Re: [PATCH] cex: fix reporting of null nonterminals

[Akim Demaille]

Hi Vincent,

Yes, indeed, it was the empty rules that made the reports confusing.
Good hunch, good catch!

> Le 21 juin 2020 à 22:20, Vincent Imbimbo <vmi6@cornell.edu> a écrit :
> 
> I implemented this to print A ::= [ ], but A ::= [ %empty ] might be clearer.

I tend to prefer %empty, indeed.

It is true though that it looks heavier

Example              B . b c
First derivation     S ::=[ B ::=[ A ::=[ B . ] b A ::=[ ] ] C ::=[ A ::=[ ] c 
A ::=[ ] ] ]
Second derivation    S ::=[ B C ::=[ A ::=[ B ::=[ A ::=[ . ] b A ::=[ ] ] ] c 
A ::=[ ] ] ]

Example              B . b c
First derivation     S ::=[ B ::=[ A ::=[ B . ] b A ::=[ %empty ] ] C ::=[ A 
::=[ %empty ] c A ::=[ %empty ] ] ]
Second derivation    S ::=[ B C ::=[ A ::=[ B ::=[ A ::=[ %empty . ] b A ::=[ 
%empty ] ] ] c A ::=[ %empty ] ] ]

but this example is artificial.  With "real world" symbol names, it's
probably more natural.

Let's wait for a possible other opinion, but eventually, I think I'll
use %empty instead.


Thanks a lot for this fix, Vincent, installed.