[infer] python integration: parser, cfg support

Reviewed By: jvillard

Differential Revision: D5657644

fbshipit-source-id: 3c7d231

Makefile.autoconf.in

@@ -13,6 +13,7 @@ bindir = @bindir@
 BUCK = @BUCK@
 BUILD_C_ANALYZERS = @BUILD_C_ANALYZERS@
 BUILD_JAVA_ANALYZERS = @BUILD_JAVA_ANALYZERS@
+BUILD_PYTHON_ANALYZERS = @BUILD_PYTHON_ANALYZERS@
 CAML_LD_LIBRARY_PATH = @CAML_LD_LIBRARY_PATH@
 CC = @CC@
 CFLAGS = @CFLAGS@

configure.ac

@@ -70,6 +70,15 @@ AC_ARG_ENABLE(java-analyzers,
 BUILD_JAVA_ANALYZERS=$enable_java_analyzers
 AC_SUBST([BUILD_JAVA_ANALYZERS])
 
+AC_ARG_ENABLE(python-analyzers,
+  AS_HELP_STRING([--disable-python-analyzers],
+    [do not build the Python analyzers (default is to build them)]),
+  ,
+  enable_python_analyzers=yes)
+
+BUILD_PYTHON_ANALYZERS=$enable_python_analyzers
+AC_SUBST([BUILD_PYTHON_ANALYZERS])
+
 AC_ARG_WITH(fcp-clang,
   AS_HELP_STRING([--without-fcp-clang],
     [do not use $CLANG_PREFIX/bin/clang to override the default compiler (default is to override if in an infer release)]),

infer/src/Makefile

@@ -255,6 +255,7 @@ $(GENERATED_FROM_AUTOCONF): $(MAKEFILE_LIST)
 	  -e "s|@INFER_GIT_BRANCH[@]|$$INFER_GIT_BRANCH|g" \
 	  -e "s|@BUILD_C_ANALYZERS[@]|$(BUILD_C_ANALYZERS)|g" \
 	  -e "s|@BUILD_JAVA_ANALYZERS[@]|$(BUILD_JAVA_ANALYZERS)|g" \
+	  -e "s|@BUILD_PYTHON_ANALYZERS[@]|$(BUILD_PYTHON_ANALYZERS)|g" \
 	  -e "s|@OPAMSWITCH[@]|$(OPAMSWITCH)|g" \
 	  -e "s|@XCODE_SELECT[@]|$(XCODE_SELECT)|g" \
 	  -e "s|@INFER_MAN_LAST_MODIFIED[@]|$(INFER_MAN_LAST_MODIFIED)|g" \

infer/src/backend/interproc.ml

@@ -757,6 +757,8 @@ let prop_init_formals_seed tenv new_formals (prop: 'a Prop.t) : Prop.exposed Pro
          -> Exp.Sizeof {typ; nbytes= None; dynamic_length= None; subtype= Subtype.exact}
         | Config.Java
          -> Exp.Sizeof {typ; nbytes= None; dynamic_length= None; subtype= Subtype.subtypes}
+        | Config.Python
+         -> L.die InternalError "prop_init_formals_seed not implemented for Python"
       in
       Prop.mk_ptsto_lvar tenv Prop.Fld_init Sil.inst_formal (pv, texp, None)
     in

infer/src/backend/prover.ml

@@ -2306,6 +2306,8 @@ and sigma_imply tenv calc_index_frame calc_missing subs prop1 sigma2 : subst2 *
             ; "java.lang.String.value" ]
           in
           Sil.Estruct (List.map ~f:mk_fld_sexp fields, Sil.inst_none)
+      | Config.Python
+       -> L.die InternalError "mk_constant_string_hpred not implemented for Python"
     in
     let const_string_texp =
       match !Config.curr_language with
@@ -2322,6 +2324,8 @@ and sigma_imply tenv calc_index_frame calc_missing subs prop1 sigma2 : subst2 *
             ; nbytes= None
             ; dynamic_length= None
             ; subtype= Subtype.exact }
+      | Config.Python
+       -> L.die InternalError "const_string_texp not implemented for Python"
     in
     Sil.Hpointsto (root, sexp, const_string_texp)
   in

