message: syntax error, unexpected ';'

[MaxOstrowski]

Hi,

I'm using the minizinc/libminizinc tests to run with my own solver to have an additional test suite for the solver.
After some initial hickups this worked very well so far, adding ~450 more unit tests for the solver.

After changing from minizinc 2.5.5 to 2.6.0 I now recognized new behaviour that was triggered due to an update of minizinc-python.
In the libminizinc/tests/requirements.txt there is a dependency to the develop branch of minizinc-python.

When I now execute the tests with my solver, I do get the following errors on some of the tests:
pytest -k globals_all_different_int_opt --driver=../../MiniZincIDE-2.6.0-bundle-linux-x86_64/bin

E   AssertionError: expected one of
E
...
E     but got
E     
E     !Error
E     message: syntax error, unexpected ';'
E     type: SyntaxError

Executing minizinc directly with my solver
minizinc --solver flatzingo spec/unit/globals/alldifferent/globals_all_different_int_opt.mzn
or with chuffed
minizinc --solver chuffed spec/unit/globals/alldifferent/globals_all_different_int_opt.mzn
doesn't show anything special and the output is the same.

Any idea what and where this ';' could come from ?
A reproduction of the error can be found in this github action here:
https://github.com/potassco/flatzingo/runs/5360589007?check_suite_focus=true

Thanks a lot in advance.

[Dekker1]

Since this test is working on the MiniZinc CI, I would think that this is a problem with your environment. However, maybe @cyderize has more insight into the problem than I do.

[MaxOstrowski]

I agree that this is probably a problem with my environment and I would highly appreciate any pointers to hint me to a cause, as the output of the binary call minizinc with my solver seems to be indistinguishable from chuffed and gecode.

[cyderize]

I'm not sure what could cause this. It may be useful to try to run something more similar to what minizinc-python is doing. It should be something like:

minizinc --solver flatzingo --allow-multiple-assignments --json-stream --output-mode json --output-time --output-objective --output-output-item --statistics --intermediate-solutions spec/unit/globals/alldifferent/globals_all_different_int_opt.mzn

And see if there's anything suspicious about its output (it should be line-break-delimited JSON objects as described here).

[MaxOstrowski]

Thanks, this actually helps me a lot to understand this.
The call:

minizinc --solver flatzingo --json-stream --statistics spec/unit/globals/alldifferent/globals_all_different_int_opt.mzn

results in the following output:

{"type": "solution", "output": {"default": "alldiff_avi1 = array1d(5..9, [1, 2, 3, <>, 5, <>]);\n", "raw": "alldiff_avi1 = array1d(5..9, [1, 2, 3, <>, 5, <>]);\n"}, "sections": ["default", "raw"]}
{"type": "error", "what": "syntax error", "location": {"filename": "statistics received from solver", "firstLine": 1, "firstColumn": 10, "lastLine": 1, "lastColumn": 10}, "message": "syntax error, unexpected ';'"}
{"type": "error", "what": "syntax error", "location": {"filename": "statistics received from solver", "firstLine": 1, "firstColumn": 66, "lastLine": 1, "lastColumn": 66}, "message": "syntax error, unexpected '(', expecting end of file"}

I had some troubles understanding the error messages, but with some try an error I could identify the cause:

minizinc --solver flatzingo --statistics spec/unit/globals/alldifferent/globals_all_different_int_opt.mzn

without the --json_stream produces:

alldiff_avi1 = array1d(5..9, [1, 2, 3, <>, 5, <>]);
----------
%%%mzn-stat: models=1+
%%%mzn-stat: time=0.028
%%%mzn-stat: choices=3       
%%%mzn-stat: conflicts=0
%%%mzn-stat: rules=721      (Original: 715)
%%%mzn-stat: boolVariables=50
%%%mzn-stat: nogoods=52
%%%mzn-stat: init_total=9.4099e-05
%%%mzn-stat: init_simplify=1.8194e-05
%%%mzn-stat: init_translate=2.69e-07
%%%mzn-stat: csp_constraints=0
%%%mzn-stat: intVariables=2
%%%mzn-stat: csp_clauses=90
%%%mzn-stat: csp_literals=38
%%%mzn-stat: csp_clauses=0
%%%mzn-stat: csp_literals=0
%%%mzn-stat: csp_time_propagation=5.583e-06
%%%mzn-stat: csp_time_check=2.8e-07
%%%mzn-stat: csp_time_undo=0
%%%mzn-stat: csp_refined_reason=0
%%%mzn-stat: csp_introduced_reason=0
%%%mzn-stat: csp_literals_introduced=0
%%%mzn-stat: nSolutions=1
%%%mzn-stat-end

The following lines:

%%%mzn-stat: models=1+

and

%%%mzn-stat: rules=721      (Original: 715)

do trigger this bug.

According to the Minizinc Handbook, statistics have the form:

%%%mzn-stat: <name>=<value>

which does not impose any restrictions
Is there a restriction of what <name> and <value> can be or is this simply a bug in the json converter...

@cyderize Thanks a lot for the very useful hint that enabled me to pinpoint this.

[cyderize]

I believe in https://www.minizinc.org/doc-2.6.0/en/fzn-spec.html#statistics-output it does mention that

The <name> describes the kind of statistics gathered, and the <value> can be any value of a MiniZinc type.

So the values are supposed to be of a valid MiniZinc type. Previously this was not really enforced, so maybe we should make that clearer in the documentation.

[MaxOstrowski]

Thanks a lot, I actually just had to scroll down to find this restriction.
Now almost all tests are excepted by my solver. Thanks!