| 1 |
B() processes a file containing a grammar definition. Different |
| 2 |
from Bison++ and Bison grammar files, b()'s grammar file consists of only |
| 3 |
two sections. The general form of a b() grammar file is as follows: |
| 4 |
verb( |
| 5 |
Bisonc++ directives |
| 6 |
%% |
| 7 |
Grammar rules |
| 8 |
) |
| 9 |
Readers familiar with Bison may note that there is no em(C declaration |
| 10 |
section) and no section to define em(Additional C code). With b() these |
| 11 |
sections are superfluous since, due to the fact that a b() parser is a class, |
| 12 |
all additionally required code can be incorporated into the parser class |
| 13 |
itself. Also, bf(C++) classes normally only require declarations that can be |
| 14 |
provided in the classes' header files, so also the `additional C code' section |
| 15 |
was omitted from the grammar file. |
| 16 |
|
| 17 |
The `%%' is a punctuation that appears in every b() grammar file to |
| 18 |
separate the two sections. |
| 19 |
|
| 20 |
The b() directives section contains all necessary directive specifications, |
| 21 |
like directives declaring the names of the terminal and nonterminal symbols, |
| 22 |
and directives describing operator precedence, and the data types of semantic |
| 23 |
values of various symbols. |
| 24 |
|
| 25 |
The file's second section (grammar rules) is used to define how to construct |
| 26 |
each nonterminal symbol from its parts. |
| 27 |
|
| 28 |
One special directive is availble that may be used in both the directives |
| 29 |
section and in the grammar rules section. This directive is tt(%include), |
| 30 |
allowing you to split long grammar specifications into smaller, more |
| 31 |
comprehensible and accessible chunks. The tt(%include) directive is discussed |
| 32 |
in more detail in section ref(INCLUDE). |