On 11 May 2015 at 13:49, D. Hugh Redelmeier <[email protected]> wrote:

> | From: Paul Wouters <[email protected]>
>
> | But since we don't export a shared library, we don't have a devel
> | package that would include the headers and developer man pages....
>
> Right.  Simplest solution that kind of works:  build those manpages,
> but don't install them.  Why build them?  Because they are hard to
> read otherwise.  This wasn't true in Henry's day (because we wrote
> directly in *roff).
>

FYI, I've pushed the following:

- any existing library make file rules that build/install man page have
been removed; i.e., library man pages no longer get installed

- some of library manpages that appear to have been generated from .xml
source have also been removed; the process is fragile so I'm being slightly
cautious

If a library man page needs to be build, then perhaps "make foo.8" will
work (I'm claiming plausible deny-ability has testing would likely result
in more work ... :-)

I also have to wonder if the knowledge contained in some of these man pages
should be folded back into the source; and, to take this off topic, if even
the libraries that only programs/pluto/ uses should be deemed failed
experiments and moved back to programs/pluto :-)
_______________________________________________
Swan-dev mailing list
[email protected]
https://lists.libreswan.org/mailman/listinfo/swan-dev

Reply via email to