[DB] Remove source_file_from_string, source_file_to_string no longer returns real path

Summary:
source_file_[to|from]_string were dangerous. While removing source_file_to_string is hard/impossible, source file should never be created from string.
Also include many random changes related to `source_file`:
- improve comments in DB.mli
- define behavior of changed-files-index and improve its description
- move some of the "dangerous" code inline to discourage its reuse

This mostly concludes cleanup of DB.source_file, the last bit is to unify filtering by filename (we have duplicated logic in `InferConfig`, `CLocation` and `JMain`)

Reviewed By: jvillard

Differential Revision: D4258795

fbshipit-source-id: 36735a8

infer/src/backend/InferPrint.re

@@ -477,18 +477,6 @@ let module IssuesJson = {
   let is_first_item = ref true;
   let pp_json_open fmt () => F.fprintf fmt "[@?";
   let pp_json_close fmt () => F.fprintf fmt "]\n@?";
-  let expand_links_under_buck_out file =>
-    if (Utils.string_is_prefix Config.buck_generated_folder file) {
-      try {
-        let file = Unix.readlink file;
-        let source_file = DB.source_file_from_string file;
-        DB.source_file_to_rel_path source_file
-      } {
-      | Unix.Unix_error _ => file
-      }
-    } else {
-      file
-    };
 
   /** Write bug report in JSON format */
   let pp_issues_of_error_log fmt error_filter _ proc_loc_opt procname err_log => {
@@ -522,7 +510,7 @@ let module IssuesJson = {
         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 source_file);
+        let file = 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/backend/inferconfig.ml

@@ -262,7 +262,7 @@ let test () =
   directory_iter
     (fun path ->
        if DB.is_source_file path then
-         let source_file = (DB.source_file_from_string path) in
+         let source_file = DB.source_file_from_abs_path path in
          let matching = matching_analyzers source_file in
          if matching <> [] then
            let matching_s = join_strings ", " (IList.map fst matching) in

infer/src/base/Config.ml

@@ -650,7 +650,8 @@ and calls_csv =
 
 and changed_files_index =
   CLOpt.mk_path_opt ~long:"changed-files-index" ~exes:CLOpt.[Toplevel] ~meta:"file"
-    "Specify the file containing the list of files from which reactive analysis should start"
+    "Specify the file containing the list of source files from which reactive analysis should \
+     start. Source files should be specified relative to project root or be absolute"
 
 and check_duplicate_symbols =
   CLOpt.mk_bool ~long:"check-duplicate-symbols"

infer/src/base/DB.ml

@@ -40,12 +40,6 @@ module SourceFileMap = Map.Make(OrderedSourceFile)
 
 module SourceFileSet = Set.Make(OrderedSourceFile)
 
-let source_file_from_string path =
-  if Filename.is_relative path then
-    RelativeProjectRoot path
-  else
-    Absolute path
-
 let rel_path_from_abs_path root fname =
   let relative_complemented_fname = filename_to_relative root fname in
   if string_is_prefix root fname &&
@@ -53,7 +47,6 @@ let rel_path_from_abs_path root 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 =
   (* IMPORTANT: results of realpath are cached to not ruin performance *)
   let fname_real = realpath fname in
@@ -75,8 +68,8 @@ let curr_encoding = `Enc_crc
 
 let source_file_to_string fname =
   match fname with
+  | RelativeInferModel path -> "INFER_MODEL/" ^ path
   | RelativeProjectRoot path
-  | RelativeInferModel path
   | Absolute path -> path
 
 let source_file_pp fmt fname =
@@ -96,7 +89,7 @@ let source_file_line_count source_file =
 let source_file_to_rel_path fname =
   match fname with
   | RelativeProjectRoot path -> path
-  | _ -> source_file_to_abs_path fname |> filename_to_relative Config.project_root
+  | _ -> source_file_to_abs_path fname
 
 (** string encoding of a source file (including path) as a single filename *)
 let source_file_encoding source_file =
@@ -155,6 +148,20 @@ let source_file_of_header header_file =
     | _ -> None in
   Option.map source_file_from_abs_path file_opt
 
+let changed_source_files_set =
+  let create_source_file path =
+    if Filename.is_relative path then
+      (* sources in changed-files-index may be specified relative to project root *)
+      RelativeProjectRoot path
+    else
+      source_file_from_abs_path path in
+  Option.map_default read_file None Config.changed_files_index |>
+  Option.map (
+    IList.fold_left
+      (fun changed_files line -> SourceFileSet.add (create_source_file line) changed_files)
+      SourceFileSet.empty
+  )
+
 (** {2 Source Dirs} *)
 
 (** source directory: the directory inside the results dir corresponding to a source file *)
@@ -405,11 +412,3 @@ let fold_paths_matching ~dir ~p ~init ~f =
     matcher function p *)
 let paths_matching dir p =
   fold_paths_matching ~dir ~p ~init:[] ~f:(fun x xs -> x :: xs)
-
-let changed_source_files_set =
-  Option.map_default read_file None Config.changed_files_index |>
-  Option.map (
-    IList.fold_left
-      (fun changed_files line -> SourceFileSet.add (source_file_from_string line) changed_files)
-      SourceFileSet.empty
-  )

infer/src/base/DB.mli

@@ -94,21 +94,21 @@ val source_file_line_count : source_file -> int
 (** empty source file *)
 val source_file_empty : source_file
 
