diff options
Diffstat (limited to 'Master/texmf-dist/source/latex/l3experimental/l3str/l3str-convert.dtx')
| -rw-r--r-- | Master/texmf-dist/source/latex/l3experimental/l3str/l3str-convert.dtx | 325 |
1 files changed, 133 insertions, 192 deletions
diff --git a/Master/texmf-dist/source/latex/l3experimental/l3str/l3str-convert.dtx b/Master/texmf-dist/source/latex/l3experimental/l3str/l3str-convert.dtx index 66a01368167..6fa42db903e 100644 --- a/Master/texmf-dist/source/latex/l3experimental/l3str/l3str-convert.dtx +++ b/Master/texmf-dist/source/latex/l3experimental/l3str/l3str-convert.dtx @@ -27,7 +27,6 @@ \documentclass[full]{l3doc} \usepackage{amsmath} \begin{document} - \tableofcontents \DocInput{\jobname.dtx} \end{document} %</driver> @@ -47,7 +46,7 @@ % }^^A % } % -% \date{Released 2018/03/05} +% \date{Released 2018-04-30} % % \maketitle % @@ -151,7 +150,7 @@ % in Tables~\ref{tab:encodings} and~\ref{tab:escapings}, respectively. % The default escaping is to input and output bytes directly. The % special case of an empty \meta{name} indicates the use of -% \enquote{native} strings, 8-bit for pdf\TeX{}, and Unicode strings +% \enquote{native} strings, 8-bit for \pdfTeX{}, and Unicode strings % for the other two engines. % % For example, @@ -190,23 +189,6 @@ % encoding. Instead, the \meta{false code} is performed. % \end{function} % -% \section{Internal string functions} -% -% \begin{function}{\__str_hexadecimal_use:NTF} -% \begin{syntax} -% \cs{__str_hexadecimal_use:NTF} \meta{token} \Arg{true code} \Arg{false code} -% \end{syntax} -% If the \meta{token} is a hexadecimal digit (upper case or lower -% case), its upper-case version is left in the input stream, -% \emph{followed} by the \meta{true code}. Otherwise, the \meta{false -% code} is left in the input stream. -% \begin{texnote} -% This function fails on some inputs if the escape character is a -% hexadecimal digit. We are thus careful to set the escape -% character to a known (safe) value before using it. -% \end{texnote} -% \end{function} -% % \section{Possibilities, and things to do} % % Encoding/escaping-related tasks. @@ -252,7 +234,7 @@ % \end{macrocode} % % \begin{macrocode} -\ProvidesExplPackage{l3str-convert}{2018/03/05}{} +\ProvidesExplPackage{l3str-convert}{2018-04-30}{} {L3 Experimental string encoding conversions} % \end{macrocode} % @@ -371,14 +353,14 @@ \prg_new_conditional:Npnn \@@_if_contains_char:NN #1#2 { T , TF } { \exp_after:wN \@@_if_contains_char_aux:NN \exp_after:wN #2 - #1 { \__prg_break:n { ? \fi: } } - \__prg_break_point: + #1 { \prg_break:n { ? \fi: } } + \prg_break_point: \prg_return_false: } \prg_new_conditional:Npnn \@@_if_contains_char:nN #1#2 { TF } { - \@@_if_contains_char_aux:NN #2 #1 { \__prg_break:n { ? \fi: } } - \__prg_break_point: + \@@_if_contains_char_aux:NN #2 #1 { \prg_break:n { ? \fi: } } + \prg_break_point: \prg_return_false: } \cs_new:Npn \@@_if_contains_char_aux:NN #1#2 @@ -389,7 +371,7 @@ \@@_if_contains_char_aux:NN #1 } \cs_new:Npn \@@_if_contains_char_true: - { \__prg_break:n { \prg_return_true: \use_none:n } } + { \prg_break:n { \prg_return_true: \use_none:n } } % \end{macrocode} % \end{macro} % \end{macro} @@ -434,9 +416,7 @@ \if_int_compare:w 1 < "1 \token_to_str:N #1 \exp_stop_f: #1 \prg_return_true: \else: - \if_case:w \__int_eval:w - \exp_after:wN ` \token_to_str:N #1 - `a - \__int_eval_end: + \if_case:w \int_eval:n { \exp_after:wN ` \token_to_str:N #1 - `a } A \or: B \or: C @@ -485,7 +465,6 @@ % \begin{macro}[EXP]{\@@_output_byte:n} % \begin{macro}[EXP]{\@@_output_byte:w} % \begin{macro}[EXP]{\@@_output_hexadecimal:n} -% \begin{macro}[EXP]{\@@_output_hexadecimal:w} % \begin{macro}[EXP]{\@@_output_end:} % Those functions must be used carefully: feeding them a value outside % the range $[-1,255]$ will attempt to use the undefined token list @@ -501,24 +480,21 @@ { \exp_after:wN \exp_after:wN \exp_after:wN \use_i:nnn - \cs:w c_@@_byte_ \__int_value:w \__int_eval:w + \cs:w c_@@_byte_ \int_eval:w } \cs_new:Npn \@@_output_hexadecimal:n #1 - { \@@_output_hexadecimal:w #1 \@@_output_end: } -\cs_new:Npn \@@_output_hexadecimal:w { \exp_after:wN \exp_after:wN \exp_after:wN \use_none:n - \cs:w c_@@_byte_ \__int_value:w \__int_eval:w + \cs:w c_@@_byte_ \int_eval:n {#1} _tl \cs_end: } \cs_new:Npn \@@_output_end: - { \__int_eval_end: _tl \cs_end: } + { \scan_stop: _tl \cs_end: } % \end{macrocode} % \end{macro} % \end{macro} % \end{macro} % \end{macro} -% \end{macro} % % \begin{macro}[rEXP]{\@@_output_byte_pair_be:n} % \begin{macro}[rEXP]{\@@_output_byte_pair_le:n} @@ -561,8 +537,8 @@ { \exp_after:wN \@@_convert_gmap_loop:NN \exp_after:wN #1 - \g_@@_result_tl { ? \__prg_break: } - \__prg_break_point: + \g_@@_result_tl { ? \prg_break: } + \prg_break_point: } } \cs_new:Npn \@@_convert_gmap_loop:NN #1#2 @@ -586,8 +562,8 @@ { \exp_after:wN \@@_convert_gmap_internal_loop:Nww \exp_after:wN #1 - \g_@@_result_tl \s__tl \q_stop \__prg_break: \s__tl - \__prg_break_point: + \g_@@_result_tl \s__tl \q_stop \prg_break: \s__tl + \prg_break_point: } } \cs_new:Npn \@@_convert_gmap_internal_loop:Nww #1 #2 \s__tl #3 \s__tl @@ -706,7 +682,7 @@ { \group_begin: #1 - \tl_gset:Nx \g_@@_result_tl { \@@_to_other_fast:n {#4} } + \tl_gset:Nx \g_@@_result_tl { \__kernel_str_to_other_fast:n {#4} } \exp_after:wN \@@_convert:wwwnn \tl_to_str:n {#5} /// \q_stop { decode } { unescape } @@ -856,8 +832,8 @@ \cs_new:Npn \@@_convert_lowercase_alphanum:n #1 { \exp_after:wN \@@_convert_lowercase_alphanum_loop:N - \tl_to_str:n {#1} { ? \__prg_break: } - \__prg_break_point: + \tl_to_str:n {#1} { ? \prg_break: } + \prg_break_point: } \cs_new:Npn \@@_convert_lowercase_alphanum_loop:N #1 { @@ -932,8 +908,8 @@ \cs_new:Npn \@@_filter_bytes:n #1 { \@@_filter_bytes_aux:N #1 - { ? \__prg_break: } - \__prg_break_point: + { ? \prg_break: } + \prg_break_point: } \cs_new:Npn \@@_filter_bytes_aux:N #1 { @@ -996,101 +972,59 @@ \cs_new_protected:Npn \@@_convert_decode_: { \@@_convert_gmap:N \@@_decode_native_char:N } \cs_new:Npn \@@_decode_native_char:N #1 - { #1 \s__tl \__int_value:w `#1 \s__tl } + { #1 \s__tl \int_value:w `#1 \s__tl } % \end{macrocode} % \end{macro} % \end{macro} % % \begin{macro}{\@@_convert_encode_:} -% The conversion from an internal string to native character tokens is -% very different in pdf\TeX{} and in other engines. For Unicode-aware -% engines, we need the definitions to be read when the null byte has -% category code $12$, so we set that inside a group. -% \begin{macrocode} -\group_begin: - \char_set_catcode_other:n { 0 } - \bool_lazy_any:nTF - { - \sys_if_engine_luatex_p: - \sys_if_engine_xetex_p: - } -% \end{macrocode} -% \begin{macro}{\@@_encode_native_loop:w} -% \begin{macro}{\@@_encode_native_flush:} -% \begin{macro}[rEXP]{\@@_encode_native_filter:N} -% In Unicode-aware engines, since building particular characters -% cannot be done expandably in \TeX{}, we cannot hope to get a -% linear-time function. However, we get quite close using the -% \pkg{l3tl-build} module, which abuses \tn{toks} to reach an almost -% linear time. We produce an arbitrary character of category code 12 -% and add that character -% to the end of the token list being built. At the end of the loop, -% put the token list together with \cs{__tl_build_end:}. Note that we -% use an \texttt{x}-expanding assignment because it is slightly -% faster. Unicode-aware engines will never incur an overflow because -% the internal string is guaranteed to only contain code points in -% $[0,\hexnum{10FFFF}]$. -% \begin{macrocode} - { - \cs_new_protected:Npn \@@_convert_encode_: - { - \__tl_gbuild_x:Nw \g_@@_result_tl - \exp_after:wN \@@_encode_native_loop:w - \g_@@_result_tl \s__tl { \q_stop \__prg_break: } \s__tl - \__prg_break_point: - \__tl_build_end: - } - \cs_new_protected:Npn \@@_encode_native_loop:w #1 \s__tl #2 \s__tl - { - \use_none_delimit_by_q_stop:w #2 \q_stop - \exp_args:Nx \__tl_build_one:n - { \char_generate:nn {#2} {12} } - \@@_encode_native_loop:w - } - } -% \end{macrocode} % \begin{macro}[EXP]{\@@_encode_native_char:n} -% Since pdf\TeX{} only supports 8-bit characters, and we have a table -% of all bytes, the conversion can be done in linear time within an -% \texttt{x}-expanding assignment. Look out for character codes larger -% than $255$, those characters are replaced by |?|, and raise a flag, -% which then triggers a pdf\TeX{}-specific error. +% The conversion from an internal string to native character tokens +% basically maps \cs{char_generate:nn} through the code-points, but in +% non-Unicode-aware engines we use a fall-back character |?