dialyzer: Compact 'file' annotations in Core code

File annotations {file, File} in Core Erlang code is replaced by
{file, FakeFile}, where FakeFile is "0", "1", ..., in order to save
memory. When a warning message is created, FakeFile is translated to
the original File (a bit awkward but easy to do).

The Core code stored in ETS tables needs less space, but more
important is that a less heap memory is used in the dataflow,
warnings, and typesig phases, allowing more parallelism.

4 files changed, 69 insertions, 20 deletions

diff --git a/lib/dialyzer/src/dialyzer_analysis_callgraph.erl b/lib/dialyzer/src/dialyzer_analysis_callgraph.erl
index fed9a7d1f2..4e18058993 100644
--- a/lib/dialyzer/src/dialyzer_analysis_callgraph.erl
+++ b/lib/dialyzer/src/dialyzer_analysis_callgraph.erl
@@ -628,14 +628,14 @@ format_bad_calls([{{_, _, _}, {_, module_info, A}}|Left], CodeServer, Acc)
 format_bad_calls([{FromMFA, {M, F, A} = To}|Left], CodeServer, Acc) ->
   {_Var, FunCode} = dialyzer_codeserver:lookup_mfa_code(FromMFA, CodeServer),
   Msg = {call_to_missing, [M, F, A]},
-  {File, Line} = find_call_file_and_line(FunCode, To),
+  {File, Line} = find_call_file_and_line(FromMFA, FunCode, To, CodeServer),
   WarningInfo = {File, Line, FromMFA},
   NewAcc = [{?WARN_CALLGRAPH, WarningInfo, Msg}|Acc],
   format_bad_calls(Left, CodeServer, NewAcc);
 format_bad_calls([], _CodeServer, Acc) ->
   Acc.
 
-find_call_file_and_line(Tree, MFA) ->
+find_call_file_and_line({Module, _, _}, Tree, MFA, CodeServer) ->
   Fun =
     fun(SubTree, Acc) ->
 	case cerl:is_c_call(SubTree) of
@@ -648,7 +648,7 @@ find_call_file_and_line(Tree, MFA) ->
 		case {cerl:concrete(M), cerl:concrete(F), A} of
 		  MFA ->
 		    Ann = cerl:get_ann(SubTree),
-		    [{get_file(Ann), get_line(Ann)}|Acc];
+		    [{get_file(CodeServer, Module, Ann), get_line(Ann)}|Acc];
 		  {erlang, make_fun, 3} ->
 		    [CA1, CA2, CA3] = cerl:call_args(SubTree),
 		    case
@@ -664,7 +664,8 @@ find_call_file_and_line(Tree, MFA) ->
 			of
 			  MFA ->
 			    Ann = cerl:get_ann(SubTree),
-			    [{get_file(Ann), get_line(Ann)}|Acc];
+			    [{get_file(CodeServer, Module, Ann),
+                              get_line(Ann)}|Acc];
 			  _ ->
 			    Acc
 			end;
@@ -684,8 +685,10 @@ get_line([Line|_]) when is_integer(Line) -> Line;
 get_line([_|Tail]) -> get_line(Tail);
 get_line([]) -> -1.
 
-get_file([{file, File}|_]) -> File;
-get_file([_|Tail]) -> get_file(Tail).
+get_file(Codeserver, Module, [{file, FakeFile}|_]) ->
+  dialyzer_codeserver:translate_fake_file(Codeserver, Module, FakeFile);
+get_file(Codeserver, Module, [_|Tail]) ->
+  get_file(Codeserver, Module, Tail).
 
