[DB.source_file] Explicitly specify what path is relative to, keep cpp model paths relative

Summary:
1. Always store cpp model source_file with relative path. This will make them cache friendly independent of infer location
2. Distinguish between "relative to project root" and "relative to infer models src"
3. Unify `source_file_from_path` used by java and C frontends into one function. There are no improvements to that logic yet
4. Move `is_cpp_model_file` to use `source_file` instead of `filename`

Reviewed By: jberdine

Differential Revision: D4204548

fbshipit-source-id: 6e21771

infer/src/IR/AttributesTable.re

@@ -112,9 +112,7 @@ let get_correct_type_from_objc_class_name type_name =>
 /** Returns true if the method is defined as a C++ model */
 let pname_is_cpp_model callee_pname =>
   switch (load_attributes callee_pname) {
-  | Some attrs =>
-    let file = DB.source_file_to_string attrs.ProcAttributes.loc.Location.file;
-    DB.file_is_in_cpp_model file
+  | Some attrs => DB.source_file_is_cpp_model attrs.ProcAttributes.loc.Location.file
   | None => false
   };
 

infer/src/backend/InferPrint.re

@@ -489,18 +489,6 @@ let module IssuesJson = {
     } else {
       file
     };
-  let make_cpp_models_path_relative file => {
-    let abs_file = DB.source_file_to_abs_path file;
-    if (Utils.string_is_prefix Config.cpp_models_dir abs_file) {
-      if (Config.debug_mode || Config.debug_exceptions) {
-        Some (DB.rel_source_file_from_abs_path Config.cpp_models_dir abs_file)
-      } else {
-        None
-      }
-    } else {
-      Some file
-    }
-  };
 
   /** Write bug report in JSON format */
   let pp_issues_of_error_log fmt error_filter _ proc_loc_opt procname err_log => {
@@ -522,16 +510,19 @@ let module IssuesJson = {
         | Some proc_loc => (proc_loc.Location.file, proc_loc.Location.line)
         | None => (loc.Location.file, 0)
         };
-      let file_opt = make_cpp_models_path_relative source_file;
+      let should_report_source_file =
+        not (DB.source_file_is_infer_model source_file) ||
+        Config.debug_mode || Config.debug_exceptions;
       if (
         in_footprint &&
         error_filter source_file error_desc error_name &&
-        Option.is_some file_opt && should_report ekind error_name error_desc && is_file source_file
+        should_report_source_file &&
+        should_report ekind error_name error_desc && is_file source_file
       ) {
         let kind = Exceptions.err_kind_string ekind;
         let bug_type = Localise.to_string error_name;
         let procedure_id = Procname.to_filename procname;
-        let file = expand_links_under_buck_out (DB.source_file_to_string (Option.get file_opt));
+        let file = expand_links_under_buck_out (DB.source_file_to_string source_file);
         let json_ml_loc =
           switch ml_loc_opt {
           | Some (file, lnum, cnum, enum) when Config.reports_include_ml_loc =>

infer/src/base/Config.ml

@@ -259,12 +259,16 @@ let models_dir =
 let models_jar =
   lib_dir // "java" // "models.jar"
 
-let cpp_extra_include_dir =
-  let dir = bin_dir // Filename.parent_dir_name // "models" // "cpp" // "include" in
+let models_src_dir =
+  let dir = bin_dir // Filename.parent_dir_name // "models" in
   Utils.filename_to_absolute dir (* Normalize the path *)
 
-let cpp_models_dir =
-  cpp_extra_include_dir // "infer_model"
+let relative_cpp_extra_include_dir = "cpp" // "include"
+
+let cpp_extra_include_dir = models_src_dir // relative_cpp_extra_include_dir
+
+let relative_cpp_models_dir =
+  relative_cpp_extra_include_dir // "infer_model"
 
 let wrappers_dir =
   lib_dir // "wrappers"

infer/src/base/Config.mli

@@ -62,7 +62,7 @@ val captured_dir_name : string
 val checks_disabled_by_default : string list
 val clang_initializer_prefix : string
 val cpp_extra_include_dir : string
-val cpp_models_dir : string
+val relative_cpp_models_dir : string
 val csl_analysis : bool
 val current_exe : CommandLineOption.exe
 val default_failure_name : string
@@ -93,6 +93,7 @@ val max_recursion : int
 val meet_level : int
 val models_dir : string
 val models_jar : string
+val models_src_dir : string
 val multicore_dir_name : string
 val ncpu : int
 val nsnotification_center_checker_backend : bool

