mackenzie-high / snowflake Goto Github PK

This method is usually needed, so it might as well be auto-generated.

In the future, it would be nice for Snowflake to support user-defined timeouts and recursion-depth-limits in order to make the parser behave better when one accidentally creates a left-recursive grammar. In addition, it may be possible to add functionality to limit the (global) maximum number of match attempts.

Add findChild(String name) : ITreeNode and findChildren(String name) : List to the ITreeNode class. These are simply convenience methods that perform a DFS by calling the static equivalents that already exist.

Currently, users must have the snowflake library on the CLASSPATH in order to use a generated parser. This can be undesirable in some circumstances. It may be possible to create a single class file containing the generated parser, generated visitor, and a copy of the parser library embedded therein as nested classes. This would be a cheap way of providing the desired functionality.

Suppose that you have the following grammar rules:

ESCAPE_B = "\b";
ESCAPE_T = "\t";
ESCAPE_N = "\n";
ESCAPE_F = "\f";
ESCAPE_R = "\r";
ESCAPE_SL = "\";
ESCAPE_SQ = "'";

Then, the following parser configuration code will be generated.
g.str("ESCAPE_B", "\b");
g.str("ESCAPE_F", "\f");
g.str("ESCAPE_N", "\n");
g.str("ESCAPE_R", "\r");
g.str("ESCAPE_SL", "\");
g.str("ESCAPE_SQ", "'");
g.str("ESCAPE_T", "\t");

However, the correct code would be:
g.str("ESCAPE_B", "\b");
g.str("ESCAPE_F", "\f");
g.str("ESCAPE_N", "\n");
g.str("ESCAPE_R", "\r");
g.str("ESCAPE_SL", "\\");
g.str("ESCAPE_SQ", "\'");
g.str("ESCAPE_T", "\t");

Work Around:
ESCAPE_B = '' , 'b';
ESCAPE_T = '' , 't';
ESCAPE_N = '' , 'n';
ESCAPE_F = '', 'f';
ESCAPE_R = '', 'r';
ESCAPE_SL = '','";
ESCAPE_SQ = '', "'";

This is a bug in the code-generator, not in the snowflake library itself.

Currently, the code-generator generates a visit() method containing an if-elif-else comprised of string equality tests. This should be converted to either a string-switch or a HashMap of lambdas.

Currently, there is not a way to force stop a parser. This functionality should be added. Then, as a separate addition, the functionality should be integrated with the GUI in order to allow the user to stop parsing attempts at will.