[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: RES: /usr/lib vs /usr/libexec

Thomas Bushnell BSG <tb@becket.net> writes:

> Most packages had files in /usr/doc.  Most packages do not have files in
> /usr/lib at all, and most of those that do, wouldn't need to be changed.

Changing from /usr/doc to /usr/share/doc was a fairly simple and
straightforward change in a whole bunch of packages.

Changing from /usr/lib to /usr/libexec is not a simple change unless
upstream already supports it, and upstream support for it is, in my
experience, rare, restricted to a few BSD-derived programs and some core
GNU software.  Most Autoconf-based software I've built doesn't support it,
let alone the software that doesn't even use Autoconf, and adding support
is frequently going to require patching the upstream source.

Furthermore, it's a change that is far more likely to have backward
compatibility ramifications than the /usr/doc change.  Thinking it over, I
stand by my statement.  I think the /usr/doc phaseout was simpler.

Of course, this is all back-of-the-envelope speculation, and again, the
difficulty of the transition is not my primary objection.  My primary
objection is that Debian follows the FHS, for very good reasons as stated
in previous messages, and this isn't in the FHS.  My secondary objection
is that you've not clearly demonstrated why this is an improvement, either
through benchmarks showing the performance difference or through
statistical analysis showing how overall Debian packaging would be
simpler, although until you address the primary objection, that wouldn't
convince me anyway.

For me, this is a closed issue until you change the FHS.  (Something that
I don't think is very likely to happen, but best of luck to you.)

Russ Allbery (rra@stanford.edu)             <http://www.eyrie.org/~eagle/>

Reply to: