Re: Reductions during Bison error handling

[Hans Aberg]

At 18:42 -0600 2002/05/23, Richard Stallman wrote:
>    A program relying on an unsupported Bison feature, perhaps undocumented,
>    discovered by trial and error or reading the sources or whatever, and
>    breaks when Bison is updated must expect to be rewritten.
>
>For some programs, that policy is acceptable.  The nature of the job
>Bison does requires a more cautious policy, as I explained before.  I
>expect the Bison maintainers to give compatibility a very high
>priority.

I think that if Bison does not have this supported/unsupported policy, then
it may attempt to move towards to that cautiously: After all, the idea is
to increase compatibility and ease of user. One advantage is that if people
rely on undocumented features, they can report them to Bug Bison and ask
for proper support.

Also, Bison has a number of undocumented features that now can only be
found out by the grapevine, for example features under development but
which need user input. It might be beneficial for that needed user input to
put these features into the manual and mark them "unsupported".

  Hans Aberg