erl_tar: Fix handling of date and time

Since aa0c4b0df7cdc, erl_tar would write the local time (instead of
the POSIX time) into the tar header for the archived files. When
extracting the tar file, the extracted file could be set to a future
time (depending on the time zone).

We could do a minimal fix, but this seems to be a good time
to rewrite the time handling to use the new features that
allow file info to be read and written in the POSIX time
format.

First reported here: https://github.com/erlang/rebar3/issues/1554

3 files changed, 64 insertions, 32 deletions

diff --git a/lib/stdlib/src/erl_tar.erl b/lib/stdlib/src/erl_tar.erl
index 168ea4002c..76f0b38108 100644
--- a/lib/stdlib/src/erl_tar.erl
+++ b/lib/stdlib/src/erl_tar.erl
@@ -176,7 +176,7 @@ check_extract(Name, #read_opts{files=Files}) ->
 -type tar_entry() :: {filename(),
                       typeflag(),
                       non_neg_integer(),
-                      calendar:datetime(),
+                      tar_time(),
                       mode(),
                       uid(),
                       gid()}.
@@ -274,8 +274,13 @@ mode_to_string(Mode, [_|T], Acc) ->
 mode_to_string(_, [], Acc) ->
     Acc.
 
-%% Converts a datetime tuple to a readable string
-time_to_string({{Y, Mon, Day}, {H, Min, _}}) ->
+%% Converts a tar_time() (POSIX time) to a readable string
+time_to_string(Secs0) ->
+    Epoch = calendar:datetime_to_gregorian_seconds(?EPOCH),
+    Secs = Epoch + Secs0,
+    DateTime0 = calendar:gregorian_seconds_to_datetime(Secs),
+    DateTime = calendar:universal_time_to_local_time(DateTime0),
+    {{Y, Mon, Day}, {H, Min, _}} = DateTime,
     io_lib:format("~s ~2w ~s:~s ~w", [month(Mon), Day, two_d(H), two_d(Min), Y]).
 
 two_d(N) ->
@@ -452,7 +457,8 @@ add(Reader, NameOrBin, NameInArchive, Options)
 
 do_add(#reader{access=write}=Reader, Name, NameInArchive, Options)
   when is_list(NameInArchive), is_list(Options) ->
-    Opts = #add_opts{read_info=fun(F) -> file:read_link_info(F) end},
+    RF = fun(F) -> file:read_link_info(F, [{time, posix}]) end,
+    Opts = #add_opts{read_info=RF},
     add1(Reader, Name, NameInArchive, add_opts(Options, Opts));
 do_add(#reader{access=read},_,_,_) ->
     {error, eacces};
@@ -460,7 +466,8 @@ do_add(Reader,_,_,_) ->
     {error, {badarg, Reader}}.
 
 add_opts([dereference|T], Opts) ->
-    add_opts(T, Opts#add_opts{read_info=fun(F) -> file:read_file_info(F) end});
+    RF = fun(F) -> file:read_file_info(F, [{time, posix}]) end,
+    add_opts(T, Opts#add_opts{read_info=RF});
 add_opts([verbose|T], Opts) ->
     add_opts(T, Opts#add_opts{verbose=true});
 add_opts([{chunks,N}|T], Opts) ->
@@ -503,7 +510,7 @@ add1(#reader{}=Reader, Name, NameInArchive, #add_opts{read_info=ReadInfo}=Opts)
     end;
 add1(Reader, Bin, NameInArchive, Opts) when is_binary(Bin) ->
     add_verbose(Opts, "a ~ts~n", [NameInArchive]),
-    Now = calendar:now_to_local_time(erlang:timestamp()),
+    Now = os:system_time(seconds),
     Header = #tar_header{
                 name = NameInArchive,
                 size = byte_size(Bin),
@@ -612,7 +619,7 @@ build_header(#tar_header{}=Header, Opts) ->
        devmajor=Devmaj,
        devminor=Devmin
       } = Header,
-    Mtime = datetime_to_posix(Header#tar_header.mtime),
+    Mtime = Header#tar_header.mtime,
 
     Block0 = ?ZERO_BLOCK,
     {Block1, Pax0} = write_string(Block0, ?V7_NAME, ?V7_NAME_LEN, Name, ?PAX_PATH, #{}),
@@ -770,14 +777,6 @@ join_split_ustar_path([Part|Rest], {ok, Name, nil}) ->
 join_split_ustar_path([Part|Rest], {ok, Name, Acc}) ->
     join_split_ustar_path(Rest, {ok, Name, <<Acc/binary,$/,Part/binary>>}).
 