infer/src/base/DB.ml

@@ -24,14 +24,18 @@ let count_newlines (path: string): int =
 
 type source_file =
   | Absolute of string
-  | Relative of string
+  | RelativeProjectRoot of string (* relative to project root *)
+  | RelativeInferModel of string (* relative to infer models *)
 
 let source_file_compare sf1 sf2 =
   match sf1, sf2 with
   | Absolute p1, Absolute p2 -> string_compare p1 p2
   | Absolute _, _ -> -1
   | _, Absolute _ -> 1
-  | Relative p1, Relative p2 -> string_compare p1 p2
+  | RelativeProjectRoot p1, RelativeProjectRoot p2 -> string_compare p1 p2
+  | RelativeProjectRoot _, _ -> -1
+  | _, RelativeProjectRoot _ -> 1
+  | RelativeInferModel p1, RelativeInferModel p2 -> string_compare p1 p2
 
 let source_file_equal sf1 sf2 =
   (source_file_compare sf1 sf2) = 0
@@ -48,7 +52,7 @@ module SourceFileSet = Set.Make(OrderedSourceFile)
 
 let source_file_from_string path =
   if Filename.is_relative path then
-    Relative path
+    RelativeProjectRoot path
   else
     Absolute path
 
@@ -56,27 +60,40 @@ let source_file_from_string path =
 let abs_source_file_from_path fname =
   Absolute (filename_to_absolute fname)
 
-(** convert a project root directory and a full path to a rooted source file *)
-let rel_source_file_from_abs_path root fname =
+let rel_path_from_abs_path root fname =
   let relative_complemented_fname = filename_to_relative root fname in
   if string_is_prefix root fname &&
      Filename.is_relative relative_complemented_fname then
-    Relative relative_complemented_fname
-  else
-    (* The project root is not a prefix of the file name *)
-    abs_source_file_from_path fname
+    Some relative_complemented_fname
+  else None (* The project root is not a prefix of the file name *)
+
+(** convert a project root directory and a full path to a rooted source file *)
+let source_file_from_abs_path fname =
+  if Filename.is_relative fname then
+    (failwithf
+       "ERROR: Path %s is relative, when absolute path was expected .@."
+       fname);
+  match rel_path_from_abs_path Config.project_root fname with
+  | Some path -> RelativeProjectRoot path
+  | None -> (
+      match rel_path_from_abs_path Config.models_src_dir fname with
+      | Some path -> RelativeInferModel path
+      | None -> Absolute fname (* fname is absolute already *)
+    )
 
 let curr_encoding = `Enc_crc
 
 let source_file_to_string fname =
   match fname with
-  | Relative path
+  | RelativeProjectRoot path
+  | RelativeInferModel path
   | Absolute path -> path
 
 (* Checking if the path exists may be needed only in some cases, hence the flag check_exists *)
 let source_file_to_abs_path fname =
   match fname with
-  | Relative path -> Filename.concat Config.project_root path
+  | RelativeProjectRoot path -> Filename.concat Config.project_root path
+  | RelativeInferModel path -> Filename.concat Config.models_src_dir path
   | Absolute path -> path
 
 let source_file_line_count source_file =
@@ -90,24 +107,39 @@ let inode_equal sf1 sf2 =
 
 let source_file_to_rel_path fname =
   match fname with
-  | Relative path -> path
-  | Absolute path -> filename_to_relative Config.project_root path
+  | RelativeProjectRoot path -> path
+  | _ -> source_file_to_abs_path fname |> filename_to_relative Config.project_root
 
 (** string encoding of a source file (including path) as a single filename *)
 let source_file_encoding source_file =
+  let prefix = match source_file with
+    | RelativeProjectRoot _ -> "P"
+    | RelativeInferModel _ -> "MOD"
+    | Absolute _ -> "ABS" in
   let source_file_s = source_file_to_string source_file in
   match curr_encoding with
   | `Enc_base ->
       Filename.basename source_file_s
   | `Enc_path_with_underscores ->
-      Escape.escape_path source_file_s
+      prefix ^ Escape.escape_path source_file_s
   | `Enc_crc ->
       let base = Filename.basename source_file_s in
-      let dir = Filename.dirname source_file_s in
+      let dir = prefix ^ Filename.dirname source_file_s in
       string_append_crc_cutoff ~key:dir base
 
 let source_file_empty = Absolute ""
 
