Add basic support for starting ocaml repl with infer code

Summary: Make it possible to run infer code from within `ocaml`/`utop`. Integration is really basic, but we can extend it if we find it useful.

Reviewed By: jberdine

Differential Revision: D3736029

fbshipit-source-id: 4cebb7c

.gitignore

@@ -87,6 +87,7 @@ buck-out/
 /infer/lib/specs/clean_models
 /scripts/checkCopyright
 /infer/etc/clang_ast.dict
+/infer/src/toplevel.mlpack
 
 #atdgen stubs
 /infer/src/backend/jsonbug_*

Makefile

@@ -125,7 +125,10 @@ check_missing_mli:
 	for x in `find infer/src -name "*.ml"`; do \
 		test -f "$$x"i || echo Missing "$$x"i; done'
 
-test: test_build ocaml_unit_test buck_test inferTraceBugs_test
+inferScriptMode_test: toplevel
+	INFER_REPL_BINARY=ocaml ./scripts/infer_repl ./infer/tests/repl/infer_batch_script.ml
+
+test: test_build ocaml_unit_test buck_test inferTraceBugs_test inferScriptMode_test
 	$(MAKE) -C $(SRC_DIR) mod_dep.dot
 
 test_xml: test_build ocaml_unit_test buck_test_xml inferTraceBugs_test
@@ -133,6 +136,9 @@ test_xml: test_build ocaml_unit_test buck_test_xml inferTraceBugs_test
 
 quick-test: test_this_build ocaml_unit_test
 
+toplevel:
+	$(MAKE) -C $(SRC_DIR) toplevel
+
 uninstall:
 	$(REMOVE_DIR) $(DESTDIR)$(libdir)/infer/
 	$(REMOVE) $(DESTDIR)$(bindir)/inferTraceBugs
@@ -274,7 +280,7 @@ conf-clean: clean
 
 .PHONY: all buck_test buck_test_xml clean clang_plugin clang_setup infer inferTraceBugs
 .PHONY: inferTraceBugs_test install ocaml_unit_test check_missing_mli src_build test test_xml
-.PHONY: test_build uninstall
+.PHONY: test_build toplevel uninstall
 
 
 # print any variable for Makefile debugging

infer/src/Makefile

@@ -169,7 +169,7 @@ INFER_CONFIG_TARGETS += $(INFERLLVM_MAIN).native
 DEPENDENCIES += llvm
 endif
 
-.PHONY: all infer init version sanitize clean
+.PHONY: all infer init version sanitize clean toplevel
 
 all: infer
 
@@ -251,6 +251,27 @@ mod_dep.dot: $(DEPENDENCIES_DIR)/ocamldot/ocamldot $(ml_src_files) $(re_src_file
 mod_dep.pdf: mod_dep.dot
 	dot -Tpdf -o mod_dep.pdf mod_dep.dot
 
+roots_grep_regex:=$(foreach root,$(roots),-e $(root)$$)
+dirs_find_regex:=$(foreach dir, $(DEPENDENCIES),-path "./$(dir)/*" -o)
+toplevel: init $(STACKTREE_ATDGEN_STUBS) $(INFERPRINT_ATDGEN_STUBS) $(CLANG_ATDGEN_STUBS) $(INFER_CLANG_FCP_MIRRORED_FILES)
+	# We need to pack all infer modules into another module to avoid name clashes
+	# with some of them coming from ocaml libraries (Ident for example).
+	# To do that, we generate .mlpack file with source files. Steps:
+	# 1. find all interesting .re and .ml files - they need to be in one of
+	#    directories listed in $(DEPENDENCIES)
+	# 2. remove './' from the beginning of each line
+	# 3. remove extension from all files
+	# 4. make first letter of filename uppercase to produce valid ocaml module
+	# 5. filter out root modules since they run code when loading the module.
+	find . \( -name "*.ml" -o -name "*.mly" -o -name "*.mll" -o -name "*.re" \) \
+		\( $(dirs_find_regex) -false \) \
+		-not \( -path "./unit/*" -o -path "./llvm/*" -o -path "./facebook/scripts/eradicate_stats.ml" \) \
+	| cut -c 3- \
+	| rev | cut -f 2- -d '.' | rev \
+	| awk 'BEGIN { FS = "/"; OFS = "/" } ; {$$NF=toupper(substr($$NF,1,1))substr($$NF,2); print $$0}' \
+	| grep -v $(roots_grep_regex) > toplevel.mlpack
+	$(OCAMLBUILD_ALL) -build-dir $(INFER_BUILD_DIR) toplevel.cmo
+
 define gen_atdgen_rules
 # generate files using atdgen
 # parameters:

infer/src/backend/CommandLineOption.ml

@@ -17,7 +17,7 @@ module YBU = Yojson.Basic.Util
 
 (** Each command line option may appear in the --help list of any executable, these tags are used to
     specify which executables for which an option will be documented. *)
-type exe = Analyze | Clang | Java | Llvm | Print | StatsAggregator | Toplevel
+type exe = Analyze | Clang | Java | Llvm | Print | StatsAggregator | Toplevel | Interactive
 
