Re: [exim-dev] XSL include path OS-independence

On 2010-07-06 at 08:45 +0100, Nigel Metheringham wrote:
> However the PDF/PS production route (if using sdop) didn't actually
> need these tweaks. And Mike Cardwell's HTML generation also doesn't.

Sorry, I should perhaps have been clearer: I know PDF/PS doesn't, I was
confirming they weren't broken by it; I myself only ever actively use
the .txt format and I reckoned that one would be sufficient to prove
correctness.

> So that just leaves text and texinfo - and I'm wondering if that too can
> be handled by a simpler method note requiring the installation of a ton
> of docbook processing stuff.

I applaud the motivation for trying to use docbook, but endorse any move
to get rid of it in practice. The performance and quality of sdop are
eye-opening.

> At present its starting to look as though the doc generation is getting
> simpler (and faster) and less likely to break horribly when the moon is not
> in the right phase.

Good. Now, if I could figure out why the .txt generation is producing
C2 characters after the section numbers in the text body (not in the
index), I'd be happy. Setting LC_ALL=C does not help. I can just strip
them out again in Tidytxt but that's hacky. It's w3m introducing them;
w3m/0.5.2 -- anyone else seeing this?

-Phil