Message3152

Author malte
Recipients jendrik, malte, moritz
Date 2014-04-30.00:44:18
Content
As in most programming language grammars, I would suggest to separate the
"AlphaNumeric" token into two kinds of tokens: pure numbers vs. identifiers. So
for example "3=foo()" should not be a valid "KeyValue" because 3 is not a valid
identifier, and similarly 22(foo) would not be a call because 22 is not an
identifier.

Identifier would be the regex [a-zA-Z_][a-zA-Z0-9_]* (as e.g. in C).

Number would be [0-9][0-9]* (for integers) or [0-9][0-9]*.[0-9][0-9]* (for
floating-point numbers).

I think the grammar would use Identifier in all places where it currently uses
AlphaNumeric, except in RHSParam, where both Identified and Number would be allowed.
History
Date User Action Args
2014-04-30 00:44:18maltesetmessageid: <1398811458.47.0.977127455482.issue288@unibas.ch>
2014-04-30 00:44:18maltesetrecipients: + malte, jendrik, moritz
2014-04-30 00:44:18maltelinkissue288 messages
2014-04-30 00:44:18maltecreate