Re: bison version in graph output file

[tys lefering]

On Sun, May 2, 2010 02:55, Joel E. Denny wrote:
> What about something like
>   "LALR(1) Automaton for parser.y"
> where LALR(1) could also be IELR(1) or canonical LR(1)?
Good idea to add some more parser info.
because it will show up on google searches
more people will become aware GNU bison nowadays
has more types of parsers available.

>> for grep it can be idea to put bison version string
>> and parser name in one line
> I think we could potentially grow many pieces of information like this
> (for example, whether it's LALR(1) and whether default reductions are
> enabled), but how much should be on one line?
> What about using something like "grep -A3 Bison *.dot"?
That is also oke. creating collection of dot graphs for
testing usage saw it is better if there is a way to find
the GNU bison dot graphs easy. this way it works.

>> noticed there is no statement about the copyright status
>> of gcc and bison generated graph output files, assuming
>> it is "undefined" and maybe not good situation for the
>> fsf free software foundation.
> I don't know, but I don't believe we've ever worried about that before.
gcc graphs have the gcc internal (rtl) information
and assuming these are GNU GPL v3 because gcc is GNU GPL v3.

bison graphs have the information also found in the
data tables in the generated parser code and because
the generated parsers are "GNU GPL with special exception"
assuming the graph data also has this copyright.

> However, I'm still thinking that it might be nice to have more information
> (like the type of the automaton) and that we should make all the same
> additions to the ASCII and XHTML output.
some extra information is good idea, and easy to add from within bison.

Thanks for the patches and the good care for
nice graph output from GNU bison.