lexical analysis - Conflict while parsing a set of expressions -

i parse set of expressions: r[3]c, r[2]c, r[3]c-r[2]c... there conflict cannot solve...

here part of lexer.mll:

  rule token = parse   | 'r'            { r }   | 'c'            { c }   | "rc"           { rc }   | ['0'-'9']+ lxm { integer (int_of_string lxm) }   | '+'            { plus }   | '-'            { minus }   | '['            { lbracket }   | ']'            { rbracket }   | eof            { eof }   ... 

a part of parser.mly:

main:    e_expression eof                { $1 };  e_expression: | ec = e_cell { ee_rc (rc.cell ec) } | e_expression minus e_expression { ee_string_eel ("minus", [$1; $3]) }  e_cell: | r lbracket r = index rbracket c c = index { (rc.i_relative r, rc.i_absolute c) } | r lbracket r = index rbracket c { (rc.i_relative r, rc.i_relative 0) }  index: | integer { $1 } | minus integer { printf.printf "%n\n" 8; 0 - $2 } 

this code curiously not work r[3]c-r[2]c, here parser.conflicts, can not understand.

if comment line line | r lbracket r = index rbracket c c = index ... in e_cell, code can parse r[3]c-r[2]c, 3 , 2 index, `r[3]c , r[2]c e_cell, , r[3]c-r[2]c e_expression.

could help?

so problem seems when sees "-" token after ], parser unsure whether making index, or if separating 2 expressions.

i.e. when parser reaches r[3]c-, isn't sure whether needs wait integer complete e_cell , reduce, or reduce , start work on e_expression.

the best way solve move negative integer code lexer. don't have ocamllex install handy, think changing

['0'-'9']+ 

to

'-'? ['0'-'9']+  

would work, , remove negative integer case index (obviously cause problem printf statement, can make internal logic more complex account this.