| rather +% than nothing when given a character code outside $[0,255]$. We +% detect the presence of bad characters using a flag and only produce +% a single error after the \texttt{x}-expanding assignment. % \begin{macrocode} - { - \cs_new_protected:Npn \@@_convert_encode_: - { - \flag_clear:n { str_error } - \@@_convert_gmap_internal:N \@@_encode_native_char:n - \@@_if_flag_error:nnx { str_error } - { pdfTeX-native-overflow } { } - } - \cs_new:Npn \@@_encode_native_char:n #1 - { - \if_int_compare:w #1 > \c_@@_max_byte_int - \flag_raise:n { str_error } - ? - \else: - \@@_output_byte:n {#1} - \fi: - } - \__kernel_msg_new:nnnn { str } { pdfTeX-native-overflow } - { Character~code~too~large~for~pdfTeX. } - { - The~pdfTeX~engine~only~supports~8-bit~characters:~ - valid~character~codes~are~in~the~range~[0,255].~ - To~manipulate~arbitrary~Unicode,~use~LuaTeX~or~XeTeX. - } - } +\bool_lazy_any:nTF + { + \sys_if_engine_luatex_p: + \sys_if_engine_xetex_p: + } + { + \cs_new_protected:Npn \@@_convert_encode_: + { \@@_convert_gmap_internal:N \@@_encode_native_char:n } + \cs_new:Npn \@@_encode_native_char:n #1 + { \char_generate:nn {#1} {12} } + } + { + \cs_new_protected:Npn \@@_convert_encode_: + { + \flag_clear:n { str_error } + \@@_convert_gmap_internal:N \@@_encode_native_char:n + \@@_if_flag_error:nnx { str_error } + { native-overflow } { } + } + \cs_new:Npn \@@_encode_native_char:n #1 + { + \if_int_compare:w #1 > \c_@@_max_byte_int + \flag_raise:n { str_error } + ? + \else: + \char_generate:nn {#1} {12} + \fi: + } + \__kernel_msg_new:nnnn { str } { native-overflow } + { Character~code~too~large~for~this~engine. } + { + This~engine~only~support~8-bit~characters:~ + valid~character~codes~are~in~the~range~[0,255].~ + To~manipulate~arbitrary~Unicode,~use~LuaTeX~or~XeTeX. + } + } % \end{macrocode} % \end{macro} % \end{macro} -% \end{macro} -% \end{macro} -% End the group to restore the catcode of the null byte. -% \begin{macrocode} -\group_end: -% \end{macrocode} -% \end{macro} % % \subsubsection{\texttt{clist}} % @@ -1195,12 +1129,12 @@ \int_zero:N \l_@@_internal_int \exp_last_unbraced:Nx \@@_decode_eight_bit_load:nn { \tl_use:c { c_@@_encoding_#1_tl } } - { \q_stop \__prg_break: } { } - \__prg_break_point: + { \q_stop \prg_break: } { } + \prg_break_point: \exp_last_unbraced:Nx \@@_decode_eight_bit_load_missing:n { \tl_use:c { c_@@_encoding_#1_missing_tl } } - { \q_stop \__prg_break: } - \__prg_break_point: + { \q_stop \prg_break: } + \prg_break_point: \flag_clear:n { str_error } \@@_convert_gmap:N \@@_decode_eight_bit_char:N \@@_if_flag_error:nnx { str_error } { decode-8-bit } {#1} @@ -1211,7 +1145,7 @@ \use_none_delimit_by_q_stop:w #1 \q_stop \tex_dimen:D "#1 = \l_@@_internal_int sp \scan_stop: \tex_skip:D \l_@@_internal_int = "#1 sp \scan_stop: - \tex_toks:D \l_@@_internal_int \exp_after:wN { \__int_value:w "#2 } + \tex_toks:D \l_@@_internal_int \exp_after:wN { \int_value:w "#2 } \int_incr:N \l_@@_internal_int \@@_decode_eight_bit_load:nn } @@ -1233,7 +1167,7 @@ \tex_the:D \tex_toks:D \tex_dimen:D \fi: \fi: - \__int_value:w `#1 \s__tl + \int_value:w `#1 \s__tl } % \end{macrocode} % \end{macro} @@ -1253,8 +1187,8 @@ \int_zero:N \l_@@_internal_int \exp_last_unbraced:Nx \@@_encode_eight_bit_load:nn { \tl_use:c { c_@@_encoding_#1_tl } } - { \q_stop \__prg_break: } { } - \__prg_break_point: + { \q_stop \prg_break: } { } + \prg_break_point: \flag_clear:n { str_error } \@@_convert_gmap_internal:N \@@_encode_eight_bit_char:n \@@_if_flag_error:nnx { str_error } { encode-8-bit } {#1} @@ -1374,11 +1308,11 @@ % \begin{itemize} % \item \texttt{bytes} (default), non-bytes are filtered out, and % bytes are left untouched (this is defined by default); -% \item \texttt{hex} or \texttt{hexadecimal}, as per the pdf\TeX{} +% \item \texttt{hex} or \texttt{hexadecimal}, as per the \pdfTeX{} % primitive \tn{pdfescapehex} -% \item \texttt{name}, as per the pdf\TeX{} primitive +% \item \texttt{name}, as per the \pdfTeX{} primitive % \tn{pdfescapename} -% \item \texttt{string}, as per the pdf\TeX{} primitive +% \item \texttt{string}, as per the \pdfTeX{} primitive % \tn{pdfescapestring} % \item \texttt{url}, as per the percent encoding of urls. % \end{itemize} @@ -1405,8 +1339,8 @@ \@@_output_byte:w " \exp_last_unbraced:Nf \@@_unescape_hex_auxi:N { \tl_to_str:N \g_@@_result_tl } - 0 { ? 0 - 1 \__prg_break: } - \__prg_break_point: + 0 { ? 0 - 1 \prg_break: } + \prg_break_point: \@@_output_end: } \@@_if_flag_error:nnx { str_error } { unescape-hex } { } @@ -1484,8 +1418,8 @@ \tl_gset:Nx \g_@@_result_tl { \exp_after:wN #3 \g_@@_result_tl - #1 ? { ? \__prg_break: } - \__prg_break_point: + #1 ? { ? \prg_break: } + \prg_break_point: } \@@_if_flag_error:nnx { str_byte } { non-byte } { #2 } \@@_if_flag_error:nnx { str_error } { unescape-#2 } { } @@ -1576,14 +1510,14 @@ \tl_gset:Nx \g_@@_result_tl { \exp_after:wN \@@_unescape_string_newlines:wN - \g_@@_result_tl \__prg_break: ^^M ? - \__prg_break_point: + \g_@@_result_tl \prg_break: ^^M ? + \prg_break_point: } \tl_gset:Nx \g_@@_result_tl { \exp_after:wN \@@_unescape_string_loop:wNNN - \g_@@_result_tl #1 ?? { ? \__prg_break: } - \__prg_break_point: + \g_@@_result_tl #1 ?? { ? \prg_break: } + \prg_break_point: } \@@_if_flag_error:nnx { str_byte } { non-byte } { string } \@@_if_flag_error:nnx { str_error } { unescape-string } { } @@ -1877,7 +1811,7 @@ \exp_after:wN \use_none_delimit_by_q_stop:w \fi: \exp_after:wN \@@_encode_utf_viii_loop:wwnnw - \__int_value:w \int_div_truncate:nn {#1} {64} ; #1 ; + \int_value:w \int_div_truncate:nn {#1} {64} ; #1 ; #5 \q_stop \@@_output_byte:n { #2 - 64 * ( #1 - 2 ) } } @@ -1922,11 +1856,14 @@ \flag_clear_new:n { str_overflow } \__kernel_msg_new:nnnn { str } { utf8-decode } { - Invalid~UTF-8~string: \exp_last_unbraced:Nf \use_none:n - \@@_if_flag_times:nT { str_missing } { ,~missing~continuation~byte } - \@@_if_flag_times:nT { str_extra } { ,~extra~continuation~byte } - \@@_if_flag_times:nT { str_overlong } { ,~overlong~form } - \@@_if_flag_times:nT { str_overflow } { ,~code~point~too~large } + Invalid~UTF-8~string: + \exp_last_unbraced:Nf \use_none:n + { + \@@_if_flag_times:nT { str_missing } { ,~missing~continuation~byte } + \@@_if_flag_times:nT { str_extra } { ,~extra~continuation~byte } + \@@_if_flag_times:nT { str_overlong } { ,~overlong~form } + \@@_if_flag_times:nT { str_overflow } { ,~code~point~too~large } + } . } { @@ -2026,7 +1963,7 @@ % need to check whether we reached the end of the string. In a correct % \textsc{utf-8} string, this happens automatically when the % \texttt{_start} auxiliary leaves its first argument in the input -% stream: the end-marker begins with \cs{__prg_break:}, which ends +% stream: the end-marker begins with \cs{prg_break:}, which ends % the loop. On the other hand, if the end is reached when looking for % a continuation byte, the \cs{use_none:n} |#3| construction removes % the first token from the end-marker, and leaves the \texttt{_end} @@ -2043,8 +1980,8 @@ \tl_gset:Nx \g_@@_result_tl { \exp_after:wN \@@_decode_utf_viii_start:N \g_@@_result_tl - { \__prg_break: \@@_decode_utf_viii_end: } - \__prg_break_point: + { \prg_break: \@@_decode_utf_viii_end: } + \prg_break_point: } \@@_if_flag_error:nnx { str_error } { utf8-decode } { } } @@ -2054,7 +1991,7 @@ \if_int_compare:w `#1 < "C0 \exp_stop_f: \s__tl \if_int_compare:w `#1 < "80 \exp_stop_f: - \__int_value:w `#1 + \int_value:w `#1 \else: \flag_raise:n { str_extra } \flag_raise:n { str_error } @@ -2062,7 +1999,7 @@ \fi: \else: \exp_after:wN \@@_decode_utf_viii_continuation:wwN - \__int_value:w \__int_eval:w `#1 - "C0 \exp_after:wN \__int_eval_end: + \int_value:w \int_eval:n { `#1 - "C0 } \exp_after:wN \fi: \s__tl \use_none_delimit_by_q_stop:w {"80} {"800} {"10000} {"110000} \q_stop @@ -2077,9 +2014,7 @@ "C0 \exp_stop_f: #3 \exp_after:wN \@@_decode_utf_viii_aux:wNnnwN - \__int_value:w \__int_eval:w - #1 * "40 + `#3 - "80 - \exp_after:wN \__int_eval_end: + \int_value:w \int_eval:n { #1 * "40 + `#3 - "80 } \exp_after:wN \else: \s__tl \flag_raise:n { str_missing } @@ -2107,7 +2042,7 @@ \@@_decode_utf_viii_overflow:w #1 \fi: \exp_after:wN \@@_decode_utf_viii_continuation:wwN - \__int_value:w \__int_eval:w #1 - #4 \exp_after:wN \__int_eval_end: + \int_value:w \int_eval:n { #1 - #4 } \exp_after:wN \fi: \s__tl #2 {#4} #5 @@ -2126,7 +2061,7 @@ \flag_raise:n { str_missing } \flag_raise:n { str_error } \int_use:N \c_@@_replacement_char_int \s__tl - \__prg_break: + \prg_break: } % \end{macrocode} % \end{macro} @@ -2246,10 +2181,13 @@ } \__kernel_msg_new:nnnn { str } { utf16-decode } { - Invalid~UTF-16~string: \exp_last_unbraced:Nf \use_none:n - \@@_if_flag_times:nT { str_missing } { ,~missing~trail~surrogate } - \@@_if_flag_times:nT { str_extra } { ,~extra~trail~surrogate } - \@@_if_flag_times:nT { str_end } { ,~odd~number~of~bytes } + Invalid~UTF-16~string: + \exp_last_unbraced:Nf \use_none:n + { + \@@_if_flag_times:nT { str_missing } { ,~missing~trail~surrogate } + \@@_if_flag_times:nT { str_extra } { ,~extra~trail~surrogate } + \@@_if_flag_times:nT { str_end } { ,~odd~number~of~bytes } + } . } { @@ -2299,7 +2237,7 @@ % big-endianness. The three endianness cases are based on a common % auxiliary whose first argument is $1$ for big-endian and $2$ for % little-endian, and whose second argument, delimited by the scan mark -% \cs{s__stop}, is expanded once (the string may be long; passing +% \cs{s_stop}, is expanded once (the string may be long; passing % \cs{g_@@_result_tl} as an argument before expansion is cheaper). % % The \cs{@@_decode_utf_xvi:Nw} function defines \cs{@@_tmp:w} to @@ -2309,13 +2247,13 @@ % \cs{@@_decode_utf_xvi_pair:NN} described below. % \begin{macrocode} \cs_new_protected:cpn { @@_convert_decode_utf16be: } - { \@@_decode_utf_xvi:Nw 1 \g_@@_result_tl \s__stop } + { \@@_decode_utf_xvi:Nw 1 \g_@@_result_tl \s_stop } \cs_new_protected:cpn { @@_convert_decode_utf16le: } - { \@@_decode_utf_xvi:Nw 2 \g_@@_result_tl \s__stop } + { \@@_decode_utf_xvi:Nw 2 \g_@@_result_tl \s_stop } \cs_new_protected:cpn { @@_convert_decode_utf16: } { \exp_after:wN \@@_decode_utf_xvi_bom:NN - \g_@@_result_tl \s__stop \s__stop \s__stop + \g_@@_result_tl \s_stop \s_stop \s_stop } \cs_new_protected:Npn \@@_decode_utf_xvi_bom:NN #1#2 { @@ -2327,7 +2265,7 @@ { \@@_decode_utf_xvi:Nw 1 #1#2 } } } - \cs_new_protected:Npn \@@_decode_utf_xvi:Nw #1#2 \s__stop + \cs_new_protected:Npn \@@_decode_utf_xvi:Nw #1#2 \s_stop { \flag_clear:n { str_error } \flag_clear:n { str_missing } @@ -2338,7 +2276,7 @@ { \exp_after:wN \@@_decode_utf_xvi_pair:NN #2 \q_nil \q_nil - \__prg_break_point: + \prg_break_point: } \@@_if_flag_error:nnx { str_error } { utf16-decode } { } } @@ -2398,7 +2336,7 @@ \@@_decode_utf_xvi_pair_end:Nw #1 \fi: \if_case:w - \__int_eval:w ( \@@_tmp:w #1#2 - "D6 ) / 4 \__int_eval_end: + \int_eval:n { ( \@@_tmp:w #1#2 - "D6 ) / 4 } \scan_stop: \or: \exp_after:wN \@@_decode_utf_xvi_quad:NNwNN \or: \exp_after:wN \@@_decode_utf_xvi_extra:NNw \fi: @@ -2440,7 +2378,7 @@ \else: \@@_decode_utf_xvi_error:nNN { end } #1 \prg_do_nothing: \fi: - \__prg_break: + \prg_break: } \cs_new:Npn \@@_decode_utf_xvi_extra:NNw #1#2 \s__tl #3 \s__tl { \@@_decode_utf_xvi_error:nNN { extra } #1#2 } @@ -2538,9 +2476,12 @@ \flag_clear_new:n { str_end } \__kernel_msg_new:nnnn { str } { utf32-decode } { - Invalid~UTF-32~string: \exp_last_unbraced:Nf \use_none:n - \@@_if_flag_times:nT { str_overflow } { ,~code~point~too~large } - \@@_if_flag_times:nT { str_end } { ,~truncated~string } + Invalid~UTF-32~string: + \exp_last_unbraced:Nf \use_none:n + { + \@@_if_flag_times:nT { str_overflow } { ,~code~point~too~large } + \@@_if_flag_times:nT { str_end } { ,~truncated~string } + } . } { @@ -2576,7 +2517,7 @@ % % The structure is similar to \textsc{utf-16} decoding functions. If % the endianness is not given, test the first $4$ bytes of the string -% (possibly \cs{s__stop} if the string is too short) for the presence +% (possibly \cs{s_stop} if the string is too short) for the presence % of a byte-order mark. If there is a byte-order mark, use that % endianness, and remove the $4$ bytes, otherwise default to % big-endian, and leave the $4$ bytes in place. The @@ -2588,22 +2529,22 @@ % inside an \texttt{x}-expanding assignment to \cs{g_@@_result_tl}. % % The \texttt{_loop} auxiliary first checks for the end-of-string -% marker \cs{s__stop}, calling the \texttt{_end} auxiliary if +% marker \cs{s_stop}, calling the \texttt{_end} auxiliary if % appropriate. Otherwise, leave the \meta{4~bytes} \cs{s__tl} behind, % then check that the code point is not overflowing: the leading byte % must be $0$, and the following byte at most $16$. % % In the ending code, we check that there remains no byte: there -% should be nothing left until the first \cs{s__stop}. Break the map. +% should be nothing left until the first \cs{s_stop}. Break the map. % \begin{macrocode} \cs_new_protected:cpn { @@_convert_decode_utf32be: } - { \@@_decode_utf_xxxii:Nw 1 \g_@@_result_tl \s__stop } + { \@@_decode_utf_xxxii:Nw 1 \g_@@_result_tl \s_stop } \cs_new_protected:cpn { @@_convert_decode_utf32le: } - { \@@_decode_utf_xxxii:Nw 2 \g_@@_result_tl \s__stop } + { \@@_decode_utf_xxxii:Nw 2 \g_@@_result_tl \s_stop } \cs_new_protected:cpn { @@_convert_decode_utf32: } { \exp_after:wN \@@_decode_utf_xxxii_bom:NNNN \g_@@_result_tl - \s__stop \s__stop \s__stop \s__stop \s__stop + \s_stop \s_stop \s_stop \s_stop \s_stop } \cs_new_protected:Npn \@@_decode_utf_xxxii_bom:NNNN #1#2#3#4 { @@ -2615,7 +2556,7 @@ { \@@_decode_utf_xxxii:Nw 1 #1#2#3#4 } } } - \cs_new_protected:Npn \@@_decode_utf_xxxii:Nw #1#2 \s__stop + \cs_new_protected:Npn \@@_decode_utf_xxxii:Nw #1#2 \s_stop { \flag_clear:n { str_overflow } \flag_clear:n { str_end } @@ -2624,14 +2565,14 @@ \tl_gset:Nx \g_@@_result_tl { \exp_after:wN \@@_decode_utf_xxxii_loop:NNNN - #2 \s__stop \s__stop \s__stop \s__stop - \__prg_break_point: + #2 \s_stop \s_stop \s_stop \s_stop + \prg_break_point: } \@@_if_flag_error:nnx { str_error } { utf32-decode } { } } \cs_new:Npn \@@_decode_utf_xxxii_loop:NNNN #1#2#3#4 { - \if_meaning:w \s__stop #4 + \if_meaning:w \s_stop #4 \exp_after:wN \@@_decode_utf_xxxii_end:w \fi: #1#2#3#4 \s__tl @@ -2652,7 +2593,7 @@ \s__tl \@@_decode_utf_xxxii_loop:NNNN } - \cs_new:Npn \@@_decode_utf_xxxii_end:w #1 \s__stop + \cs_new:Npn \@@_decode_utf_xxxii_end:w #1 \s_stop { \tl_if_empty:nF {#1} { @@ -2661,7 +2602,7 @@ #1 \s__tl \int_use:N \c_@@_replacement_char_int \s__tl } - \__prg_break: + \prg_break: } % \end{macrocode} % \end{macro} |