bug#37659: rx additions: anychar, unmatchable, unordered-or

[Paul Eggert]

On 10/22/19 8:14 AM, Mattias Engdegård wrote:
> 'regexp-opt' always generates a regexp preferring long matches. This is 
> undocumented, but useful enough that I would be surprised if this property 
> wasn't exploited (perhaps unknowingly) by callers. It's quite natural: given a 
> set of strings, surely the caller want them all to be candidates for a match, 
> even if there is no following anchoring pattern.

Yes, the longstanding tradition is that regular expressions are greedy.

> Thus, instead of 'unordered-or', define the operator in terms of long matches: 
> 'or-max' (working name) would work like 'or' but guarantee a longest match, and 
> only permit strings and 'or-max' forms as arguments.

That's an odd restriction. I'm not sure it's a good idea to add an 
operator with such a restriction. That is, I know why the restriction is 
there (it's because of limitations in the Emacs regexp matcher), but 
it's not clear that users should have to know and understand these details.

Moreover, if greed is the longstanding tradition for regexp-opt, 
shouldn't plain "or" be greedy, to be consistent with other operators? 
That is true for POSIX regular expressions involving "|". For example, 
the shell command:

echo abbc |
awk '{n=split($0, a, /b|bb/); for (i=1;i<=n;i++) print a[i]}'

outputs the two lines "a" and "c" (not the three lines "a", "", and "c") 
because the "b|bb" matches greedily.

If it's too much trouble to make plain "or" greedy, I suggest just 
documenting it as possibly being greedy and possibly not (that is, 
document it as being unordered, even if it happens to be ordered now). 
This will give us more opportunity for optimization later.

More generally, surely it would be better to improve the underlying 
Emacs regular expression matcher to have a greedy "or", or a stingy 
"or", or whatever.