Plans for next: prediction, nulling events

[Jeffrey Kegler]

With an indexed release soon, that question of what to do next comes
up. I plan to move on to implementing events for predicted symbols, and
for nulled symbols.

Predicted symbol events will trigger when a symbol is expected. These
differ from the lexeme pauses in two ways. First predicted symbols
events won't be restricted to lexemes. Second, lexeme pauses look ahead
at the input -- they only take effect if the lexeme is actually there.
Prediction events will trigger without looking ahead.

Nulled symbol events will trigger whenever a symbol is nulled -- that
is, recognized as zero-length. Among other things, this will allow
straightforward implementation of zero-width assertions.

I had been thinking of moving on to ASF's (abstract syntax forests).
But completion events have me immersed in the same sections of Libmarpa
that I will need to work on for prediction and nulled events, and it
saves time to work with code while it is fresh in my mind.

Yours comments are welcome and appreciated, jeffrey

[Durand Jean-Damien]

Since prediction would not do look-ahead, do you think there would be a use for a "start" event, an event that is triggered when a lexeme is truely starting ?
I.e. something like. event nameOfStartEvent= start <lexeme>
Just ideas since you work are on this area currently.
In fact I was missing the "start", and had to implemented backward search using progress() to recover the G1 location of it.
Thanks, Regards, Jean-Damien.

[Jeffrey Kegler]

Doesn't pause => before for lexemes do what you want here? -- jeffrey

Durand Jean-Damien wrote:

Since prediction would not do look-ahead, do you think there would be a use for a "start" event, an event that is triggered when a lexeme is truely starting ?
I.e. something like. event nameOfStartEvent= start <lexeme>

Going forward, perhaps a position event could be envisaged, i.e. event NameOfPositionEvent = position <lexeme:d> where d is a digit.

Just ideas since you work are on this area currently.
In fact I was missing the "start", and had to implemented backward search using progress() to recover the G1 location of it.
Thanks, Regards, Jean-Damien.

Le mercredi 15 mai 2013 03:10:56 UTC+2, Jeffrey Kegler a écrit :

--
You received this message because you are subscribed to the Google Groups "marpa parser" group.
To unsubscribe from this group and stop receiving emails from it, send an email to marpa-parser...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

[Durand Jean-Damien]

Yes, plus a lookup to progress() then I determine the G1 rule where is the parser
/ JD.

Le mercredi 15 mai 2013 03:10:56 UTC+2, Jeffrey Kegler a écrit :