- Made some chars const as proposed by Stefan Huehner <stefan@huehner.org>.
- Synced parser and keyword lists.
- Copied two token parsing from backend parser to ecpg parser.
- Also added a test case for this.