 let exes = [
   ("InferAnalyze", Analyze);
@@ -27,13 +27,16 @@ let exes = [
   ("InferPrint", Print);
   ("InferStatsAggregator", StatsAggregator);
   ("infer", Toplevel);
+  ("interactive", Interactive);
 ]
 
 let all_exes = IList.map snd exes
 
 let current_exe =
-  try IList.assoc string_equal (Filename.basename Sys.executable_name) exes
-  with Not_found -> Toplevel
+  if !Sys.interactive then Interactive
+  else
+    try IList.assoc string_equal (Filename.basename Sys.executable_name) exes
+    with Not_found -> Toplevel
 
 
 type desc = {
@@ -442,7 +445,7 @@ let decode_env_to_argv env =
 
 let prepend_to_argv args =
   let cl_args = match Array.to_list Sys.argv with _ :: tl -> tl | [] -> [] in
-  (Sys.executable_name, args @ cl_args)
+  args @ cl_args
 
 (** [prefix_before_rest (prefix @ ["--" :: rest])] is [prefix] where "--" is not in [prefix]. *)
 let prefix_before_rest args =
@@ -553,7 +556,10 @@ let parse ?(incomplete=false) ?(accept_unknown=false) ?config_file env_var exe_u
       (try Unix.getenv "INFERCLANG_ARGS" with Not_found -> "") in
   let env_args = c_args @ env_args in
   (* end transitional support for INFERCLANG_ARGS *)
-  let exe_name, env_cl_args = prepend_to_argv env_args in
+  let exe_name = Sys.executable_name in
+  let env_cl_args = match current_exe with
+    | Interactive -> env_args
+    | _ -> prepend_to_argv env_args in
   let all_args = match config_file with
     | None -> env_cl_args
     | Some path ->
@@ -575,7 +581,7 @@ let parse ?(incomplete=false) ?(accept_unknown=false) ?config_file env_var exe_u
     | Arg.Bad usage_msg -> Pervasives.prerr_string usage_msg; exit 2
     | Arg.Help usage_msg -> Pervasives.print_string usage_msg; exit 0
   in
-  parse_loop () ;
+  parse_loop ();
   if not incomplete then
     Unix.putenv env_var (encode_argv_to_env (prefix_before_rest all_args)) ;
   curr_usage

infer/src/backend/CommandLineOption.mli

@@ -11,7 +11,7 @@
 
 open! Utils
 
-type exe = Analyze | Clang | Java | Llvm | Print | StatsAggregator | Toplevel
+type exe = Analyze | Clang | Java | Llvm | Print | StatsAggregator | Toplevel | Interactive
 
 val current_exe : exe
 

infer/src/backend/config.ml

@@ -1183,6 +1183,8 @@ let exe_usage (exe : CLOpt.exe) =
        Aggregates all the perf stats generated by Buck on each target"
   | Toplevel ->
       version_string
+  | Interactive ->
+      "Usage: interactive ocaml toplevel. To pass infer config options use env variable"
 
 let post_parsing_initialization () =
   F.set_margin !margin ;

infer/tests/repl/infer_batch_script.ml

@@ -0,0 +1,22 @@
+(*
+ * Copyright (c) 2016 - 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.
+ *)
+
+(* Example of ocaml script starting with infer code. To execute a scipt run:
+   ./scripts/infer_repl <path/to/this/script.ml>
+   It's used as a basic integration test *)
+
+(* "import" infer code *)
+#use "toplevel_init";;
+
+let _ = Ident.create_fresh Ident.knormal in
+let ident = Ident.create_fresh Ident.knormal in
+let e = Exp.Var ident in
+print_endline (Sil.exp_to_string e);
+(* pass --flavors flag to change the value *)
+print_endline (string_of_bool Config.flavors)

scripts/infer_repl

@@ -0,0 +1,20 @@
+#!/bin/bash
+
+SCRIPT_DIR="$( cd "$( dirname "${BASH_SOURCE[0]}" )" && pwd )"
+
+BUILD_DIR=$SCRIPT_DIR/../infer/_build/infer
+# to build new toplevel, run `make toplevel`
+# -init option is used only in interactive mode
+# in batch mode, scripts need to import toplevel_init themselves
+# It can be done by adding #use "toplevel_init";; to the beginning
+# of a script.
+# NOTE: $SCRIPT_DIR is added search path for batch scripts
+#       so they can be located anywhere and still find toplevel_init
+#       file. In interactive mode $SCRIPT_DIR isn't needed
+
+# by default utop is used, pass `INFER_UTOP_BINARY` to change the toplevel
+# binary (to `ocaml` for example)
+if [ -z "$INFER_UTOP_BINARY" ]; then
+  INFER_REPL_BINARY="utop"
+fi
+$INFER_REPL_BINARY -init $SCRIPT_DIR/toplevel_init -I $BUILD_DIR -I $SCRIPT_DIR $@

scripts/toplevel_init

@@ -0,0 +1,9 @@
+(* load dependencies *)
+#use "topfind";;
+#require "sawja";;
+#require "atdgen";;
+#require "extlib";;
+
+(* load infer code *)
+#load_rec "toplevel.cmo";;
+open Toplevel;;