From 8e7441c27485bd416abdb0253c9c92620d671065 Mon Sep 17 00:00:00 2001 From: Richard Carlsson Date: Tue, 6 May 2014 10:34:07 +0200 Subject: Make the scanned form of the io_lib format strings available for processing This adds three new functions to io_lib - scan_format/2, unscan_format/1, and build_text/1 - which expose the parsed form of the format control sequences to make it possible to easily modify or filter the input to io_lib:format/2. This can e.g. be used in order to replace unbounded-size control sequences like ~w or ~p with corresponding depth-limited ~W and ~P before doing the actual formatting. --- lib/stdlib/src/io_lib_format.erl | 112 +++++++++++++++++++++++++++++++++------ 1 file changed, 95 insertions(+), 17 deletions(-) (limited to 'lib/stdlib/src/io_lib_format.erl') diff --git a/lib/stdlib/src/io_lib_format.erl b/lib/stdlib/src/io_lib_format.erl index 89ae6fb187..015afb317a 100644 --- a/lib/stdlib/src/io_lib_format.erl +++ b/lib/stdlib/src/io_lib_format.erl @@ -1,7 +1,7 @@ %% %% %CopyrightBegin% %% -%% Copyright Ericsson AB 1996-2013. All Rights Reserved. +%% Copyright Ericsson AB 1996-2014. All Rights Reserved. %% %% The contents of this file are subject to the Erlang Public License, %% Version 1.1, (the "License"); you may not use this file except in @@ -20,10 +20,9 @@ %% Formatting functions of io library. --export([fwrite/2,fwrite_g/1,indentation/2]). +-export([fwrite/2,fwrite_g/1,indentation/2,scan/2,unscan/1,build/1]). -%% fwrite(Format, ArgList) -> string(). -%% Format the arguments in ArgList after string Format. Just generate +%% Format the arguments in Args after string Format. Just generate %% an error if there is an error in the arguments. %% %% To do the printing command correctly we need to calculate the @@ -37,15 +36,84 @@ %% and it also splits the handling of the control characters into two %% parts. -fwrite(Format, Args) when is_atom(Format) -> - fwrite(atom_to_list(Format), Args); -fwrite(Format, Args) when is_binary(Format) -> - fwrite(binary_to_list(Format), Args); +-spec fwrite(Format, Data) -> FormatList when + Format :: io:format(), + Data :: [term()], + FormatList :: [char() | io_lib:format_spec()]. + fwrite(Format, Args) -> - Cs = collect(Format, Args), + build(scan(Format, Args)). + +%% Build the output text for a pre-parsed format list. + +-spec build(FormatList) -> io_lib:chars() when + FormatList :: [char() | io_lib:format_spec()]. + +build(Cs) -> Pc = pcount(Cs), build(Cs, Pc, 0). +%% Parse all control sequences in the format string. + +-spec scan(Format, Data) -> FormatList when + Format :: io:format(), + Data :: [term()], + FormatList :: [char() | io_lib:format_spec()]. + +scan(Format, Args) when is_atom(Format) -> + scan(atom_to_list(Format), Args); +scan(Format, Args) when is_binary(Format) -> + scan(binary_to_list(Format), Args); +scan(Format, Args) -> + collect(Format, Args). + +%% Revert a pre-parsed format list to a plain character list and a +%% list of arguments. + +-spec unscan(FormatList) -> {Format, Data} when + FormatList :: [char() | io_lib:format_spec()], + Format :: io:format(), + Data :: [term()]. + +unscan(Cs) -> + {print(Cs), args(Cs)}. + +args([#{args := As} | Cs]) -> + As ++ args(Cs); +args([_C | Cs]) -> + args(Cs); +args([]) -> + []. + +print([#{control_char := C, width := F, adjust := Ad, precision := P, + pad_char := Pad, encoding := Encoding, strings := Strings} | Cs]) -> + print(C, F, Ad, P, Pad, Encoding, Strings) ++ print(Cs); +print([C | Cs]) -> + [C | print(Cs)]; +print([]) -> + []. + +print(C, F, Ad, P, Pad, Encoding, Strings) -> + [$~] ++ print_field_width(F, Ad) ++ print_precision(P) ++ + print_pad_char(Pad) ++ print_encoding(Encoding) ++ + print_strings(Strings) ++ [C]. + +print_field_width(none, _Ad) -> ""; +print_field_width(F, left) -> integer_to_list(-F); +print_field_width(F, right) -> integer_to_list(F). + +print_precision(none) -> ""; +print_precision(P) -> [$. | integer_to_list(P)]. + +print_pad_char($\s) -> ""; % default, no need to make explicit +print_pad_char(Pad) -> [$., Pad]. + +print_encoding(unicode) -> "t"; +print_encoding(latin1) -> "". + +print_strings(false) -> "l"; +print_strings(true) -> "". + collect([$~|Fmt0], Args0) -> {C,Fmt1,Args1} = collect_cseq(Fmt0, Args0), [C|collect(Fmt1, Args1)]; @@ -60,7 +128,10 @@ collect_cseq(Fmt0, Args0) -> {Encoding,Fmt4,Args4} = encoding(Fmt3, Args3), {Strings,Fmt5,Args5} = strings(Fmt4, Args4), {C,As,Fmt6,Args6} = collect_cc(Fmt5, Args5), - {{C,As,F,Ad,P,Pad,Encoding,Strings},Fmt6,Args6}. + FormatSpec = #{control_char => C, args => As, width => F, adjust => Ad, + precision => P, pad_char => Pad, encoding => Encoding, + strings => Strings}, + {FormatSpec,Fmt6,Args6}. encoding([$t|Fmt],Args) -> true = hd(Fmt) =/= $l, @@ -136,17 +207,19 @@ collect_cc([$i|Fmt], [A|Args]) -> {$i,[A],Fmt,Args}. pcount(Cs) -> pcount(Cs, 0). -pcount([{$p,_As,_F,_Ad,_P,_Pad,_Enc,_Str}|Cs], Acc) -> pcount(Cs, Acc+1); -pcount([{$P,_As,_F,_Ad,_P,_Pad,_Enc,_Str}|Cs], Acc) -> pcount(Cs, Acc+1); +pcount([#{control_char := $p}|Cs], Acc) -> pcount(Cs, Acc+1); +pcount([#{control_char := $P}|Cs], Acc) -> pcount(Cs, Acc+1); pcount([_|Cs], Acc) -> pcount(Cs, Acc); pcount([], Acc) -> Acc. -%% build([Control], Pc, Indentation) -> string(). +%% build([Control], Pc, Indentation) -> io_lib:chars(). %% Interpret the control structures. Count the number of print %% remaining and only calculate indentation when necessary. Must also %% be smart when calculating indentation for characters in format. -build([{C,As,F,Ad,P,Pad,Enc,Str}|Cs], Pc0, I) -> +build([#{control_char := C, args := As, width := F, adjust := Ad, + precision := P, pad_char := Pad, encoding := Enc, + strings := Str} | Cs], Pc0, I) -> S = control(C, As, F, Ad, P, Pad, Enc, Str, I), Pc1 = decr_pc(C, Pc0), if @@ -162,10 +235,14 @@ decr_pc($p, Pc) -> Pc - 1; decr_pc($P, Pc) -> Pc - 1; decr_pc(_, Pc) -> Pc. -%% indentation(String, Indentation) -> Indentation. + %% Calculate the indentation of the end of a string given its start %% indentation. We assume tabs at 8 cols. +-spec indentation(String, StartIndent) -> integer() when + String :: io_lib:chars(), + StartIndent :: integer(). + indentation([$\n|Cs], _I) -> indentation(Cs, 0); indentation([$\t|Cs], I) -> indentation(Cs, ((I + 8) div 8) * 8); indentation([C|Cs], I) when is_integer(C) -> @@ -366,7 +443,6 @@ float_data([D|Cs], Ds) when D >= $0, D =< $9 -> float_data([_|Cs], Ds) -> float_data(Cs, Ds). -%% fwrite_g(Float) %% Writes the shortest, correctly rounded string that converts %% to Float when read back with list_to_float/1. %% @@ -374,6 +450,8 @@ float_data([_|Cs], Ds) -> %% in Proceedings of the SIGPLAN '96 Conference on Programming %% Language Design and Implementation. +-spec fwrite_g(float()) -> string(). + fwrite_g(0.0) -> "0.0"; fwrite_g(Float) when is_float(Float) -> @@ -642,7 +720,7 @@ prefixed_integer(Int, F, Adj, Base, Pad, Prefix, Lowercase) term([Prefix|S], F, Adj, none, Pad) end. -%% char(Char, Field, Adjust, Precision, PadChar) -> string(). +%% char(Char, Field, Adjust, Precision, PadChar) -> chars(). char(C, none, _Adj, none, _Pad) -> [C]; char(C, F, _Adj, none, _Pad) -> chars(C, F); -- cgit v1.2.3