Subj : Re: File Syntax checking: need logical help
To   : comp.programming
From : =?iso-8859-1?q?C=E9dric_LEMAIRE?=
Date : Tue Aug 02 2005 02:27 am

Use a parser generator to describe your file structure. Example with
CodeWorker, an open source parsing tool and code generator freely
available at "http://www.codeworker.org":

file_structure ::= #continue A;
A ::= /*BNF sequence for what is specific to A*/ #continue B H;
B ::= /*BNF sequence for what is specific to B*/ #continue C E F G;
C ::= /*BNF sequence for what is specific to C*/ #continue [[C1 | C2 |
C3] C]? [D]?;
D ::= /*BNF sequence for what is specific to D*/;
E ::= [E1 &| E2] | E3;
E1 ::= /*BNF sequence for what is specific to E1*/;
E2 ::= /*BNF sequence for what is specific to E2*/;
E3 ::= /*BNF sequence for what is specific to E3*/;
F ::= /*BNF sequence for what is specific to F*/;
G ::= /*BNF sequence for what is specific to G*/;
H ::= /*BNF sequence for what is specific to H*/;

This is an extended-BNF script you have to execute with the CodeWorker
procedure parseAsBNF().
'#continue' means that the rest of the BNF sequence must match the
input file (similar to your 'mandatory', but placed according to
another logic).

The comment /*BNF sequence for what is specific to ...*/ might contain:
  - the reading of a fixed record:  [#readChar]10 // for a length of
.... 10!
  - the reading of a variable length: #readChars(my_length)
  - the reading of some usual tokens: #readInteger, #readCString,
#readIdentifier ...
  - a jump to a particular pattern: ->P // read up to encountering and
consuming P
  - a combination (sequences, alternatives) of them,
  - ...

Note about E ::= [E1 &| E2] | E3:
the operator 'A &| B' means 'A | A B | B' in CodeWorker (A and/or B).

.