RPM Community Forums

Mailing List Message of <rpm-devel>

Re: Paths for per-interpreter initialization?

From: Jeff Johnson <n3npq@mac.com>
Date: Fri 08 May 2009 - 18:05:52 CEST
Message-id: <5AE2FB24-D28F-43CE-9720-2560BE544F6D@mac.com>

On May 8, 2009, at 4:35 AM, devzero2000 wrote:

> OK. Got the point.
>

So (because the solution for paths is obvious and simple) I've
added bin/ and lib/ directories to /usr/lib/rpm, and install
the JS shell in bin/js, and the JS bindings for RPM in lib/libjsm.so.

That solution should surprise noone, is entirely consistent with
existing practice like /opt/APPLICATION/{bin,lib}/, and solves
the tyranny of having to discuss and justtify each and every path used  
by
rpm to death for usage cases that may never be important.

For multilib, one can have /usr/{lib,lib64}/rpm multilib paths.
That's consistent with how many interpreters are providing multilib.

But I'm _SURE_ someone will find FHS flaws with the scheme. Isn't
it ironic that choices like "executables in bin/, libraries in lib/"
is somehow in need of discussion on uglix these days because of  
"standard"
carpet-baggers.

Now to figger how to use mount -bind (or FUSE) to propagate /usr/lib/ 
rpm/bin
into chroot's, and to permit relative paths like "sh" instead of
"/bin/sh" so that cross-packaging might be contemplated without
the rigors of switching everything to using embedded interpreters like  
lua.

73 de Jeff
Received on Fri May 8 18:06:21 2009
Driven by Jeff Johnson and the RPM project team.
Hosted by OpenPKG and Ralf S. Engelschall.
Powered by FreeBSD and OpenPKG.