Subject: Re: proposal for new format of pkgsrc README.html
To: None <email@example.com>
From: James K. Lowden <firstname.lastname@example.org>
Date: 05/11/2005 23:10:34
Jeremy C. Reed wrote:
> On Wed, 11 May 2005, James K. Lowden wrote:
> > * Include DESCR in the body as a quotation.
> Yes, that is good.
> Have a look at
> pkgsrc/mk/bsd.pkg.mk for various targets for readme or README generation
Thank you, Jeremy, as ever.
You didn't mention pkgsrc/mk/scripts/genreadme.awk. Yikes! 500 lines of
awk? That's 2 orders of magnitude longer than my longest awk script. I
expected to be working with XML and jade, not (for all its virtues) awk.
What's the philosophy here? Are we free to redo this with more powerful
tools, or does the HTML in pkgsrc have to be generated by tools in the
If I were approaching this problem fresh, I'd do something like:
1. Start with XML templates (instead of HTML ones)
2. Extract the data (e.g. dependencies) in a neutral form.
3. Convert the data to XML entities (included in #1).
4. Process with jade, per usual.
Apart from ease of maintenance, we'd have PDF output....