lynx-dev keys and terminology and code[s] once again (was new command format, was transport)

[Klaus Weide]

On Wed, 7 Jul 1999, Philip Webb wrote:

> 990707 Klaus Weide commented on my suggestion for a new command `:' :
> i will respond later today, but meanwhile a detail which isn't clear:
> you mentioned `lynxactioncode' (SIC) in your original discussion re `t',
> but a quick look at Users Guide &  lynx.cfg  finds no mention:

I invented the terminology of 'lynxactioncode' and 'lynxkeycode' in
the February thread "Keys, an attempt to understand", see there for an
introduction to the gory details.  I found other terms used in various
places were too ambiguous or used inconsistently.  The 'lynx' prefix
to make it clear I was talking about something assigned by lynx not
ncurses etc.  'action' instead of 'function' to avoid any possible
confusion with C function names in the Lynx code, or with 'function'
as in 'function key', other functions of 'function'.  Explicitly
'code' because I was really talking about the C code representation
(basically just bit patterns of some integer type), as opposed to more
abstract notions.

I've stuck with the terms since then, at least in code comments
and some lynx-dev messages.  Some others seem to have picked them up.

>  lynx.cfg  speaks of `functions' & refers the user to  LYKeymap.c ,

'function' (in lynx.cfg)  == 'lynxactioncode', more or less.
Sometimes also called a 'keyboard command' or a 'keystroke command' or
just 'command', all these terms can be found in the UG.

> which has a table of values in the form  LYK_SOME_FUNCTION ,
> in both cases giving  > 1  mapping of some functions.

There are two different types of things there - well more, but let's
stop before 'funckey[]'.
First, there are mapping tables.  Then, there is a list 'revmap[]'.
The former don't attempt to provide any kind of listing, they are
just the default hardwired lynxkeycode-->lynxactioncode mappings.
The latter don't give any " > 1 " mapping, so I'm not sure what you
mean.

> so is there anywhere a complete non-repeating listing of `functions'?

The LYKeymapCode enum in the second half of LYKeymap.h (which has the same
order as 'revmap[]' in LYKeymap.c). It is non-repeating for any given
compilation.  If it weren't the compiler would complain.  Ignore the
conditional #define lines...

> what should we (all) consider the correct term for them?

I think I'll stick with 'lynxactioncode', at least for the purpose of
these kinds of exchanges (when looking at code).  [I'm aware that
doesn't answer your question.  Maybe the UG and lynx.cfg should
continue using the terms it uses, even if it's not consistent, if it
is understandable for the intended audience.]

> BTW not only did this come up with `t' recently,
> but earlier with KD's textarea-editing feature,
> which introduced several new functions which later caused confusion,
> possibly resolved by you to some extent in 2-8-3dev.n .

There is still stuff to clean up, primarily IMO statusline messages.

  Klaus