m (Remove RHEL4 information) |
|||
Line 271: | Line 271: | ||
</pre> | </pre> | ||
This snippet is nearly the same as the one for upgrading. Why can't we just combine this portion with the %pre portion? The answer is that we want to delete any old versions of the schema during an upgrade. But this has to happen before we install the new version (in the %post script) otherwise we end up removing the schema that the upgrading package installs. However, if it really is a removal that will leave no other instances of this package on the system, we have to clean up the schema before deleting it. | This snippet is nearly the same as the one for upgrading. Why can't we just combine this portion with the %pre portion? The answer is that we want to delete any old versions of the schema during an upgrade. But this has to happen before we install the new version (in the %post script) otherwise we end up removing the schema that the upgrading package installs. However, if it really is a removal that will leave no other instances of this package on the system, we have to clean up the schema before deleting it. | ||
[[Category:EPEL]] | [[Category:EPEL]] |
Revision as of 03:16, 9 May 2013
This page contains guidelines which are no longer relevant to Fedora, but still apply to EPEL packages (EL-4, EL-5, and/or EL-6). These guidelines are designed to avoid conflict with the larger Fedora Packaging Guidelines, but should any conflicts occur, these guidelines should take precedence (on EPEL packages).
As a reminder, these guidelines only apply to EPEL packages, not to Fedora packages.
Limited Arch Packages
When RHEL ships a package for only a subset of available arches, it's possibly for EPEL to ship that SAME package in order to satisfy dependencies in the other arches in EPEL. In order to do this:
- Make sure the package is not shipped on all arches. EPEL5: (ppc/i386/x86_64) EPEL6: (ppc64/i686/x86_64).
- Make sure the package meets the Fedora licensing and distribution rules. Nothing non-free or under an unacceptable license.
- Notify the epel-devel list of your intention to add this package.
- Change the release of the package to have a leading 0. EXAMPLE: RHEL has foobar-1.0-1, you change it to foobar-1.0-0.1 for EPEL.
- Add a Changelog entry that the package was added to EPEL and has a 0 leading version to keep it older than RHEL.
- Submit a SCM request asking for the el4/5/6 branch you need.
- Import and build your package, submit as update.
- Watch the RHEL version of the package. When it updates, you should update the EPEL version too. You should never update other than that.
NOTE: Do not add ExclusiveArch tags, this will break building on the other architectures!
EPEL 6 and earlier
Provides and Requires Filtering
Generic Filtering on EPEL6
On EPEL6, the version of rpm is too old to support the Fedora methods of filtering Provides and Requires. Please the older guidelines instead: EPEL:Packaging_Autoprovides_and_Requires_Filtering
Perl Provides and Requires on EPEL5 and older
Unfortunately, the modern macros for Provides and Requires Filtering (Packaging:AutoProvidesAndRequiresFiltering) do not work for EPEL 5 or older. There are two mechanisms for filtering Perl Provides and Requires in EPEL, either In %prep or via External scripts.
In %prep (preferred)
Filtering can be done entirely in the SPEC file, in the %prep section:
%prep %setup -q -n Foo-%{version} cat << \EOF > %{name}-prov #!/bin/sh %{__perl_provides} $* |\ sed -e '/perl(unwanted_provide)/d' EOF %global __perl_provides %{_builddir}/%{name}-%{version}/%{name}-prov chmod +x %{__perl_provides} cat << \EOF > %{name}-req #!/bin/sh %{__perl_requires} $* |\ sed -e '/perl(unwanted_require)/d' EOF %global __perl_requires %{_builddir}/%{name}-%{version}/%{name}-req chmod +x %{__perl_requires}
External filtering
Or the script can be placed in an external file and referenced from the specfile. This is worse than the above because the full path of the to-be-overridden script needs to be hardcoded into the file, ignoring the system rpmbuild config. It is, however, the method used by a significant number of existing packages.
Source98: filter-provides.sh Source99: filter-requires.sh %global __perl_provides %{SOURCE98} %global __perl_requires %{SOURCE99}
where filter-provides.sh contains:
#!/bin/sh /usr/lib/rpm/perl.prov $* | sed -e '/perl(unwanted_provide)/d'
and filter-requires.sh contains:
#!/bin/sh /usr/lib/rpm/perl.req $* | sed -e '/perl(unwanted_require)/d'
PHP PEAR Macros
On EPEL (EL-4/EL-5/EL-6), the "%{pear_macrodir}" macro is not defined. The simplest fix is to add this line to your spec file:
%{!?pear_metadir: %global pear_metadir %{pear_phpdir}}
Alternately, simply use %{pear_phpdir} instead.
EPEL 5 and earlier
Building packages for EPEL 5 on non EL5 machines
If you are receiving md5sum errors when building on non EL5 machines please try the following steps:
1) fedpkg clone <package> 2) fedpkg switch-branch el5 3) <make changes> 4) fedpkg srpm 5) koji build --scratch dist-5E-epel file.el5.src.rpm 6) Verify build completes 7) fedpkg commit 8) fedpkg push 9) fedpkg build
PHP ABI Check Handling
For Fedora EPEL 5:
%global php_apiver %((echo 0; php -i 2>/dev/null | sed -n 's/^PHP API => //p') | tail -1) BuildRequires: php-devel Requires: php-api = %{php_apiver}
There is no way of checking the ABI with packages for Fedora EPEL 5.
For a spec file which is compatible with both Fedora and EPEL 5:
%global php_apiver %((echo 0; php -i 2>/dev/null | sed -n 's/^PHP API => //p') | tail -1) %if 0%{?php_zend_api} Requires: php(zend-abi) = %{php_zend_api} Requires: php(api) = %{php_core_api} %else Requires: php-api = %{php_apiver} %endif
No API/ABI dependencies are available for Fedora EPEL 4 packages.
