whitespace changes

This commit is contained in:
alisdair sullivan 2010-08-01 20:05:28 -07:00
parent 4d129ca320
commit 640156033e
2 changed files with 1 additions and 19 deletions

View file

@ -39,15 +39,6 @@ parse(JSON, Opts) ->
start(JSON, [], Opts). start(JSON, [], Opts).
%% this code is mostly autogenerated and mostly ugly. apologies. for more insight on
%% Callbacks or Opts, see the comments accompanying decoder/2 (in jsx.erl). Stack
%% is a stack of flags used to track depth and to keep track of whether we are
%% returning from a value or a key inside objects. all pops, peeks and pushes are
%% inlined. the code that handles naked values and comments is not optimized by the
%% compiler for efficient matching, but you shouldn't be using naked values or comments
%% anyways, they are horrible and contrary to the spec
start(<<S/?encoding, Rest/binary>>, Stack, Opts) when ?is_whitespace(S) -> start(<<S/?encoding, Rest/binary>>, Stack, Opts) when ?is_whitespace(S) ->
start(Rest, Stack, Opts); start(Rest, Stack, Opts);
start(<<?start_object/?encoding, Rest/binary>>, Stack, Opts) -> start(<<?start_object/?encoding, Rest/binary>>, Stack, Opts) ->
@ -254,10 +245,8 @@ key(Bin, Stack, Opts) ->
%% representation of the string being parsed. using a list of integers representing %% representation of the string being parsed. using a list of integers representing
%% unicode codepoints is faster than constructing binaries, many of which will be %% unicode codepoints is faster than constructing binaries, many of which will be
%% converted back to lists by the user anyways %% converted back to lists by the user anyways
%% string uses partial_utf/1 to cease parsing when invalid encodings are encountered %% string uses partial_utf/1 to cease parsing when invalid encodings are encountered
%% rather than just checking remaining binary size like other states %% rather than just checking remaining binary size like other states
string(<<?quote/?encoding, Rest/binary>>, [key|_] = Stack, Opts, Acc) -> string(<<?quote/?encoding, Rest/binary>>, [key|_] = Stack, Opts, Acc) ->
{event, {key, lists:reverse(Acc)}, fun() -> colon(Rest, Stack, Opts) end}; {event, {key, lists:reverse(Acc)}, fun() -> colon(Rest, Stack, Opts) end};
string(<<?quote/?encoding, Rest/binary>>, Stack, Opts, Acc) -> string(<<?quote/?encoding, Rest/binary>>, Stack, Opts, Acc) ->
@ -329,7 +318,6 @@ partial_utf(_) -> true.
%% only thing to note here is the additional accumulator passed to escaped_unicode used %% only thing to note here is the additional accumulator passed to escaped_unicode used
%% to hold the codepoint sequence. unescessary, but nicer than using the string %% to hold the codepoint sequence. unescessary, but nicer than using the string
%% accumulator %% accumulator
escape(<<$b/?encoding, Rest/binary>>, Stack, Opts, Acc) -> escape(<<$b/?encoding, Rest/binary>>, Stack, Opts, Acc) ->
string(Rest, Stack, Opts, "\b" ++ Acc); string(Rest, Stack, Opts, "\b" ++ Acc);
escape(<<$f/?encoding, Rest/binary>>, Stack, Opts, Acc) -> escape(<<$f/?encoding, Rest/binary>>, Stack, Opts, Acc) ->
@ -364,7 +352,6 @@ escape(Bin, Stack, Opts, Acc) ->
%% as it represents a valid unicode codepoint. this means non-characters %% as it represents a valid unicode codepoint. this means non-characters
%% representable in 16 bits are not converted (the utf16 surrogates and the two %% representable in 16 bits are not converted (the utf16 surrogates and the two
%% special non-characters). any other option and no conversion is done %% special non-characters). any other option and no conversion is done
escaped_unicode(<<D/?encoding, Rest/binary>>, escaped_unicode(<<D/?encoding, Rest/binary>>,
Stack, Stack,
?escaped_unicode_to_ascii(Opts), ?escaped_unicode_to_ascii(Opts),
@ -413,7 +400,6 @@ escaped_unicode(Bin, Stack, Opts, String, Acc) ->
%% upon encountering a low pair json/hex encoded value, check to see if there's a high %% upon encountering a low pair json/hex encoded value, check to see if there's a high
%% value already in the accumulator %% value already in the accumulator
check_acc_for_surrogate([D, C, B, A, $u, ?rsolidus|Rest]) check_acc_for_surrogate([D, C, B, A, $u, ?rsolidus|Rest])
when ?is_hex(D), ?is_hex(C), ?is_hex(B), ?is_hex(A) -> when ?is_hex(D), ?is_hex(C), ?is_hex(B), ?is_hex(A) ->
case erlang:list_to_integer([A, B, C, D], 16) of case erlang:list_to_integer([A, B, C, D], 16) of
@ -426,14 +412,12 @@ check_acc_for_surrogate(_) ->
false. false.
%% stole this from the unicode spec %% stole this from the unicode spec
surrogate_to_codepoint(High, Low) -> surrogate_to_codepoint(High, Low) ->
(High - 16#d800) * 16#400 + (Low - 16#dc00) + 16#10000. (High - 16#d800) * 16#400 + (Low - 16#dc00) + 16#10000.
%% like strings, numbers are collected in an intermediate accumulator before %% like strings, numbers are collected in an intermediate accumulator before
%% being emitted to the callback handler %% being emitted to the callback handler
negative(<<$0/?encoding, Rest/binary>>, Stack, Opts, Acc) -> negative(<<$0/?encoding, Rest/binary>>, Stack, Opts, Acc) ->
zero(Rest, Stack, Opts, "0" ++ Acc); zero(Rest, Stack, Opts, "0" ++ Acc);
negative(<<S/?encoding, Rest/binary>>, Stack, Opts, Acc) when ?is_nonzero(S) -> negative(<<S/?encoding, Rest/binary>>, Stack, Opts, Acc) when ?is_nonzero(S) ->
@ -804,8 +788,7 @@ null(Bin, Stack, Opts) ->
%% any unicode character is valid in a comment except the */ sequence which ends %% any unicode character is valid in a comment except the */ sequence which ends
%% the comment. they're implemented as a closure called when the comment ends that %% the comment. they're implemented as a closure called when the comment ends that
%% returns execution to the point where the comment began. comments are not %% returns execution to the point where the comment began. comments are not
%% recorded in any way, simply parsed. %% reported in any way, simply parsed.
maybe_comment(<<?star/?encoding, Rest/binary>>, Resume) -> maybe_comment(<<?star/?encoding, Rest/binary>>, Resume) ->
comment(Rest, Resume); comment(Rest, Resume);
maybe_comment(Bin, Resume) -> maybe_comment(Bin, Resume) ->

View file

@ -82,7 +82,6 @@ parse_opts([{encoding, _}|Rest], Opts) ->
%% encoding detection %% encoding detection
%% first check to see if there's a bom, if not, use the rfc4627 method for determining %% first check to see if there's a bom, if not, use the rfc4627 method for determining
%% encoding. this function makes some assumptions about the validity of the stream %% encoding. this function makes some assumptions about the validity of the stream
%% which may delay failure later than if an encoding is explicitly provided %% which may delay failure later than if an encoding is explicitly provided