 -spec dump_callgraph(dialyzer_callgraph:callgraph(), #analysis_state{}, #analysis{}) ->
   'ok'.
diff --git a/lib/dialyzer/src/dialyzer_behaviours.erl b/lib/dialyzer/src/dialyzer_behaviours.erl
index 5623929a43..524ae047e2 100644
--- a/lib/dialyzer/src/dialyzer_behaviours.erl
+++ b/lib/dialyzer/src/dialyzer_behaviours.erl
@@ -62,9 +62,9 @@ check_callbacks(Module, Attrs, Records, Plt, Codeserver) ->
     _ ->
       MFA = {Module,module_info,0},
       {_Var,Code} = dialyzer_codeserver:lookup_mfa_code(MFA, Codeserver),
-      File = get_file(cerl:get_ann(Code)),
+      File = get_file(Codeserver, Module, cerl:get_ann(Code)),
       State = #state{plt = Plt, filename = File, behlines = BehLines,
-		     codeserver = Codeserver, records = Records},
+                     codeserver = Codeserver, records = Records},
       Warnings = get_warnings(Module, Behaviours, State),
       [add_tag_warning_info(Module, W, State) || W <- Warnings]
   end.
@@ -213,12 +213,15 @@ add_tag_warning_info(Module, {_Tag, [_B, Fun, Arity|_R]} = Warn, State) ->
     dialyzer_codeserver:lookup_mfa_code({Module, Fun, Arity},
 					State#state.codeserver),
   Anns = cerl:get_ann(FunCode),
-  WarningInfo = {get_file(Anns), get_line(Anns), {Module, Fun, Arity}},
+  File = get_file(State#state.codeserver, Module, Anns),
+  WarningInfo = {File, get_line(Anns), {Module, Fun, Arity}},
   {?WARN_BEHAVIOUR, WarningInfo, Warn}.
 
 get_line([Line|_]) when is_integer(Line) -> Line;
 get_line([_|Tail]) -> get_line(Tail);
 get_line([]) -> -1.
 
-get_file([{file, File}|_]) -> File;
-get_file([_|Tail]) -> get_file(Tail).
+get_file(Codeserver, Module, [{file, FakeFile}|_]) ->
+  dialyzer_codeserver:translate_fake_file(Codeserver, Module, FakeFile);
+get_file(Codeserver, Module, [_|Tail]) ->
+  get_file(Codeserver, Module, Tail).
diff --git a/lib/dialyzer/src/dialyzer_codeserver.erl b/lib/dialyzer/src/dialyzer_codeserver.erl
index 786ed229df..a5bb4e209c 100644
--- a/lib/dialyzer/src/dialyzer_codeserver.erl
+++ b/lib/dialyzer/src/dialyzer_codeserver.erl
@@ -60,7 +60,8 @@
 	 new/0,
 	 set_next_core_label/2,
 	 set_temp_records/2,
-	 store_temp_records/3]).
+	 store_temp_records/3,
+         translate_fake_file/3]).
 
 -export_type([codeserver/0, fun_meta_info/0, contracts/0]).
 
@@ -183,13 +184,15 @@ insert(Mod, ModCode, CS) ->
   Exports = cerl:module_exports(ModCode),
   Attrs = cerl:module_attrs(ModCode),
   Defs = cerl:module_defs(ModCode),
+  {Files, SmallDefs} = compress_file_anno(Defs),
   As = cerl:get_ann(ModCode),
   Funs =
     [{{Mod, cerl:fname_id(Var), cerl:fname_arity(Var)},
-      Val, {Var, cerl_trees:get_label(Fun)}} || Val = {Var, Fun} <- Defs],
+      Val, {Var, cerl_trees:get_label(Fun)}} || Val = {Var, Fun} <- SmallDefs],
   Keys = [Key || {Key, _Value, _Label} <- Funs],
   ModEntry = {Mod, {Name, Exports, Attrs, Keys, As}},
-  true = ets:insert(CS#codeserver.code, [ModEntry|Funs]),
+  ModFileEntry = {{mod, Mod}, Files},
+  true = ets:insert(CS#codeserver.code, [ModEntry, ModFileEntry|Funs]),
   CS.
 
 -spec get_temp_exported_types(codeserver()) -> sets:set(mfa()).
@@ -404,9 +407,39 @@ finalize_contracts(#codeserver{temp_contracts = TempContDict,
   true = ets:delete(TempCallDict),
   CS#codeserver{temp_contracts = clean, temp_callbacks = clean}.
 
