Re: lynx-dev lynxcfg.html ideas

[Vlad Harchev]

On Fri, 4 Jun 1999, Leonid Pauzner wrote:

> 2-Jun-99 12:12 Vlad Harchev wrote:
> 
> >   Anyway, I done conversion in ~8 hours of one day, so I was so tired to 
> > make
> > spelling errors.
> >   I'll place body.in and the patch to my homepage soon and will announce
> > lynx-dev about location.
> 
> Thinking a little more on it I found out that we may want
> a more Spartan design for lynxcfg.html by the cost of having
> an automatic conversion lynx.cfg <-> lynxcfg.html as expected.

 Seems this will be reconsidered by you when you'll inspect the patch 
 (the file scripts-without-body.tgz)
 Currently (in patch) the following scheme is used

 body.in+auxilary files -> lynxcfg.html
 body.in+auxilary files + special definition of macros -> temporary html 
   -> (with use of lynx to render, sed to turn in comments) ->  lynx.cfg

 So, to modify appearance of generated lynx.cfg, only cpp macros should be
 redefined/tweaked. 

 As I said, it's possible to emit any text uncommented to lynx.cfg with the
use of macro @EMIT(x).
 Search for it in body.in.
 
> My proposal is to write two tiny C programs (I am not familiar with sed)
> which read one file and generate another based on the assumptions
> of first (few) characters of each line. I will try to do this tomorrow.
> 
> The convertion lynx.cfg -> lynxcfg.html
> based on lynx.cfg lines syntax looks like this
> (rules applied in order):
> 
> 
> 1) the line is blank (or spases or tabs only) - may be a sign of
> a delimiter between different logical sections. may be not (?)
> 
> 2) the line begins with character != '#' - an active option,
> the name ended with ':' delimiter.
> 
> 3) the line begins with character == '#' and next coming
> character exist and not a space or tab - inactive option,
> the name ended with ':' delimiter.
> 
> both 2) and 3) may be found several times in lynx.cfg but assumed
> from the same section.
> 
> 
> Others lines belongs to comments and begins with "#" symbol.
> 
> 4) no characters after '#' (except probably a number of spaces/ or tabs only) 
> -
> a sign of starting a new paragraph in the comment - [markup].
> 
> 5) more than one space or a tab after '#' (text not flushed left) -
> a sign of authors markup - [show verbatim].
> 
> 6) the comment strings begins with "# " and no more tabs or spaces before
> next coming text - no special markup, [the text may be formated to any display
> width].
> 
> 
> Now we only need to *guess* the start/end of each section and its name
> based on rules 1) 2) 3).
> 
> Also we may add manually a sweety murkup for <em>... </em> etc.
> and set a logical sections names according to lynxcfg_toc.html (hmm),
> also do something with the default values and "see also" section
> if we want to be smart enough.  Without this sweety things we got
> a Spartan design but automatic procedure both ways.
> 
> 
> 

 Best regards,
  -Vlad