+let source_file_is_infer_model source_file = match source_file with
+  | RelativeProjectRoot _ | Absolute _ -> false
+  | RelativeInferModel _ -> true
+
+(** Returns true if the file is a C++ model *)
+let source_file_is_cpp_model file =
+  match file with
+  | RelativeInferModel path ->
+      string_is_prefix Config.relative_cpp_models_dir path
+  | _ -> false
+
 (** {2 Source Dirs} *)
 
 (** source directory: the directory inside the results dir corresponding to a source file *)
@@ -347,13 +379,6 @@ let mark_file_updated fname =
   let near_future = Unix.gettimeofday () +. 1. in
   Unix.utimes fname near_future near_future
 
-
-(** Returns true if the file is a C++ model *)
-let file_is_in_cpp_model file =
-  let normalized_file_dir = filename_to_absolute (Filename.dirname file) in
-  let normalized_cpp_models_dir = filename_to_absolute Config.cpp_models_dir in
-  string_is_prefix normalized_cpp_models_dir normalized_file_dir
-
 (** Fold over all file paths recursively under [dir] which match [p]. *)
 let fold_paths_matching ~dir ~p ~init ~f =
   let rec paths path_list dir =

infer/src/base/DB.mli

@@ -104,8 +104,7 @@ val source_file_empty : source_file
 (** convert a path to a source file, turning it into an absolute path if necessary *)
 val abs_source_file_from_path : string -> source_file
 
-(** convert a project root directory and an absolute path to a source file *)
-val rel_source_file_from_abs_path : string -> string -> source_file
+val source_file_from_abs_path : string -> source_file
 
 (** string encoding of a source file (including path) as a single filename *)
 val source_file_encoding : source_file -> string
@@ -122,6 +121,11 @@ val source_file_to_abs_path : source_file -> string
 (** get the relative path of a source file *)
 val source_file_to_rel_path : source_file -> string
 
+val source_file_is_infer_model : source_file -> bool
+
+(** Returns true if the file is a C++ model *)
+val source_file_is_cpp_model : source_file -> bool
+
 (** {2 Source Dirs} *)
 
 (** source directory: the directory inside the results dir corresponding to a source file *)
@@ -162,9 +166,6 @@ val global_tenv_fname : filename
 (** Check if a path is a Java, C, C++ or Objectve C source file according to the file extention *)
 val is_source_file: string -> bool
 
-(** Returns true if the file is a C++ model *)
-val file_is_in_cpp_model : string -> bool
-
 (** Fold over all file paths recursively under [dir] which match [p]. *)
 val fold_paths_matching :
   dir:filename -> p:(filename -> bool) -> init:'a -> f:(filename -> 'a -> 'a) -> 'a

infer/src/clang/Capture.re

@@ -67,7 +67,7 @@ let run_clang_frontend ast_source => {
     switch ast_decl {
     | Clang_ast_t.TranslationUnitDecl (_, _, _, info) =>
       Config.arc_mode := info.Clang_ast_t.tudi_arc_enabled;
-      let source_file = CLocation.source_file_from_path info.Clang_ast_t.tudi_input_path;
+      let source_file = DB.source_file_from_abs_path info.Clang_ast_t.tudi_input_path;
       init_global_state_for_capture_and_linters source_file;
       let lang =
         switch info.Clang_ast_t.tudi_input_kind {

infer/src/clang/ComponentKit.ml

@@ -26,7 +26,7 @@ let is_in_main_file translation_unit_context an =
       false
   | Some file ->
       DB.source_file_equal
-        (CLocation.source_file_from_path file)
+        (DB.source_file_from_abs_path file)
         translation_unit_context.CFrontend_config.source_file
 
 let is_ck_context (context: CLintersContext.context) an =

infer/src/clang/cFrontend_utils.ml

@@ -658,7 +658,7 @@ struct
   let get_rel_file_path file_opt =
     match file_opt with
     | Some file ->
-        DB.source_file_to_rel_path (DB.rel_source_file_from_abs_path Config.project_root file)
+        DB.source_file_to_string (DB.source_file_from_abs_path file)
     | None -> ""
 
   let is_cpp_translation translation_unit_context =

infer/src/clang/cLocation.ml

@@ -11,21 +11,10 @@ open! Utils
 
 (** Module for function to retrieve the location (file, line, etc) of instructions *)
 
