Re: forward-comment and syntax-ppss

[Andreas Röhler]

On 08.12.2016 22:47, Stefan Monnier wrote:
> > Can parse-partial-sexp parse backwards?
> No.  Two reasons: one is that we don't have code that (attempts to)
> do(es) that, the other is that it's fundamentally difficult/impossible
> to do it reliably.
>
> More specifically, the patch "forward-comment and syntax-ppss" is
> specifically handling a case where we really don't know how to parse
> backward: when moving backward, forward-comment generally tries to find
> the beginning of a comment by "parsing backward", but in some cases it
> really can't figure it out and (currently) resorts to calling
> (parse-partial-sexp (point-min) POS).

Another argument why syntax-ppss shouldn' t be used at all, thanks.


>    Since it may potentially do that
> for every "comment-end" marker it finds (e.g. every \n in most major
> modes), there are pathological cases where the performance in large
> buffers can become really poor, hence the need for a cache.

Performance of parse-partial-sexp was very fast in its original design. 
May our masters of circular dependencies have managed to slow it down? 
At any case: the question is, if you need the info from point-min. If 
yes, there is not other way than calling it from there. If not, don't 
ask for it, give a nearer boundary.