UWN

Now, while `(\+)//1` is detected as a non-supported construct, the rule is still asserted as a fact. ``` ?- [user]. c--> \+d. error(representation_error(dcg_body),phrase/3). % Warning: overwriting -->/2 because the clauses...

``` ulrich@gupu:/opt/gupu/bakaq-scryer-prolog$ target/release/scryer-prolog ?- [user]. get_heap_limit(Lim) :- '$get_heap_limit'(Lim). set_heap_limit(Lim) :- '$set_heap_limit'(Lim). ?- set_heap_limit(100000). thread 'main' panicked at 'index out of bounds: the len is 87987 but the index is 8935141660703064304',...

Invalid syntax. What else to expect?

(One reason why you were not able to reproduce it with exactly the same values is that you invoked Scryer with a minimally different path: `target/release/scryer-prolog` vs. `./target/`...)

@infogulch, one minor improvement: Instead of `/*/bin/env` rather use `/*usr/bin/env` which should reduce the search overhead.

... and avoids exploits with `/tmp/bin/env` ...

The [manual](https://www.scryer.pl/clpz#labeling/2) states > Labeling is always complete, always terminates, and yields no redundant solutions. Is this still the case here?

The inference limit does not help much, space is out much too early: ``` ?- call_with_inference_limit((A#=max(2,max(A,A^A))),8500,R). Killed 5.235u 12.181s 0:25.13 69.2% 0+0k 15816+16io 113pf+0w ``` Yes, a clean resource error...

There are so many cases where everything works as expected, even ``` ?- A#>max(2,A^A^A). clpz:(A in 3..sup), clpz:(A#>=_A), clpz:(A^_B#=_C), clpz:(A^A#=_B), clpz:(_A in 7625597484988..sup), clpz:(_C+1#=_A), clpz:(_C in 7625597484987..sup), clpz:(_B in 27..sup)....

There must be a way to do this in rust when even `ps` and `top` know this