Correct C# Grammar too Sharp for ISO

Short Description
C# code could been successfully parsed (SGLR) with it. Grammar recovery-a technique for extracting a complete grammar out of an …

Website: www.cs.vu.nl | Filesize: 81kb

Content
Correct C# Grammar too Sharp for ISO
Vadim Zaytsev
Vrije Universiteit Amsterdam, The Netherlands,
vadim@cs.vu.nl
Introduction. The most used programming language nowadays is COBOL.
At the Free Unversity in Amsterdam we have done numerous transformations on
COBOL, parsed and transformed millions of lines of code. COBOL is standardised,
but vendors usually deviate from the standard, making their own dialects.
In order to parse code, we need a working grammar, which should be derived
from the compiler documentation. However, documentation is never complete
nor error-free, and special techniques are needed to obtain correct grammars:
grammar recovery and grammar (re)engineering. One can argue whether this
happens because of COBOL decades-long evolution and legacy.
Recently we started thinking about transforming C# code, too. C# is quite
different from COBOL, it is a very sharp modern language, the latest big accomplishment
in programming languages design. C# was produced by a big
corporation and submitted as a specification to both ECMA International1 and
ISO2. C# compiler provided by Microsoft claims to fully implement the standard.
Thus, one might think that this standard is of much better quality that
COBOL’s, making it easier to use it in parser construction….