Improve onSymbol API

nimterop/cimport.nim

@@ -174,18 +174,33 @@ macro cSkipSymbol*(skips: varargs[string]): untyped =
 
 macro cPlugin*(body): untyped =
   ## When `cOverride() <cimport.html#cOverride.m,>`_ and `cSkipSymbol() <cimport.html#cSkipSymbol.m%2Cvarargs[string]>`_
-  ## are not adequate, the `cPlugin() <cimport.html#cPlugin.m,>`_ macro can be used to customize the generated Nim output.
+  ## are not adequate, the `cPlugin() <cimport.html#cPlugin.m,>`_ macro can be used
-  ## The following callbacks are available at this time.
+  ## to customize the generated Nim output. The following callbacks are available at
+  ## this time.
   ##
+  ## .. code-block:: nim
+  ##
+  ##     proc onSymbol(sym: var Symbol): Result {.exportc, dynlib.}
+  ##
+  ## `onSymbol()` can be used to handle symbol name modifications required due to invalid
+  ## characters like `_` or to rename duplicate types. It can also be used to remove prefixes
+  ## and suffixes. The symbol name and type is provided to the callback and the name can be
+  ## modified. Symbol types can be any of the following:
+  ## - `nskConst` for constants
+  ## - `nskType` for type identifiers, including primitive
+  ## - `nskParam` for param names
+  ## - `nskField` for struct field names
+  ## - `nskEnumField` for enum (field) names, though they are in the global namespace as `nskConst`
+  ## - `nskProc` - for proc names
   runnableExamples:
     cPlugin:
       import strutils
 
-      proc onSymbol*(sym: string): string {.exportc, dynlib.} =
+      proc onSymbol*(sym: var Symbol): Result {.exportc, dynlib.} =
-        return sym.strip(chars={'_'})
+        sym.name = sym.name.strip(chars={'_'})
 
   let
-    data = body.repr
+    data = "import nimterop/cimport\n\n" & body.repr
     hash = data.hash()
     phash = if hash<0: -hash else: hash
     path = getTempDir() / "nimterop_" & $phash & ".nim"

nimterop/getters.nim

@@ -90,11 +90,17 @@ proc getType*(str: string): string =
 template checkUnderscores(str, errmsg: string): untyped =
   doAssert str[0] != '_' and str[^1] != '_', errmsg
 
-proc getIdentifier*(str: string): string =
+proc getIdentifier*(str: string, kind: NimSymKind): string =
   doAssert str.len != 0, "Blank identifier error"
 
   if gStateRT.onSymbol != nil:
-    result = gStateRT.onSymbol(str)
+    var
+      sym = Symbol(name: str, kind: kind)
+      res = gStateRT.onSymbol(sym)
+
+    doAssert res.error == 0, res.message
+
+    result = sym.name
     checkUnderscores(result, &"Identifier '{str}' still contains leading/trailing underscores '_'  after 'cPlugin:onSymbol()': result '{result}'")
   else:
     result = str

nimterop/globals.nim

@@ -45,9 +45,13 @@ type
       tonim*: proc (ast: ref Ast, node: TSNode)
     regex*: Regex
 
-  Symbol = object
+  Symbol* = object
-    name: string
+    name*: string
-    kind: NimSymKind
+    kind*: NimSymKind
+
+  Result* = object
+    error*: int
+    message*: string
 
   State = object
     compile*, defines*, headers*, includeDirs*, searchDirs*, symOverride*: seq[string]
@@ -63,7 +67,7 @@ type
     when not declared(CIMPORT):
       grammar*: seq[tuple[grammar: string, call: proc(ast: ref Ast, node: TSNode) {.nimcall.}]]
 
-    onSymbol*: proc(sym: string): string {.cdecl.}
+    onSymbol*: proc(sym: var Symbol): Result {.cdecl.}
 var
   gStateCT {.compiletime, used.}: State
   gStateRT {.used.}: State
@@ -79,4 +83,4 @@ type CompileMode = enum
 const modeDefault {.used.} = $cpp # TODO: USE this everywhere relevant
 
 when not declared(CIMPORT):
-  export gAtoms, gExpressions, gEnumVals, Kind, Ast, Symbol, State, gStateRT, nBl, CompileMode, modeDefault
+  export gAtoms, gExpressions, gEnumVals, Kind, Ast, State, gStateRT, nBl, CompileMode, modeDefault

nimterop/grammar.nim

@@ -1,4 +1,4 @@
-import sets, strformat, strutils, tables
+import macros, sets, strformat, strutils, tables
 
 import regex
 
@@ -14,7 +14,7 @@ proc initGrammar() =
   """,
     proc (ast: ref Ast, node: TSNode) =
       let
-        name = gStateRT.data[0].val.getIdentifier()
+        name = gStateRT.data[0].val.getIdentifier(nskConst)
         val = gStateRT.data[1].val.getLit()
 
       if val.nBl and gStateRT.consts.addNewIdentifer(name):
@@ -67,7 +67,7 @@ proc initGrammar() =
     """
 
