test: Disable no-arg callable statement tests in simple query mode

[sehrope]

I took a look into fixing #2399 but it's more complicated than I thought. This PR doesn't actually fix anything yet. It just disables those three new tests in simple query mode so as not to break the CI. Nothing has changed recently in that part of the driver, those are just new tests that exposed existing broken behavior.

Here's the block in Parser where the extra parameter is being added:

pgjdbc/pgjdbc/src/main/java/org/postgresql/core/Parser.java

Lines 1167 to 1204 in d281a78

	int opening = s.indexOf('(') + 1;
	if (opening == 0) {
	// here the function call has no parameters declaration eg : "{ ? = call pack_getValue}"
	sb.append(outParamBeforeFunc ? "(?)" : "()");
	} else if (outParamBeforeFunc) {
	// move the single out parameter into the function call
	// so that it can be treated like all other parameters
	boolean needComma = false;
	// the following loop will check if the function call has parameters
	// eg "{ ? = call pack_getValue(?) }" vs "{ ? = call pack_getValue() }
	for (int j = opening + prefixLength; j < sb.length(); j++) {
	char c = sb.charAt(j);
	if (c == ')') {
	break;
	}
	if (!Character.isWhitespace(c)) {
	needComma = true;
	break;
	}
	}
	// insert the return parameter as the first parameter of the function call
	if (needComma) {
	sb.insert(opening + prefixLength, "?,");
	} else {
	sb.insert(opening + prefixLength, "?");
	}
	}
	if (!suffix.isEmpty()) {
	sql = sb.append(suffix).toString();
	} else {
	sql = sb.toString();
	}
	return new JdbcCallParseInfo(sql, isFunction);
	}

That "parameter" is actually an out parameter to reflect the functions return value. However removing it ends up breaking subsequent invocations to CallableStatement.registerOutParameter(...) as then the driver thinks there are no out parameters.

The reason this all works fine in extended mode is that the server "helpfully" ignores any void parameters and pretends they're not part of the command. Honestly it's very odd and I think it's a server bug and fixing it there would likely break working (but errant) things like this driver's callable statement interface.

I think the fix for this would have to touch both the parsing and parameter parsing. Not sure when I'll be able to follow up on this so if anyone else wants to take a look please go ahead. That code is pretty old and we haven't heard any complaints about it so I doubt there's anyone actually running into this issue as you'd have to be both using simple mode and JDBC call syntax.

I ran the omni suite on my fork and it completed successfully (https://github.com/sehrope/pgjdbc/runs/4939369066?check_suite_focus=true) so once this cycles through CI here I'll merge it in to make CI green again.

[davecramer]

Thx!