I've been at this for a good two hours now but I can't get the compiler to stop giving me this warning.
Warning: production type_anno -> is never reduced.
Below is my parser.mly code.
(*Optional type annotation*)
type_anno:
| {None}
| COL;name = ID {Some name}
;
statement:
|(*SEND stmt*) SEND; chn = ID; O_TBRACE; var1 = ID; COMMA; var2 = ID; C_TBRACE {Send(chn,var1,var2)}
|(*RECV stmt*) O_TBRACE; var1 = ID; COMMA; var2=ID; C_TBRACE; ARW; RECV; chn = ID; S_COL ; stmt = statement {Recv(var1,var2,chn, stmt)}
|(*SELECT stmt*) chn=ID; DOT; lbl=ID; O_TBRACE; l1 = ID; C_TBRACE; S_COL; stmt = statement {Select(chn, lbl, l1, stmt)}
|(*CASE stmt*) CASE; chn =ID; O_PAREN; branches = branch_statement C_PAREN {Case(chn,branches)}
|(*CAST stmt*) CAST; chn = ID;O_TBRACE;l1=ID;C_TBRACE; {Cast(chn,l1)}
|(*SHIFT stmt*) l1 = ID ; ARW; SHIFT; chn=ID; S_COL;stmt=statement {Shift(chn, l1, stmt)}
|(*ACCEPT stmt*) l1 = ID ; ARW; ACC; chn=ID; S_COL;stmt=statement {Accept(chn, l1, stmt)}
|(*ACQUIRE stmt*) l1 = ID ; ARW; ACQ; chn=ID; S_COL;stmt=statement {Acquire(chn, l1, stmt)}
|(*DETACH stmt*) DET; chn=ID; O_TBRACE; l1=ID;C_TBRACE;{Detach(chn,l1)}
|(*RELEASE stmt*) l1 = ID ; ARW; REL; chn=ID; S_COL;stmt=statement {Release(chn, l1, stmt)}
|(*CLOSE stmt*) CLOSE; chn=ID;{Close(chn)}
|(*WAIT stmt*) WAIT; chn=ID;stmt=statement;{Wait(chn,stmt)}
|(*SPAWN stmt*) SPAWN; ref =ID; ARW; fn_id =ID ;inp = params; S_COL; stmt=statement {Spawn(ref,fn_id,inp,stmt)}
|(*NEW stmt*) chn =ID;anno=type_anno; ARW; NEW; stmt =statement ; S_COL ;stmt0 = statement {match anno with Some a -> NewA(chn,a,stmt, stmt0) | None -> New(chn,stmt,stmt0)}
|(*OUTPUT stmt*) PRINT; O_PAREN;var= ID; C_PAREN {Out(var)}
|(*FWD stmt*) FWD; O_PAREN; chn1=ID; COMMA;chn2 = ID; C_PAREN {Fwd(chn1,chn2)}
|(*SPL stmt*) O_TBRACE; var1 = ID; COMMA; var2=ID; C_TBRACE; ARW; SPL; chn = ID; S_COL ; stmt = statement {Split(var1,var2,chn, stmt)}
CodePudding user response:
The more important part of the warning emitted by menhir is that your grammar has a shift/reduce conflict that was resolved in favor of the shift.
Due to this conflict, the production
type_anno:
| {None}
can never be reduced since it is overridden by all productions that accepts ARW
after an ID
in the statement
rule
|(*SHIFT stmt*) l1 = ID ; ARW; ...
|(*ACCEPT stmt*) l1 = ID ; ARW; ...
|(*ACQUIRE stmt*) l1 = ID ; ARW; ...
|(*RELEASE stmt*) l1 = ID ; ARW; ...
...
|(*NEW stmt*) chn =ID;anno=type_anno; ARW; NEW; ...
You can check the issue in menhir
interpret mode that your parser as written rejects
ID ARW NEW
The smallest fix for the shift/reduce conflict is to make the rule type_anno
an inline rule:
%inline type_anno:
| {None}
| COL;name = ID {Some name}
;