-datetime_to_posix(DateTime) ->
-    Epoch = calendar:datetime_to_gregorian_seconds(?EPOCH),
-    Secs = calendar:datetime_to_gregorian_seconds(DateTime),
-    case Secs - Epoch of
-        N when N < 0 -> 0;
-        N -> N
-    end.
-
 write_octal(Block, Pos, Size, X) ->
     Octal = zero_pad(format_octal(X), Size-1),
     if byte_size(Octal) < Size ->
@@ -984,7 +983,7 @@ do_get_format(#header_v7{}=V7, Bin)
 
 unpack_format(Format, #header_v7{}=V7, Bin, Reader)
   when is_binary(Bin), byte_size(Bin) =:= ?BLOCK_SIZE ->
-    Mtime = posix_to_erlang_time(parse_numeric(V7#header_v7.mtime)),
+    Mtime = parse_numeric(V7#header_v7.mtime),
     Header0 = #tar_header{
                  name=parse_string(V7#header_v7.name),
                  mode=parse_numeric(V7#header_v7.mode),
@@ -1051,9 +1050,9 @@ unpack_modern(Format, #header_v7{}=V7, Bin, #tar_header{}=Header0)
                             Star = to_star(V7, Bin),
                             Prefix0 = parse_string(Star#header_star.prefix),
                             Atime0 = Star#header_star.atime,
-                            Atime = posix_to_erlang_time(parse_numeric(Atime0)),
+                            Atime = parse_numeric(Atime0),
                             Ctime0 = Star#header_star.ctime,
-                            Ctime = posix_to_erlang_time(parse_numeric(Ctime0)),
+                            Ctime = parse_numeric(Ctime0),
                             {Prefix0, H1#tar_header{
                                         atime=Atime,
                                         ctime=Ctime
@@ -1313,11 +1312,6 @@ is_header_only_type(?TYPE_LINK)    -> true;
 is_header_only_type(?TYPE_DIR)     -> true;
 is_header_only_type(_) -> false.
 
-posix_to_erlang_time(Sec) ->
-    OneMillion = 1000000,
-    Time = calendar:now_to_datetime({Sec div OneMillion, Sec rem OneMillion, 0}),
-    erlang:universaltime_to_localtime(Time).
-
 foldl_read(#reader{access=read}=Reader, Fun, Accu, #read_opts{}=Opts)
   when is_function(Fun,4) ->
     case foldl_read0(Reader, Fun, Accu, Opts) of
@@ -1423,7 +1417,7 @@ do_merge_pax(Header, [_Ignore|Rest]) ->
     do_merge_pax(Header, Rest).
 
 %% Returns the time since UNIX epoch as a datetime
--spec parse_pax_time(binary()) -> calendar:datetime().
+-spec parse_pax_time(binary()) -> tar_time().
 parse_pax_time(Bin) when is_binary(Bin) ->
     TotalNano = case binary:split(Bin, [<<$.>>]) of
                     [SecondsStr, NanoStr0] ->
@@ -1450,8 +1444,7 @@ parse_pax_time(Bin) when is_binary(Bin) ->
     Micro = TotalNano div 1000,
     Mega = Micro div 1000000000000,
     Secs = Micro div 1000000 - (Mega*1000000),
-    Micro2 = Micro rem 1000000,
-    calendar:now_to_datetime({Mega, Secs, Micro2}).
+    Secs.
 
 %% Given a regular file reader, reads the whole file and
 %% parses all extended attributes it contains.
@@ -1671,7 +1664,7 @@ set_extracted_file_info(Name, #tar_header{typeflag = ?TYPE_BLOCK}=Header) ->
     set_device_info(Name, Header);
 set_extracted_file_info(Name, #tar_header{mtime=Mtime,mode=Mode}) ->
     Info = #file_info{mode=Mode, mtime=Mtime},
-    file:write_file_info(Name, Info).
+    file:write_file_info(Name, Info, [{time, posix}]).
 
 set_device_info(Name, #tar_header{}=Header) ->
     Mtime = Header#tar_header.mtime,
diff --git a/lib/stdlib/src/erl_tar.hrl b/lib/stdlib/src/erl_tar.hrl
index d646d02989..cff0c2f500 100644
--- a/lib/stdlib/src/erl_tar.hrl
+++ b/lib/stdlib/src/erl_tar.hrl
@@ -55,6 +55,8 @@
                      {string(), binary()} |
                      {string(), file:filename()}].
 
