The goals for this project is to create Pike RPMs, produce all needed
documentation about how that is done and finally getting a 'make RPM' run
in the Pike source tree to build this RPM.
RPMs will be suggested and discussed in the Pike development forums.
This page will contain the RPMs, what they can do and what they should be
able to do. The project is coordinated (and was started) by
An automated RPM generation mode is necessary.
It must be possible to have multiple versions of Pike installed
simultaneously, both RPM and non-RPM builds.
Pike modules that Pike doesn't always need should be built as
separate RPMs for those who do not want a mastodont Pike RPM.
What do we have now?
Lenny Cartier () has made a
7.2.239 RPM. It is great work that
should be possible to use, but it has some errors and missing features:
It installs itselves in /usr/ (just as a Mandrake RPM should).
A general Pike RPM should install itself in /usr/local/, though.
The binary RPM doesn't include precompiled Pike modules, so the
installation of them takes way too much time.
The GMP module is not a part of the Pike RPM. This is bad, because
Pike doesn't work at all without it, but it is good because you can
then recompile just the Pike GMP source RPM if you want to optimize
it for you architecture.
The .spec file probably needs some extensive modifications before it
will work with Pike 7.3 or greater.
It doesn't install Pike in [prefix]/pike/7.2.239/, so it is not
possible to install many versions of Pike.
To handle many versions, the RPM should do like this during installation:
[prefix]/bin/pike should be a link to /etc/alternatives/pike,
which is a link to [prefix]/bin/pike-7.2, which is a link to
[prefix]/bin/pike-7.2 shall be updated by the RPM even if installed
/etc/alternatives/pike should only be updated with -U (?).