Question

I am parsing a verilog file to extract the dependencies within it. I don’t need to parse most of the module contents, but am only interested in include statements and module instantiations. I first am attempting to only extract the include statements. This is my code so far:

include_pragma = Group(Keyword("`include") + quotedString + lineEnd.suppress())
module_definition = ZeroOrMore(~Keyword("endmodule") + MatchFirst([include_pragma, restOfLine])) + Keyword("endmodule")

My point is that I will either match an include pragma, or a line of anything until I reach “endmodule”. If I try this grammer on the following string, pyparsing gets into some kind of infinite loop.

`include "InternalInclude.v"

localparam COMMA_WIDTH = 10;

localparam  UNKNOWN = 1'b0,
            KNOWN   = 1'b1;

reg     [DATA_WIDTH-1:0] TtiuPosQ2;
reg                      TtiuDetQ2;
reg     [           7:0] TtiuDetQ2Save;

assign DebugQO = {  DataQI,         // 65:34
                    TxTrainingEnQI, // 33
                    TtiuDetQ2,      // 32
                    TtiuPosQ2       // 31:0
                  };

always @(posedge Clk or posedge ResetI) begin
    if (ResetI) begin
        CommaPosQO <= {(DATA_WIDTH){1'b0}};
        CommaDetQO <= 0;
        CommaPosQ1 <= {(DATA_WIDTH){1'b0}};
        CommaPosQ2 <= {(DATA_WIDTH){1'b0}};
        CommaDetQ2 <= 0;
        StateQ <= 0;

        CommaPosSaveQ <= {(DATA_WIDTH){1'b0}};
        TtiuPosQ1 <= 0;
        TtiuPosQ2 <= 0;
        TtiuDetQ2 <= 0;
        TtiuDetQ2Save <= 8'h00;
    end
    else begin 
        CommaPosQO <= CommaPosC2;
        CommaDetQO <= CommaDetC2;
        CommaPosQ1 <= CommaPosC;
        CommaPosQ2 <= CommaPosQ1;
        CommaDetQ2 <= (| CommaPosQ1);
        StateQ <= StateC;
        CommaPosSaveQ <= CommaPosSaveC;
        TtiuPosQ1 <= TtiuPosC1;
        TtiuPosQ2 <= TtiuPosC2;
        TtiuDetQ2 <= TtiuDetC2;
        TtiuDetQ2Save <= TtiuDetQ2 ? 8'hFF : {1'b0, TtiuDetQ2Save[7:1]};
    end
end

endmodule

I am probably misunderstanding the ~ operator. Any suggestions?

update

After using the setDebug() method suggested, I found that the infinite loops prints this:

Match {Group:({"`include" quotedString using single or double quotes Suppress:(lineEnd)}) | Re:('.*')} at loc 0(1,1)
Matched {Group:({"`include" quotedString using single or double quotes Suppress:(lineEnd)}) | Re:('.*')} -> [['`include'
, '"InternalInclude.v"']]
Match {Group:({"`include" quotedString using single or double quotes Suppress:(lineEnd)}) | Re:('.*')} at loc 31(4,1)
Matched {Group:({"`include" quotedString using single or double quotes Suppress:(lineEnd)}) | Re:('.*')} -> ['']
Match {Group:({"`include" quotedString using single or double quotes Suppress:(lineEnd)}) | Re:('.*')} at loc 31(4,1)
Matched {Group:({"`include" quotedString using single or double quotes Suppress:(lineEnd)}) | Re:('.*')} -> ['']
Match {Group:({"`include" quotedString using single or double quotes Suppress:(lineEnd)}) | Re:('.*')} at loc 31(4,1)
Matched {Group:({"`include" quotedString using single or double quotes Suppress:(lineEnd)}) | Re:('.*')} -> ['']

Is something that I do causing the parse position not to move forward?

Answer 1

The problem was with using the restOfLine expression. It can match anything, including ”. So, the parser would proceed as follows:

1. Matched include "InternalInclude.v"
2. Match ” to restOfLine
3. Match the same ” to restOfLine because it did not consume the \n AND there is still ” left before the \n, as there alwasy will be 😉

To fix it, I changed restOfLine to (restOfLine + lineEnd). This forced the parser to consume the \n after matching the line. The new parser reads:

include_pragma = Group(Keyword("`include") + quotedString + lineEnd.suppress())
module_definition = ZeroOrMore(~Keyword("endmodule") + MatchFirst([include_pragma, (restOfLine + lineEnd)])) + Keyword("endmodule")