aboutsummaryrefslogtreecommitdiffstats
path: root/lib/stdlib/src/unicode.erl
diff options
context:
space:
mode:
authorLukas Larsson <[email protected]>2011-05-18 16:21:34 +0200
committerLukas Larsson <[email protected]>2011-05-18 16:21:34 +0200
commit15426ac367eed736c165a5bdbb1c051a87944f68 (patch)
treefcabce7847168a8416600fe35f94a411a5f73d6e /lib/stdlib/src/unicode.erl
parent4cd0717b717803ce8f03a12de4bf89f452ed1df7 (diff)
parentf44bbb331fb517e989d4d906b7f63ec110bbbc18 (diff)
downloadotp-15426ac367eed736c165a5bdbb1c051a87944f68.tar.gz
otp-15426ac367eed736c165a5bdbb1c051a87944f68.tar.bz2
otp-15426ac367eed736c165a5bdbb1c051a87944f68.zip
Merge branch 'dev' of super:otp into dev
* 'dev' of super:otp: (166 commits) Corrected documentation error and added examples to Users Guide In TLS 1.1, failure to properly close a connection no longer requires that a session not be resumed. This is a change from TLS 1.0 to conform with widespread implementation practice. Erlang ssl will now in TLS 1.0 conform to the widespread implementation practice instead of the specification to avoid performance issues. Add escript to bootstrap/bin Remove unused variable warning in inet_res Remove unused variable in epmd_port Remove compiler warnings in inet_drv Add SASL test suite Allow same module name in multiple applications if explicitely excluded Fix bugs concerning the option report_missing_types Fix default encoding in SAX parser. re: remove gratuitous "it " in manpage Spelling in (backward *compatibility*) comment. Improve erl_docgen's support for Dialyzer specs and types dialyzer warning on mnesia_tm Add documentation text about majority checking add mnesia_majority_test suite where_to_wlock optimization + change_table_majority/2 bug in mnesia_tm:needs_majority/2 optimize sticky_lock maj. check check majority for sticky locks ...
Diffstat (limited to 'lib/stdlib/src/unicode.erl')
-rw-r--r--lib/stdlib/src/unicode.erl51
1 files changed, 50 insertions, 1 deletions
diff --git a/lib/stdlib/src/unicode.erl b/lib/stdlib/src/unicode.erl
index 12bc60623d..a5d9965ca2 100644
--- a/lib/stdlib/src/unicode.erl
+++ b/lib/stdlib/src/unicode.erl
@@ -30,12 +30,34 @@
characters_to_binary/3,
bom_to_encoding/1, encoding_to_bom/1]).
--export_type([encoding/0]).
+-export_type([chardata/0, charlist/0, encoding/0, external_chardata/0,
+ external_charlist/0, latin1_chardata/0,
+ latin1_charlist/0, unicode_binary/0, unicode_char/0]).
-type encoding() :: 'latin1' | 'unicode' | 'utf8'
| 'utf16' | {'utf16', endian()}
| 'utf32' | {'utf32', endian()}.
-type endian() :: 'big' | 'little'.
+-type unicode_binary() :: binary().
+-type unicode_char() :: non_neg_integer().
+-type charlist() :: [unicode_char() | unicode_binary() | charlist()].
+-type chardata() :: charlist() | unicode_binary().
+-type external_unicode_binary() :: binary().
+-type external_chardata() :: external_charlist() | external_unicode_binary().
+-type external_charlist() :: [unicode_char() | external_unicode_binary()
+ | external_charlist()].
+-type latin1_binary() :: binary().
+-type latin1_char() :: byte().
+-type latin1_chardata() :: latin1_charlist() | latin1_binary().
+-type latin1_charlist() :: [latin1_char() | latin1_binary()
+ | latin1_charlist()].
+
+-spec characters_to_list(Data) -> Result when
+ Data :: latin1_chardata() | chardata() | external_chardata(),
+ Result :: list()
+ | {error, list(), RestData}
+ | {incomplete, list(), binary()},
+ RestData :: latin1_chardata() | chardata() | external_chardata().
characters_to_list(ML) ->
unicode:characters_to_list(ML,unicode).
@@ -69,6 +91,13 @@ do_characters_to_list(ML, Encoding) ->
end.
+-spec characters_to_binary(Data) -> Result when
+ Data :: latin1_chardata() | chardata() | external_chardata(),
+ Result :: binary()
+ | {error, binary(), RestData}
+ | {incomplete, binary(), binary()},
+ RestData :: latin1_chardata() | chardata() | external_chardata().
+
characters_to_binary(ML) ->
try
unicode:characters_to_binary(ML,unicode)
@@ -104,6 +133,15 @@ characters_to_binary_int(ML,InEncoding) ->
erlang:raise(error,TheError,[{Mod,characters_to_binary,L}|Rest])
end.
+-spec characters_to_binary(Data, InEncoding, OutEncoding) -> Result when
+ Data :: latin1_chardata() | chardata() | external_chardata(),
+ InEncoding :: encoding(),
+ OutEncoding :: encoding(),
+ Result :: binary()
+ | {error, binary(), RestData}
+ | {incomplete, binary(), binary()},
+ RestData :: latin1_chardata() | chardata() | external_chardata().
+
characters_to_binary(ML, latin1, latin1) when is_binary(ML) ->
ML;
characters_to_binary(ML, latin1, Uni) when is_binary(ML) and ((Uni =:= utf8) or (Uni =:= unicode)) ->
@@ -215,6 +253,13 @@ characters_to_binary_int(ML, InEncoding, OutEncoding) ->
Res
end.
+-spec bom_to_encoding(Bin) -> {Encoding, Length} when
+ Bin :: binary(),
+ Encoding :: 'latin1' | 'utf8'
+ | {'utf16', endian()}
+ | {'utf32', endian()},
+ Length :: non_neg_integer().
+
bom_to_encoding(<<239,187,191,_/binary>>) ->
{utf8,3};
bom_to_encoding(<<0,0,254,255,_/binary>>) ->
@@ -228,6 +273,10 @@ bom_to_encoding(<<255,254,_/binary>>) ->
bom_to_encoding(Bin) when is_binary(Bin) ->
{latin1,0}.
+-spec encoding_to_bom(InEncoding) -> Bin when
+ Bin :: binary(),
+ InEncoding :: encoding().
+
encoding_to_bom(unicode) ->
<<239,187,191>>;
encoding_to_bom(utf8) ->