infer/src/backend/rearrange.ml

@@ -462,6 +462,8 @@ let mk_ptsto_exp_footprint pname tenv orig_prop (lexp, typ) max_stamp inst
      -> Subtype.exact
     | Config.Java
      -> Subtype.subtypes
+    | Config.Python
+     -> L.die InternalError "Subtypes for Python not implemented"
   in
   let create_ptsto footprint_part off0 =
     match (root, off0, typ.Typ.desc) with

infer/src/base/Config.ml

@@ -43,11 +43,11 @@ let string_of_analyzer a =
 let clang_frontend_action_symbols =
   [("lint", `Lint); ("capture", `Capture); ("lint_and_capture", `Lint_and_capture)]
 
-type language = Clang | Java [@@deriving compare]
+type language = Clang | Java | Python [@@deriving compare]
 
 let equal_language = [%compare.equal : language]
 
-let string_of_language = function Java -> "Java" | Clang -> "C_CPP"
+let string_of_language = function Java -> "Java" | Clang -> "C_CPP" | Python -> "python"
 
 let ml_bucket_symbols =
   [ ("all", `MLeak_all)

infer/src/base/Config.mli

@@ -30,7 +30,7 @@ val string_to_analyzer : (string * analyzer) list
 
 val string_of_analyzer : analyzer -> string
 
-type language = Clang | Java [@@deriving compare]
+type language = Clang | Java | Python [@@deriving compare]
 
 val equal_language : language -> language -> bool
 

infer/src/base/Version.ml.in

@@ -45,6 +45,8 @@ let clang_enabled = is_yes "@BUILD_C_ANALYZERS@"
 
 let java_enabled = is_yes "@BUILD_JAVA_ANALYZERS@"
 
+let python_enabled = is_yes "@BUILD_PYTHON_ANALYZERS@"
+
 let xcode_enabled = is_not_no "@XCODE_SELECT@"
 
 let man_pages_last_modify_date = "@INFER_MAN_LAST_MODIFIED@"

infer/src/base/Version.mli

@@ -19,6 +19,8 @@ val clang_enabled : bool
 
 val java_enabled : bool
 
+val python_enabled : bool
+
 val xcode_enabled : bool
 
 val man_pages_last_modify_date : string

infer/src/integration/Driver.ml

@@ -26,6 +26,7 @@ type build_system =
   | BMake
   | BMvn
   | BNdk
+  | BPython
   | BXcode
   [@@deriving compare]
 
@@ -55,6 +56,7 @@ let build_system_exe_assoc =
   ; (BMvn, "mvn")
   ; (BMvn, "mvnw")
   ; (BNdk, "ndk-build")
+  ; (BPython, "python")
   ; (BXcode, "xcodebuild") ]
 
 let build_system_of_exe_name name =
@@ -73,6 +75,7 @@ type mode =
   | ClangCompilationDB of [`Escaped of string | `Raw of string] list
   | Javac of Javac.compiler * string * string list
   | Maven of string * string list
+  | Python of string list
   | PythonCapture of build_system * string list
   | XcodeXcpretty of string * string list
   [@@deriving compare]
@@ -92,6 +95,7 @@ let pp_mode fmt mode =
   | BuckGenrule _
   | BuckCompilationDB _
   | ClangCompilationDB _
+  | Python _
   | PythonCapture (_, _)
   | XcodeXcpretty _
    -> (* these are pretty boring, do not log anything *)
@@ -268,6 +272,9 @@ let capture ~changed_files = function
    -> L.progress "Capturing in javac mode...@." ; Javac.capture compiler ~prog ~args
   | Maven (prog, args)
    -> L.progress "Capturing in maven mode...@." ; Maven.capture ~prog ~args
+  | Python args
+   -> (* pretend prog is the root directory of the project *)
+      PythonMain.go args
   | PythonCapture (build_system, build_cmd)
    -> L.progress "Capturing in %s mode...@." (string_of_build_system build_system) ;
       let in_buck_mode = equal_build_system build_system BBuck in
@@ -461,6 +468,8 @@ let assert_supported_mode required_analyzer requested_mode_string =
      -> Version.clang_enabled
     | `Java
      -> Version.java_enabled
+    | `Python
+     -> Version.python_enabled
     | `Xcode
      -> Version.clang_enabled && Version.xcode_enabled
   in
@@ -471,6 +480,8 @@ let assert_supported_mode required_analyzer requested_mode_string =
        -> "clang"
       | `Java
        -> "java"
+      | `Python
+       -> "python"
       | `Xcode
        -> "clang and xcode"
     in
@@ -484,6 +495,8 @@ let assert_supported_build_system build_system =
    -> string_of_build_system build_system |> assert_supported_mode `Java
   | BClang | BMake | BNdk
    -> string_of_build_system build_system |> assert_supported_mode `Clang
+  | BPython
+   -> string_of_build_system build_system |> assert_supported_mode `Python
   | BXcode
    -> string_of_build_system build_system |> assert_supported_mode `Xcode
   | BBuck
@@ -528,6 +541,8 @@ let mode_of_build_command build_cmd =
        -> Javac (Javac.Javac, prog, args)
       | BMvn
        -> Maven (prog, args)
+      | BPython
+       -> Python args
       | BXcode when Config.xcpretty
        -> XcodeXcpretty (prog, args)
       | BAnt | BBuck | BGradle | BNdk | BXcode as build_system

