paulzfm/ll1pg

Yet another LL(1) parser generation tool, built in principle.

ll1pg's Issues

Better error report for undefined character terminate token

If there is a rule using undefined character terminated token. It will report confusing error:
Test : '^' { // An confusing error will report for '^' // for '^' is not a defined token } ;
For the above code, if i haven't defined %token '^', it will report
Error at (The position of quotation token): string matching regex [_A-Za-z][_A-za-z0-9]* expected but `'' found
it's somehow confusing and maybe an undefined token is better if the right side is an undefined terminate character token.

Should we warn about unused productions?

https://github.com/paulzfm/LL1-Parser-Gen/blob/688381a09a262b31091665838e89143e3d607e4f/src/main/scala/JavaCodeFile.scala#L227

Before write this line, should we check if ts is empty? If it is empty, the production is unused, and should we warn about it?

Currently, in some situation, this program may generate a Table.java like this:

// ...
                switch (lookahead) {
                    case PRINT:
                    case CASE:
                    case COMPLEX:
                    case VOID:
                    case FOR:
                    case '!':
                    case '-':
                    case CLASS:
                    case READ_LINE:
                    case WHILE:
                    case RETURN:
                    case NULL:
                    case INT:
                    case PRINT_COMP:
                    case SCOPY:
                    case '}':
                    case '@':
                    case DO:
                    case IDENTIFIER:
                    case NEW:
                    case '$':
                    case IF:
                    case THIS:
                    case INSTANCEOF:
                    case STRING:
                    case LITERAL:
                    case BRANCH:
                    case ELSE:
                    case '(':
                    case ';':
                    case '#':
                    case OD:
                    case DCOPY:
                    case BOOL:
                    case SUPER:
                    case BREAK:
                    case READ_INTEGER:
                    case '{':
                        return new Pair<>(131, Arrays.asList());           // <---
                        return new Pair<>(43, Arrays.asList(ELSE, Stmt));  // <---
                    default: return null;
                }
// ...

This two lines make Java compiler unhappy.