-let source_file_from_path path =
-  if Filename.is_relative path then
-    (failwithf
-       "ERROR: Path %s is relative. Please pass an absolute path in the -c argument.@."
-       path);
-  try
-    DB.rel_source_file_from_abs_path Config.project_root path
-  with Failure _ ->
-    Logging.err_debug "ERROR: %s should be a prefix of %s.@." Config.project_root path;
-    DB.source_file_from_string path
-
 let clang_to_sil_location trans_unit_ctx clang_loc =
   let line = Option.default (-1) clang_loc.Clang_ast_t.sl_line in
   let col = Option.default (-1) clang_loc.Clang_ast_t.sl_column in
-  let file = Option.map_default source_file_from_path
+  let file = Option.map_default DB.source_file_from_abs_path
       trans_unit_ctx.CFrontend_config.source_file clang_loc.Clang_ast_t.sl_file in
   Location.{line; col; file}
 
@@ -45,7 +34,7 @@ let should_do_frontend_check trans_unit_ctx (loc_start, _) =
   match loc_start.Clang_ast_t.sl_file with
   | Some file ->
       let equal_current_source file =
-        DB.source_file_equal (source_file_from_path file)
+        DB.source_file_equal (DB.source_file_from_abs_path file)
           trans_unit_ctx.CFrontend_config.source_file in
       equal_current_source file ||
       (file_in_project file &&  not Config.testing_mode)
@@ -62,7 +51,7 @@ let should_translate trans_unit_ctx (loc_start, loc_end) decl_trans_context ~tra
     | None -> false
   in
   let map_file_of pred loc =
-    let path_pred path = pred (source_file_from_path path) in
+    let path_pred path = pred (DB.source_file_from_abs_path path) in
     map_path_of path_pred loc
   in
   (* it's not necessary to compare inodes here because both files come from
@@ -73,8 +62,8 @@ let should_translate trans_unit_ctx (loc_start, loc_end) decl_trans_context ~tra
   let file_in_project = map_path_of file_in_project loc_end
                         || map_path_of file_in_project loc_start in
   let translate_on_demand = translate_when_used || file_in_project || Config.models_mode in
-  let file_in_models = map_path_of DB.file_is_in_cpp_model loc_end
-                       || map_path_of DB.file_is_in_cpp_model loc_start in
+  let file_in_models = map_file_of DB.source_file_is_cpp_model loc_end
+                       || map_file_of DB.source_file_is_cpp_model loc_start in
   map_file_of equal_current_source loc_end
   || map_file_of equal_current_source loc_start
   || file_in_models

infer/src/clang/cLocation.mli

@@ -24,7 +24,5 @@ val should_do_frontend_check : CFrontend_config.translation_unit_context ->
 
 val is_file_blacklisted : string -> bool
 
-val source_file_from_path : string -> DB.source_file
-
 val get_sil_location_from_range : CFrontend_config.translation_unit_context ->
   Clang_ast_t.source_range -> bool -> Location.t

infer/src/java/jClasspath.ml

@@ -72,13 +72,6 @@ let is_model procname =
 let split_classpath cp = Str.split (Str.regexp JFile.sep) cp
 
 
-let java_source_file_from_path path =
-  if Filename.is_relative path then
-    failwith "Expect absolute path for java source files"
-  else
-    DB.rel_source_file_from_abs_path Config.project_root path
-
-
 (** Add the android.jar containing bytecode at the beginning of the class path *)
 let add_android_jar paths =
   AndroidFramework.non_stub_android_jar () :: paths
@@ -130,7 +123,7 @@ let add_source_file path map =
   let basename = Filename.basename path in
   let entry =
     let current_source_file =
-      java_source_file_from_path (convert_to_absolute path) in
+      DB.source_file_from_abs_path (convert_to_absolute path) in
     try
       match StringMap.find basename map with
       | Singleton previous_source_file ->

infer/src/java/jClasspath.mli

@@ -25,10 +25,6 @@ val add_models : string -> unit
 (** Check if there is a model for the given procname *)
 val is_model : Procname.t -> bool
 
-(** create a source file from an absolute path.
-    Source files are relative if the project root is specified and absolute otherwise *)
-val java_source_file_from_path : string -> DB.source_file
-
 val split_classpath : string -> string list
 
 (** map entry for source files with potential basname collision within the same compiler call *)

infer/src/java/jMain.ml

@@ -81,7 +81,7 @@ let capture_libs linereader program tenv =
     | Javalib.JClass _ ->
         begin
           let fake_source_file =
-            JClasspath.java_source_file_from_path (JFrontend.path_of_cached_classname cn) in
+            DB.source_file_from_abs_path (JFrontend.path_of_cached_classname cn) in
           init_global_state fake_source_file;
           let call_graph, cfg =
             JFrontend.compute_class_icfg fake_source_file linereader program tenv node in