel.

(In reply to comment #22)
> Well, I rechecked the build log and I found:
> --------------------------------
> checking for ANTHY... no
> --------------------------------
> This can be turned to yes in current Fedora packages by
> installing anthy-devel.

Awesome. From this, everything else looks good and this package is APPROVED. :]

Please remember to close this bug as NEXTRELEASE after you've imported and built it.

Thank you for reviewing and approving this package!!

Well,
* Rebuild for FE-devel succeeded.
  Umm.. this was the first package after ACL system is
  introduced into cvs system and I was a bit confused,
  actually...
* SyncNeeded requested for FE-6, FE-5 (for FE-5, I have to
  recheck spec file)

Now I close this bug as CLOSED NEXTRELEASE. Thank you.

----------------------------------------------------------
btw, I still meet deluge freeze on FC-devel (not on FC-5)
I want to figure out what is the problem, however currently
I don't know the way...

Package Change Request
======================
Package Name: kazehakase
New Branches: F-8

cvs done.

Package Change Request
======================
Package Name: kazehakase
New Branches: F-12
Owners: mtasaka

Early branch request.

CVS done.