infer/src/integration/Driver.mli

@@ -23,6 +23,7 @@ type mode =
   | ClangCompilationDB of [`Escaped of string | `Raw of string] list
   | Javac of Javac.compiler * string * string list
   | Maven of string * string list
+  | Python of string list
   | PythonCapture of build_system * string list
   | XcodeXcpretty of string * string list
   [@@deriving compare]

infer/src/jbuild.common.in

@@ -7,6 +7,8 @@ let clang = is_yes "@BUILD_C_ANALYZERS@"
 
 let java = is_yes "@BUILD_JAVA_ANALYZERS@"
 
+let python = is_yes "@BUILD_PYTHON_ANALYZERS@"
+
 let facebook = is_yes "@IS_FACEBOOK_TREE@"
 
 let extra_cflags = if "@EXTRA_CFLAGS" = "" then [] else ["@EXTRA_CFLAGS@"]

infer/src/jbuild.in

@@ -17,6 +17,7 @@ let sources =
     :: ( ( if clang then ["clang"; ("unit" ^/ "clang")]
          else ["clang_stubs"; ("unit" ^/ "clang_stubs")] )
        @ [ (if java then "java" else "java_stubs")
+         ; (if python && facebook then "python" else "python_stubs")
          ; "absint"
          ; "backend"
          ; "base"
@@ -66,6 +67,8 @@ let stanzas =
   ( if clang then
       ["(ocamllex (types_lexer ctl_lexer))"; "(menhir ((modules (types_parser ctl_parser))))"]
   else [] )
+  @ ( if python && facebook then ["(ocamllex (pythonLexer))"; "(menhir ((modules (pythonParser))))"]
+    else [] )
   @ [ Format.sprintf
         {|
 (library
@@ -73,7 +76,7 @@ let stanzas =
   (flags (%s))
   (libraries (%s))
   (modules (:standard \ %s infertop))
-  (preprocess (pps (ppx_compare)))
+  (preprocess (pps (ppx_compare ppx_sexp_conv -no-check)))
  ))
 |}
         (String.concat " " infer_cflags) (String.concat " " infer_libraries)
@@ -85,7 +88,7 @@ let stanzas =
   (flags (%s -open InferModules))
   (libraries (InferModules))
   (modules (%s))
-  (preprocess (pps (ppx_compare)))
+  (preprocess (pps (ppx_compare ppx_sexp_conv -no-check)))
  ))
 |}
         (String.concat " " infer_binaries) (String.concat " " infer_cflags)

infer/src/python_stubs/pythonMain.ml

@@ -0,0 +1,10 @@
+(*
+ * Copyright (c) 2017 - present Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under the BSD style license found in the
+ * LICENSE file in the root directory of this source tree. An additional grant
+ * of patent rights can be found in the PATENTS file in the same directory.
+ *)
+
+let go _ = ()

infer/src/python_stubs/pythonMain.mli

@@ -0,0 +1,10 @@
+(*
+ * Copyright (c) 2017 - present Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under the BSD style license found in the
+ * LICENSE file in the root directory of this source tree. An additional grant
+ * of patent rights can be found in the PATENTS file in the same directory.
+ *)
+
+val go : string list -> unit