[Julia] Macro calls are parsed the same as function calls? #10495

iuliadmtru · 2024-09-02T17:42:38Z

Describe the bug
In the Julia syntax, both calls and macrocalls exist. However, Semgrep generates ASTs with the same node (Call) for both of them. For example, for basic calls:

f(x):

   ExprStmt(
     Call(
       N(
         Id(("f", ()),
           {id_info_id=8; id_flags=Ref(0); id_resolved=Ref(None);
            id_type=Ref(None); id_svalue=Ref(None); })),
       [Arg(
          N(
            Id(("x", ()),
              {id_info_id=9; id_flags=Ref(0); id_resolved=Ref(None);
               id_type=Ref(None); id_svalue=Ref(None); })))]), ())

@m(x):

   ExprStmt(
     Call(OtherExpr(("@", ()), [I(("m", ()))]),
       [Arg(
          N(
            Id(("x", ()),
              {id_info_id=7; id_flags=Ref(0); id_resolved=Ref(None);
               id_type=Ref(None); id_svalue=Ref(None); })))]), ())

This can be problematic when writing rules. For example, a struct wrapped around Base.@kwdef gets matched as a function call: https://semgrep.dev/playground/s/WAyDy. This can be avoided like this, but it seems a bit frail and I think they should be treated differently anyway.

The tree sitter seems to differentiate as well between function calls and macro calls, so I believe the problem is with the Semgrep generated AST.

As a note, macro definitions and function definitions are treated differently:

f(x) = x:

   DefStmt(
     ({
       name=EN(
              Id(("f", ()),
                {id_info_id=2; id_flags=Ref(0); id_resolved=Ref(None);
                 id_type=Ref(None); id_svalue=Ref(None); }));
       attrs=[]; tparams=None; },
      FuncDef(
        {fkind=(Function, ());
         fparams=[Param(
                    {pname=Some(("x", ())); pdefault=None; ptype=None;
                     pattrs=[];
                     pinfo={id_info_id=3; id_flags=Ref(0);
                            id_resolved=Ref(Some((Parameter, -1)));
                            id_type=Ref(None); id_svalue=Ref(None); };
                     })];
         frettype=None;
         fbody=FBExpr(
                 N(
                   Id(("x", ()),
                     {id_info_id=1; id_flags=Ref(0); id_resolved=Ref(
                      None); id_type=Ref(None); id_svalue=Ref(None); })));
         })))

macro m(x) :(x) end:

   DefStmt(
     ({
       name=EN(
              Id(("m", ()),
                {id_info_id=4; id_flags=Ref(0); id_resolved=Ref(None);
                 id_type=Ref(None); id_svalue=Ref(None); }));
       attrs=[]; tparams=None; },
      MacroDef(
        {macroparams=[("x", ())];
         macrobody=[S(
                      ExprStmt(
                        OtherExpr((":", ()),
                          [E(
                             N(
                               Id(("x", ()),
                                 {id_info_id=6; id_flags=Ref(0);
                                  id_resolved=Ref(None); id_type=Ref(
                                  None); id_svalue=Ref(None); })))]), ()))];
         })))

To Reproduce
The playground link above.

Expected behavior
Macro calls should be treated differently. Maybe they should not be matched with $X(...), but with @$X(...)?

What is the priority of the bug to you?

P0: blocking your adoption of Semgrep or workflow
P1: important to fix or quite annoying
P2: regular bug that should get fixed

Use case
This would make the generated AST more correct/complete.

The text was updated successfully, but these errors were encountered:

aryx added the lang:julia label Sep 4, 2024

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

[Julia] Macro calls are parsed the same as function calls? #10495

[Julia] Macro calls are parsed the same as function calls? #10495

[Julia] Macro calls are parsed the same as function calls? #10495

[Julia] Macro calls are parsed the same as function calls? #10495

Comments