+-spec translate_fake_file(codeserver(), module(), file:filename()) ->
+                             file:filename().
+
+translate_fake_file(#codeserver{code = Code}, Module, FakeFile) ->
+  Files = ets:lookup_element(Code, {mod, Module}, 2),
+  {FakeFile, File} = lists:keyfind(FakeFile, 1, Files),
+  File.
+
 table__lookup(TablePid, M) when is_atom(M) ->
   {Name, Exports, Attrs, Keys, As} = ets:lookup_element(TablePid, M, 2),
   Defs = [table__lookup(TablePid, Key) || Key <- Keys],
   cerl:ann_c_module(As, Name, Exports, Attrs, Defs);
 table__lookup(TablePid, MFA) ->
   ets:lookup_element(TablePid, MFA, 2).
+
+compress_file_anno(Term) ->
+  {Files, SmallTerm} = compress_file_anno(Term, []),
+  {[{FakeFile, File} || {File, {file, FakeFile}} <- Files], SmallTerm}.
+
+compress_file_anno({file, F}, Fs) when is_list(F) ->
+  case lists:keyfind(F, 1, Fs) of
+    false ->
+      I = integer_to_list(length(Fs)),
+      FileI = {file, I},
+      NFs = [{F, FileI}|Fs],
+      {NFs, FileI};
+    {F, FileI} -> {Fs, FileI}
+  end;
+compress_file_anno(T, Fs) when is_tuple(T) ->
+  {NFs, NL} = compress_file_anno(tuple_to_list(T), Fs),
+  {NFs, list_to_tuple(NL)};
+compress_file_anno([E|L], Fs) ->
+  {Fs1, NE} = compress_file_anno(E, Fs),
+  {NFs, NL} = compress_file_anno(L, Fs1),
+  {NFs, [NE|NL]};
+compress_file_anno(T, Fs) -> {Fs, T}.
diff --git a/lib/dialyzer/src/dialyzer_dataflow.erl b/lib/dialyzer/src/dialyzer_dataflow.erl
index 639ed426df..ce292e1140 100644
--- a/lib/dialyzer/src/dialyzer_dataflow.erl
+++ b/lib/dialyzer/src/dialyzer_dataflow.erl
@@ -529,7 +529,7 @@ handle_apply_or_call([{TypeOfApply, {Fun, Sig, Contr, LocalRet}}|Left],
     case is_race_analysis_enabled(State) of
       true ->
         Ann = cerl:get_ann(Tree),
-        File = get_file(Ann),
+        File = get_file(Ann, State),
         Line = abs(get_line(Ann)),
         dialyzer_races:store_race_call(Fun, ArgTypes, Args,
                                        {File, Line}, State);
@@ -3090,7 +3090,7 @@ state__add_warning(#state{warnings = Warnings, warning_mode = true} = State,
   Ann = cerl:get_ann(Tree),
   case Force of
     true ->
-      WarningInfo = {get_file(Ann),
+      WarningInfo = {get_file(Ann, State),
                      abs(get_line(Ann)),
                      State#state.curr_fun},
       Warn = {Tag, WarningInfo, Msg},
@@ -3100,7 +3100,9 @@ state__add_warning(#state{warnings = Warnings, warning_mode = true} = State,
       case is_compiler_generated(Ann) of
         true -> State;
         false ->
-          WarningInfo = {get_file(Ann), get_line(Ann), State#state.curr_fun},
+          WarningInfo = {get_file(Ann, State),
+                         get_line(Ann),
+                         State#state.curr_fun},
           Warn = {Tag, WarningInfo, Msg},
           case Tag of
             ?WARN_CONTRACT_RANGE -> ok;
@@ -3499,6 +3501,12 @@ state__put_races(Races, State) ->
 state__records_only(#state{records = Records}) ->
   #state{records = Records}.
 
+-spec state__translate_file(file:filename(), state()) -> file:filename().
+
+state__translate_file(FakeFile, State) ->
+  #state{codeserver = CodeServer, module = Module} = State,
+  dialyzer_codeserver:translate_fake_file(CodeServer, Module, FakeFile).
+
 %%% ===========================================================================
 %%%
 %%%  Races
@@ -3570,9 +3578,11 @@ get_line([Line|_]) when is_integer(Line) -> Line;
 get_line([_|Tail]) -> get_line(Tail);
 get_line([]) -> -1.
 
-get_file([]) -> [];
-get_file([{file, File}|_]) -> File;
-get_file([_|Tail]) -> get_file(Tail).
+get_file([], _State) -> [];
+get_file([{file, FakeFile}|_], State) ->
+  state__translate_file(FakeFile, State);
+get_file([_|Tail], State) ->
+  get_file(Tail, State).
 
 is_compiler_generated(Ann) ->
   lists:member(compiler_generated, Ann) orelse (get_line(Ann) < 1).