   template funcParamCommon(pname, ptyp, pptr, pout, count, i: untyped): untyped =
-    ptyp = gStateRT.data[i].val.getIdentifier()
+    ptyp = gStateRT.data[i].val.getIdentifier(nskType)
     if i+1 < gStateRT.data.len and gStateRT.data[i+1].name == "pointer_declarator":
       pptr = "ptr "
       i += 1
@@ -75,7 +75,7 @@ proc initGrammar() =
       pptr = ""
 
     if i+1 < gStateRT.data.len and gStateRT.data[i+1].name == "identifier":
-      pname = gStateRT.data[i+1].val.getIdentifier()
+      pname = gStateRT.data[i+1].val.getIdentifier(nskParam)
       i += 2
     else:
       pname = "a" & $count
@@ -104,7 +104,7 @@ proc initGrammar() =
     proc (ast: ref Ast, node: TSNode) =
       var
         i = 0
-        typ = gStateRT.data[i].val.getIdentifier()
+        typ = gStateRT.data[i].val.getIdentifier(nskType)
         name = ""
         tptr = ""
         aptr = ""
@@ -120,7 +120,7 @@ proc initGrammar() =
             i += 1
 
       if i < gStateRT.data.len:
-        name = gStateRT.data[i].val.getIdentifier()
+        name = gStateRT.data[i].val.getIdentifier(nskType)
         i += 1
 
       if gStateRT.types.addNewIdentifer(name):
@@ -144,8 +144,10 @@ proc initGrammar() =
             gStateRT.typeStr &= &"  {name}* = proc({pout}) {{.nimcall.}}\n"
         else:
           if i < gStateRT.data.len and gStateRT.data[i].name in ["identifier", "number_literal"]:
-            let
+            var
-              flen = gStateRT.data[i].val.getIdentifier()
+              flen = gStateRT.data[i].val
+            if gStateRT.data[i].name == "identifier":
+              flen = flen.getIdentifier(nskConst)
             gStateRT.typeStr &= &"  {name}* = {aptr}array[{flen}, {getPtrType(tptr&typ)}]\n"
           else:
             if name == typ:
@@ -157,7 +159,7 @@ proc initGrammar() =
   proc pDupTypeCommon(nname: string, fend: int, isEnum=false) =
     var
       dname = gStateRT.data[^1].val
-      ndname = gStateRT.data[^1].val.getIdentifier()
+      ndname = gStateRT.data[^1].val.getIdentifier(nskType)
       dptr =
         if fend == 2:
           "ptr "
@@ -175,7 +177,7 @@ proc initGrammar() =
 
   proc pStructCommon(ast: ref Ast, node: TSNode, name: string, fstart, fend: int) =
     var
-      nname = name.getIdentifier()
+      nname = name.getIdentifier(nskType)
       prefix = ""
       union = ""
 
@@ -231,7 +233,7 @@ proc initGrammar() =
             aptr = "ptr "
             i += 1
 
-        fname = gStateRT.data[i].val.getIdentifier()
+        fname = gStateRT.data[i].val.getIdentifier(nskField)
         if i+1 < gStateRT.data.len-fend and gStateRT.data[i+1].name in gEnumVals:
           let
             flen = gStateRT.data[i+1].val.getNimExpression()
@@ -348,7 +350,7 @@ proc initGrammar() =
       if name.len == 0:
         getUniqueIdentifier(gStateRT.enums, "Enum")
       else:
-        name.getIdentifier()
+        name.getIdentifier(nskType)
 
     if gStateRT.enums.addNewIdentifer(nname):
       gStateRT.enumStr &= &"\ntype {nname}* = distinct int"
@@ -363,7 +365,7 @@ proc initGrammar() =
           continue
 
         let
-          fname = gStateRT.data[i].val.getIdentifier()
+          fname = gStateRT.data[i].val.getIdentifier(nskEnumField)
 
         if i+1 < gStateRT.data.len-fend and
           gStateRT.data[i+1].name in gEnumVals:
@@ -447,7 +449,7 @@ proc initGrammar() =
   """,
     proc (ast: ref Ast, node: TSNode) =
       var
-        ftyp = gStateRT.data[0].val.getIdentifier()
+        ftyp = gStateRT.data[0].val.getIdentifier(nskType)
         fptr = ""
         i = 1
 
@@ -464,7 +466,7 @@ proc initGrammar() =
 
         var
           fname = gStateRT.data[i].val
-          fnname = fname.getIdentifier()
+          fnname = fname.getIdentifier(nskProc)
           pout, pname, ptyp, pptr = ""
           count = 1
 

tests/tmath.nim

@@ -14,8 +14,8 @@ cAddStdDir()
 cPlugin:
   import strutils
 
-  proc onSymbol*(sym: string): string {.exportc, dynlib.} =
+  proc onSymbol*(sym: var Symbol): Result {.exportc, dynlib.} =
-    return sym.strip(chars={'_'})
+    sym.name = sym.name.strip(chars={'_'})
 
 cImport cSearchPath("math.h")
 

tests/tnimterop_c.nim

@@ -12,8 +12,8 @@ cCompile cSearchPath("test.c")
 cPlugin:
   import strutils
 
-  proc onSymbol*(sym: string): string {.exportc, dynlib.} =
+  proc onSymbol*(sym: var Symbol): Result {.exportc, dynlib.} =
-    return sym.strip(chars={'_'})
+    sym.name = sym.name.strip(chars={'_'})
 
 cImport cSearchPath "test.h"