Re: yytname woes

[Christian Schoenebeck]

On Tuesday, November 14, 2023 6:16:56 PM CET James K. Lowden wrote:
> On Wed, 15 Nov 2023 12:35:31 +0100
> Christian Schoenebeck <schoenebeck@crudebyte.com> wrote:
[...]
> Bison has seen a lot of technical improvement over the years.  There is
> a C++ interface, a push interface, and a choice of generated parsers
> in a a variety of target languages. And there is a rich array of
> debugging/programming features. For me, and I think for many others,
> there is no better yacc.  
> 
> Documenting all that is a challenge. I think the manual is useful but
> needs rewriting. It oscillates between tutorial and reference, and
> tucks in tangental advice.  And it leaves crucial terms undefined, such
> as "internal Bison token code".  I've been using Bison daily for the
> last two years, and still have no idea what a "skeleton" is,

"skeleton" refers to the common structure and parts of generated parsers that 
are independent of the specified tokens and grammar (i.e. language), so simply 
spoken the portions that are (more or less) copied verbatim, unlike all the 
tables for instance which are always generated dynamically based on the 
specified tokens and grammar and therefore are filled with completely 
different values for individual generated parsers.

> or when exactly I can depend on $$ = $1 having been executed.

You mean the exact *order* of actions being executed, not whether or not they 
are executed at all, right?

/Christian