PHP PECL Module Scriptlets
On EPEL (EL-4/EL-5), the Fedora scriptlets for properly registering and unregistering the module have to be wrapped with conditionals checking for the existence of %{pecl_install} and %{pecl_uninstall}:
%if 0%{?pecl_install:1} %post %{pecl_install} %{pecl_xmldir}/%{name}.xml >/dev/null || : %endif %if 0%{?pecl_uninstall:1} %postun if [ $1 -eq 0 ] ; then %{pecl_uninstall} %{pecl_name} >/dev/null || : fi %endif
pkgconfig
rpm in EPEL5 and below does not automatically create dependencies for pkgconfig files. Packages containing pkgconfig(.pc) files must Requires: pkgconfig
(for directory ownership and usability).
python byte compilation
In EPEL5 the automatic byte compilation of python files that is performed by brp-python-bytecompile byte compiles all files that match *.py This is undesirable for program files in %{_bindir} and %{_sbindir} because the user will probably never invoke these files, only the main program file and python won't use these files. There are two workarounds:
- Rename scripts in %{_bindir} to not have a .py extension: For instance, from /usr/bin/orient.py to /usr/bin/orient.
- Use %exclude to exclude the scripts from the file listing:
%files %{_bindir}/orient.py %exclude %{_bindir}/orient.pyc %exclude %{_bindir}/orient.pyo
noarch subpackages
EL 5 and earlier do not support noarch subpackages. If your build fails due to unpackaged debuginfo files ensure that the BuildArch: noarch
is wrapped in an if to make sure its not used on EL-5 and earlier.
xz compression
Tar in EL 5 and earlier does not support extracting xz-compressed tarballs. To extract such tarballs, use the following %prep section:
%prep xzcat %{SOURCE0} | tar -xf - %setup -qDT
BuildRoot tag
rpm in EPEL5 and below does not automatically provide a value for the BuildRoot tag, so one must be provided in the spec by the packager.
The BuildRoot value MUST be below %{_tmppath}/
and MUST contain at least %{name}
, %{version}
and %{release}
. It may invoke mktemp
since this is guaranteed to exist on every system. From there, packagers are expected to use a sane BuildRoot.
The recommended values for the BuildRoot tag are (in descending order of preference) :
%(mktemp -ud %{_tmppath}/%{name}-%{version}-%{release}-XXXXXX) %{_tmppath}/%{name}-%{version}-%{release}-root-%(%{__id_u} -n) %{_tmppath}/%{name}-%{version}-%{release}-root
If unsure, simply pick the first.
Prepping BuildRoot For %install
It is important to properly prepare the BuildRoot in the %install
section of your package before it is used as rpm in EPEL5 and below does not do this automatically. Package for these releases MUST have an %install section that begins with either:
%install rm -rf %{buildroot}
or
%install rm -rf $RPM_BUILD_ROOT
This is to ensure that the BuildRoot will be created fresh during the %install
section.
Scrollkeeper
For EL-4 and EL-5, Gnome and KDE use the scrollkeeper cataloging system to keep track of documentation installed on the system. Scrollkeeper allows the help system to sort and search documentation metadata stored in .omf files. When you add documentation in these systems you need to make scrollkeeper aware that the documentation has been changed.
Note that we BuildRequires scrollkeeper as most Makefile's are setup to install the necessary scrollkeeper files only if scrollkeeper is present at install time.
BuildRequires: scrollkeeper Requires(post): scrollkeeper Requires(postun): scrollkeeper ... %post scrollkeeper-update -q -o %{_datadir}/omf/%{name} || : %postun scrollkeeper-update -q || :
These two scriptlets tell scrollkeeper to update its indexes to account for the new scrollkeeper files.
GConf Scriptlets
In Fedora, we now use macros for our GConf2 scriptlets, but for EL-4 and EL-5, this is not an option. For those targets, please use the old manual scriptlets, as documented below:
Requires(pre): GConf2 Requires(post): GConf2 Requires(preun): GConf2 ... %pre if [ "$1" -gt 1 ] ; then export GCONF_CONFIG_SOURCE=`gconftool-2 --get-default-source` gconftool-2 --makefile-uninstall-rule \ %{_sysconfdir}/gconf/schemas/[NAME] .schemas >/dev/null || : fi
In this section we uninstall the old schemas when we upgrade. The way we do this is first to get information about where gconf stores its values via the gconftool-2 --get-default-source line. Then we uninstall the schema from that source. If the package could be upgrading a package which had another name for the schema at one time, then we uncomment the lines to uninstall those as well.
The next section is for installing the new schema:
%post export GCONF_CONFIG_SOURCE=`gconftool-2 --get-default-source` gconftool-2 --makefile-install-rule \ %{_sysconfdir}/gconf/schemas/[NAME] .schemas > /dev/null || :
Here we do the same things as in the %pre section for upgrading except the gconftool-2 switch used is --makefile-install-rule to install the new schemas instead of the uninstall-rule to remove the old schemas.
The last section deals with deleting the schemas on package removal:
%preun if [ "$1" -eq 0 ] ; then export GCONF_CONFIG_SOURCE=`gconftool-2 --get-default-source` gconftool-2 --makefile-uninstall-rule \ %{_sysconfdir}/gconf/schemas/[NAME] .schemas > /dev/null || : fi
This snippet is nearly the same as the one for upgrading. Why can't we just combine this portion with the %pre portion? The answer is that we want to delete any old versions of the schema during an upgrade. But this has to happen before we install the new version (in the %post script) otherwise we end up removing the schema that the upgrading package installs. However, if it really is a removal that will leave no other instances of this package on the system, we have to clean up the schema before deleting it.