+(** create source file from absolute path *)
 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
 
-(** convert a source file to a string *)
+(** convert a source file to a string
+    WARNING: result may not be valid file path, do not use this function to perform operations
+             on filenames *)
 val source_file_to_string : source_file -> string
 
-(** convert a string obtained by source_file_to_string to a source file *)
-val source_file_from_string : string -> source_file
-
 (** pretty print source_file *)
 val source_file_pp : Format.formatter -> source_file -> unit
 
-(** get the full path of a source file, raise No_project_root exception when used with a relative source file and no project root specified *)
+(** get the full path of a source file *)
 val source_file_to_abs_path : source_file -> string
 
 (** get the relative path of a source file *)
@@ -126,6 +126,9 @@ val source_file_is_under_project_root : source_file -> bool
     file exists. returns None otherwise *)
 val source_file_of_header : source_file -> source_file option
 
+(** Set of files read from --changed-files-index file, None if option not specified *)
+val changed_source_files_set :  SourceFileSet.t option
+
 (** {2 Source Dirs} *)
 
 (** source directory: the directory inside the results dir corresponding to a source file *)
@@ -170,6 +173,3 @@ val fold_paths_matching :
 
 (** Return all file paths recursively under the given directory which match the given predicate *)
 val paths_matching : string -> (string -> bool) -> string list
-
-(** Set of files read from --changed-files-index file, None if option not specified *)
-val changed_source_files_set :  SourceFileSet.t option

infer/src/clang/cFrontend_utils.ml

@@ -625,12 +625,6 @@ struct
     let class_name = Ast_utils.get_class_name_from_member field_qual_name in
     Ident.create_fieldname (Mangled.mangled field_name class_name) 0
 
-  let get_rel_file_path file_opt =
-    match file_opt with
-    | Some file ->
-        DB.source_file_to_string (DB.source_file_from_abs_path file)
-    | None -> ""
-
   let is_cpp_translation translation_unit_context =
     let lang = translation_unit_context.CFrontend_config.lang in
     lang = CFrontend_config.CPP || lang = CFrontend_config.ObjCPP
@@ -660,6 +654,11 @@ struct
          | _ -> assert false)
 
   let mk_procname_from_function translation_unit_context name function_decl_info_opt =
+    let get_rel_file_path file_opt =
+      match file_opt with
+      | Some file ->
+          DB.source_file_to_string (DB.source_file_from_abs_path file)
+      | None -> "" in
     let file =
       match function_decl_info_opt with
       | Some (decl_info, function_decl_info) ->

infer/src/harness/inhabit.ml

@@ -202,22 +202,18 @@ let inhabit_call tenv (procname, receiver) cfg env =
   with Not_found -> env
 
 (** create a dummy file for the harness and associate them in the exe_env *)
-let create_dummy_harness_file harness_name =
-  let dummy_file_name =
-    let dummy_file_dir =
-      Filename.get_temp_dir_name () in
-    let file_str =
-      Procname.java_get_class_name
-        harness_name ^ "_" ^Procname.java_get_method harness_name ^ ".java" in
-    Filename.concat dummy_file_dir file_str in
-  DB.source_file_from_string dummy_file_name
+let create_dummy_harness_filename harness_name =
+  let dummy_file_dir =
+    Filename.get_temp_dir_name () in
+  let file_str =
+    Procname.java_get_class_name
+      harness_name ^ "_" ^Procname.java_get_method harness_name ^ ".java" in
+  Filename.concat dummy_file_dir file_str
 
 (** write the SIL for the harness to a file *)
 (* TODO (t3040429): fill this file up with Java-like code that matches the SIL *)
-let write_harness_to_file harness_instrs harness_file =
-  let harness_file =
-    let harness_file_name = DB.source_file_to_abs_path harness_file in
-    create_outfile harness_file_name in
+let write_harness_to_file harness_instrs harness_file_name =
+  let harness_file = create_outfile harness_file_name in
   let pp_harness fmt = IList.iter (fun instr ->
       Format.fprintf fmt "%a\n" (Sil.pp_instr pe_text) instr) harness_instrs in
   do_outf harness_file (fun outf ->
@@ -266,7 +262,7 @@ let setup_harness_cfg harness_name env cg cfg =
 let inhabit_trace tenv trace harness_name cg cfg =
   if IList.length trace > 0 then
     let source_file = Cg.get_source cg in
-    let harness_file = create_dummy_harness_file harness_name in
+    let harness_filename = create_dummy_harness_filename harness_name in
     let start_line = 1 in
     let empty_env =
       let pc = { Location.line = start_line; col = 1; file = source_file; } in
@@ -279,5 +275,5 @@ let inhabit_trace tenv trace harness_name cg cfg =
     let env'' = IList.fold_left (fun env to_call -> inhabit_call tenv to_call cfg env) empty_env trace in
     try
       setup_harness_cfg harness_name env'' cg cfg;
-      write_harness_to_file (IList.rev env''.instrs) harness_file
+      write_harness_to_file (IList.rev env''.instrs) harness_filename
     with Not_found -> ()