lilypond-devel
[Top][All Lists]
Advanced
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: creating dev/gop-web

From: Graham Percival
Subject: Re: creating dev/gop-web
Date: Fri, 26 Dec 2008 17:11:47 -0800
User-agent: Mutt/1.5.18 (2008-05-17) 
We can take this public now...

On Sat, Dec 27, 2008 at 01:43:35AM +0100, John Mandereau wrote:
> Le vendredi 26 d??cembre 2008 ?? 14:25 -0800, Graham Percival a ??crit :
> > Could one of you kind git-savvy people please create a branch
> > containing the attached (untarred) files?
> 
> Let's name it web-gop, because dev/JACK is usually the branch of
> developer named JACK, and we create a lot of branches named with gop, I
> prefer to list it near web in alphabetical order, thus web-gop.

Ok.

> Done, but I tried to integrate the files in current site directory
> structure:
> - new directory texinfo/ for texinfo sources;
> - images went to site/images (I renamed some with big suffix to keep
> existing images) -- this has broken images in HTML output, but let's
> setup a clean directory structure first.

Thanks!  Ok guys, we now have something concrete to talk about.
Checkout web-gop, then cd texinfo and make.  Beautiful, ain't it?
Well, ok, not at all beautiful.

> Maybe Texinfo stuff should even be moved into the site/ tree, because
> - it would not make makefiles more complicated,
> - it would ease HTML processing by Python scripts,
> - it would be easier to write output page to the right output directory
> and guess in advance where it will go, e.g. in order to predict correct
> links between HTML pages from Texi2HTML and HTML pages already in the
> sources.
> 
> Any thoughts?

At the moment I don't think this is a concern; we still need to
figure out if this will work at all.

That said, my initial thought is that you're right.  The directory
structure should match the website; that should make it easier to
figure out what file to modify.
 
> > After that, if somebody could spend a few minutes hacking an
> > initial stylesheet / texi2html configuration so that it looks
> > slightly less horrible, that would be great.
> 
> Hacking a texi2html configuration is not trivial (at least not for an
> analphaperl like me): we must know what we want first -- see below.

I didn't think it was; that's why I wasn't certain that we wanted
to go this route.

> > Particular concern: the texinfo manual says that image scaling is
> > only implemented in TeX, not in HTML.  Does texi2html handle this
> > yet?  The images are currently way too large.
> 
> Two ways:
> 1) use ImageMagick/convert to scale images at build time;

Seriously?  I was expecting that we could do {scale=50%} or
something like that... oh well.  It seems that the lilypond
website already requires imagemagick, so this won't add any new
dependencies.

> > Don't spend too long on it, though -- we're still at the stage of
> > determining if this is feasible/desirable.
> 
> It is, but we have two ways of integrating texi2html-generated HTML:
> 1) hack an init file to produce HTML output ready to be published like
> current lilypond pages (or any layout we'll end up with for pure HTML
> pages), this means currently adding a navigation menu at top;
> 
> 2) hack an init file to generate bare HTML pages (ie without any <html>
> or <head>), then process them like HTML pages in the sources.
> 
> I prefer 2), because it keeps all layout formatting in Python processing
> scripts.

Hmm.  My initial thought was to do #1 -- essentially just take the
left-hand navbar in the NR and rotate it 90 degrees.  But it's
certainly true that the existing website already handles the
layout formatting stuff.


I'm abstaining on all of these questions.  I have no experience in
this area, and frankly not all that much interest.  As I said on
my newly-posted GOP "help wanted" pages, I'll handle the content
and leave the presentation to other people.  I think (hope?) that
I've described the problem, our desired output (pdf and html), and
given enough content to experiment with; I'll leave the rest of
the technical questions to you guys.

I think that you (John) have more experience with our html output
than anybody else (from all your the translation work), so if
nobody has dissenting opinions, we should do whatever you think is
best.

Cheers,
- Graham
reply via email to
[Prev in Thread] Current Thread [Next in Thread]