Line and Column Number Counting¶
By default, a generated lexer stamps tokens with the line and column number to
indicate where their according match occurred. This facilitates the
localization of a token’s origin in the input stream. The underlying line and
column counting behavior can be customized with the mode tag <counter:
...>. Counting is configured in terms of character specific column number
increments, line number increments, and grid counting. The
latter corresponds to the traditional tabulator effect, i.e. jumping to the
next multiple of a grid’s base value. For example, if the current column
number is 13 and a character appears causing a grid step of 8, then it will be
set to 16 which is the next number greater 13 and dividable by 8.
The syntax to specify a counting action for a set of characters is:
<character-set> '=>' keyword [number] ';'
The character-set may be any expression that results in a character set.
The keyword \else represents the complement of all character sets which are
explicitly mentioned. keyword must be either columns, grid, or
lines. The optional number specifies the increment or grid size. If
number is not specified it is set to 1. The default counting behavior can
be defined as shown below.
mode EXAMPLE :
<counter: \else => columns 1;
[\t] => grid 4;
[\n] => lines 1; >
{
...
}
Counting and token stamping can be controlled via the command line arguments mentioned in table Command line options to disable counting features.. The disablement of a counting feature, naturally, implies the disablement of the according token class stamping.
Command line option |
Consequence |
|---|---|
–no-count |
no counting, no stamping of line and column counts |
–no-line-count |
no counting, no stamping of line counts |
–no-column-count |
no counting, no stamping of column counts |
–no-token-stamp |
no stamping of line and column counts |
–no-token-stamp-line-count |
no stamping of line counts |
–no-token-stamp-column-count |
no stamping of column counts |