+-type tar_time() :: non_neg_integer().
+
 %% The tar header, once fully parsed.
 -record(tar_header, {
           name = "" :: string(),                %% name of header file entry
@@ -62,15 +64,15 @@
           uid = 0 :: non_neg_integer(),         %% user id of owner
           gid = 0 :: non_neg_integer(),         %% group id of owner
           size = 0 :: non_neg_integer(),        %% length in bytes
-          mtime :: calendar:datetime(),         %% modified time
+          mtime :: tar_time(),                  %% modified time
           typeflag :: char(),                   %% type of header entry
           linkname = "" :: string(),            %% target name of link
           uname = "" :: string(),               %% user name of owner
           gname = "" :: string(),               %% group name of owner
           devmajor = 0 :: non_neg_integer(),    %% major number of character or block device
           devminor = 0 :: non_neg_integer(),    %% minor number of character or block device
-          atime :: calendar:datetime(),         %% access time
-          ctime :: calendar:datetime()          %% status change time
+          atime :: tar_time(),                  %% access time
+          ctime :: tar_time()                   %% status change time
          }).
 -type tar_header() :: #tar_header{}.
 
diff --git a/lib/stdlib/test/tar_SUITE.erl b/lib/stdlib/test/tar_SUITE.erl
index e9ab12e061..4061008812 100644
--- a/lib/stdlib/test/tar_SUITE.erl
+++ b/lib/stdlib/test/tar_SUITE.erl
@@ -27,7 +27,8 @@
 	 extract_from_binary_compressed/1, extract_filtered/1,
 	 extract_from_open_file/1, symlinks/1, open_add_close/1, cooked_compressed/1,
 	 memory/1,unicode/1,read_other_implementations/1,
-         sparse/1, init/1, leading_slash/1, dotdot/1]).
+         sparse/1, init/1, leading_slash/1, dotdot/1,
+         roundtrip_metadata/1]).
 
 -include_lib("common_test/include/ct.hrl").
 -include_lib("kernel/include/file.hrl").
@@ -41,7 +42,7 @@ all() ->
      extract_filtered,
      symlinks, open_add_close, cooked_compressed, memory, unicode,
      read_other_implementations,
-     sparse,init,leading_slash,dotdot].
+     sparse,init,leading_slash,dotdot,roundtrip_metadata].
 
 groups() -> 
     [].
@@ -953,6 +954,42 @@ dotdot(Config) ->
 
     ok.
 
+roundtrip_metadata(Config) ->
+    PrivDir = proplists:get_value(priv_dir, Config),
+    Dir = filename:join(PrivDir, ?FUNCTION_NAME),
+    ok = file:make_dir(Dir),
+
+    do_roundtrip_metadata(Dir, "name-does-not-matter"),
+    ok.
+
+do_roundtrip_metadata(Dir, File) ->
+    Tar = filename:join(Dir, atom_to_list(?FUNCTION_NAME)++".tar"),
+    BeamFile = code:which(compile),
+    {ok,Fd} = erl_tar:open(Tar, [write]),
+    ok = erl_tar:add(Fd, BeamFile, File, []),
+    ok = erl_tar:close(Fd),
+
+    ok = erl_tar:extract(Tar, [{cwd,Dir}]),
+
+    %% Make sure that size and modification times are the same
+    %% on all platforms.
+    {ok,OrigInfo} = file:read_file_info(BeamFile),
+    ExtractedFile = filename:join(Dir, File),
+    {ok,ExtractedInfo} = file:read_file_info(ExtractedFile),
+    #file_info{size=Size,mtime=Mtime,type=regular} = OrigInfo,
+    #file_info{size=Size,mtime=Mtime,type=regular} = ExtractedInfo,
+
+    %% On Unix platforms more fields are expected to be the same.
+    case os:type() of
+        {unix,_} ->
+            #file_info{access=Access,mode=Mode} = OrigInfo,
+            #file_info{access=Access,mode=Mode} = ExtractedInfo,
+            ok;
+        _ ->
+            ok
+    end.
+
+
 %% Delete the given list of files.
 delete_files([]) -> ok;
 delete_files([Item|Rest]) ->