diff options
Diffstat (limited to 'Master/texmf-dist/source/latex-dev/l3kernel/l3text-case.dtx')
-rw-r--r-- | Master/texmf-dist/source/latex-dev/l3kernel/l3text-case.dtx | 2336 |
1 files changed, 2336 insertions, 0 deletions
diff --git a/Master/texmf-dist/source/latex-dev/l3kernel/l3text-case.dtx b/Master/texmf-dist/source/latex-dev/l3kernel/l3text-case.dtx new file mode 100644 index 00000000000..5010114fb4b --- /dev/null +++ b/Master/texmf-dist/source/latex-dev/l3kernel/l3text-case.dtx @@ -0,0 +1,2336 @@ +% \iffalse meta-comment +% +%% File: l3text-case.dtx +% +% Copyright (C) 2020-2024 The LaTeX Project +% +% It may be distributed and/or modified under the conditions of the +% LaTeX Project Public License (LPPL), either version 1.3c of this +% license or (at your option) any later version. The latest version +% of this license is in the file +% +% https://www.latex-project.org/lppl.txt +% +% This file is part of the "l3kernel bundle" (The Work in LPPL) +% and all files in that bundle must be distributed together. +% +% ----------------------------------------------------------------------- +% +% The development version of the bundle can be found at +% +% https://github.com/latex3/latex3 +% +% for those people who are interested. +% +%<*driver> +\documentclass[full,kernel]{l3doc} +\begin{document} + \DocInput{\jobname.dtx} +\end{document} +%</driver> +% \fi +% +% \title{^^A +% The \pkg{l3text-case} module\\ Text processing (case changing)^^A +% } +% +% \author{^^A +% The \LaTeX{} Project\thanks +% {^^A +% E-mail: +% \href{mailto:latex-team@latex-project.org} +% {latex-team@latex-project.org}^^A +% }^^A +% } +% +% \date{Released 2024-04-11} +% +% \maketitle +% +% \begin{documentation} +% +% \end{documentation} +% +% \begin{implementation} +% +% \section{\pkg{l3text-case} implementation} +% +% \begin{macrocode} +%<*package> +% \end{macrocode} +% +% \begin{macrocode} +%<@@=text> +% \end{macrocode} +% +% \subsection{Case changing} +% +% \begin{variable}{\l_text_titlecase_check_letter_bool} +% Needed to determine the route used in titlecasing. +% \begin{macrocode} +\bool_new:N \l_text_titlecase_check_letter_bool +\bool_set_true:N \l_text_titlecase_check_letter_bool +% \end{macrocode} +% \end{variable} +% +% \begin{macro}[EXP] +% { +% \text_lowercase:n, +% \text_uppercase:n, +% \text_titlecase_all:n, +% \text_titlecase_first:n +% } +% \begin{macro}[EXP] +% { +% \text_lowercase:nn, +% \text_uppercase:nn, +% \text_titlecase_all:nn, +% \text_titlecase_first:nn +% } +% \begin{macro}[EXP]{\@@_change_case:nnn} +% The user level functions here are all wrappers around the internal +% functions for case changing. +% \begin{macrocode} +\cs_new:Npn \text_lowercase:n #1 + { \@@_change_case:nnn { lower } { } {#1} } +\cs_new:Npn \text_uppercase:n #1 + { \@@_change_case:nnn { upper } { } {#1} } +\cs_new:Npn \text_titlecase_all:n #1 + { \@@_change_case:nnn { title } { } {#1} } +\cs_new:Npn \text_titlecase_first:n #1 + { \@@_change_case:nnnn { title } { break } { } {#1} } +\cs_new:Npn \text_lowercase:nn #1#2 + { \@@_change_case:nnn { lower } {#1} {#2} } +\cs_new:Npn \text_uppercase:nn #1#2 + { \@@_change_case:nnn { upper } {#1} {#2} } +\cs_new:Npn \text_titlecase_all:nn #1#2 + { \@@_change_case:nnn { title } {#1} {#2} } +\cs_new:Npn \text_titlecase_first:nn #1#2 + { \@@_change_case:nnnn { title } { break } {#1} {#2} } +\cs_new:Npn \@@_change_case:nnn #1#2#3 + { \@@_change_case:nnnn {#1} {#1} {#2} {#3} } +% \end{macrocode} +% \end{macro} +% \end{macro} +% \end{macro} +% +% \begin{macro}[EXP] +% { +% \@@_change_case:nnnn , +% \@@_change_case_auxi:nnnn , +% \@@_change_case_auxii:nnnn +% } +% \begin{macro}[EXP]{\@@_change_case_BCP:nnnn} +% \begin{macro}[EXP]{\@@_change_case_BCP:nnnw} +% \begin{macro}[EXP]{\@@_change_case_BCP:nnnnnw} +% \begin{macro}[EXP] +% { +% \@@_change_case_store:n, \@@_change_case_store:o, +% \@@_change_case_store:V, \@@_change_case_store:v, +% \@@_change_case_store:e +% } +% \begin{macro}[EXP]{\@@_change_case_store:nw} +% \begin{macro}[EXP]{\@@_change_case_result:n} +% \begin{macro}[EXP]{\@@_change_case_end:w} +% \begin{macro}[EXP]{\@@_change_case_loop:nnnw} +% \begin{macro}[EXP]{\@@_change_case_break:w, \@@_change_case_break_aux:w} +% \begin{macro}[EXP] +% { +% \@@_change_case_group_lower:nnnn , +% \@@_change_case_group_upper:nnnn , +% \@@_change_case_group_title:nnnn +% } +% \begin{macro}[EXP] +% {\@@_change_case_space:nnnw, \@@_change_case_space_break:nnnw} +% \begin{macro}[EXP] +% {\@@_change_case_N_type:nnnN, \@@_change_case_N_type_aux:nnnN} +% \begin{macro}[EXP]{\@@_change_case_N_type:nnnnN} +% \begin{macro}[EXP]{\@@_change_case_math_search:nnnNNN} +% \begin{macro}[EXP]{\@@_change_case_math_loop:nnnNw} +% \begin{macro}[EXP]{\@@_change_case_math_N_type:nnnNN} +% \begin{macro}[EXP]{\@@_change_case_math_group:nnnNn} +% \begin{macro}[EXP]{\@@_change_case_math_space:nnnNw} +% \begin{macro}[EXP]{\@@_change_case_cs_check:nnnN} +% \begin{macro}[EXP]{\@@_change_case_exclude:nnnN} +% \begin{macro}[EXP]{\@@_change_case_exclude:nnnnN} +% \begin{macro}[EXP]{\@@_change_case_exclude:nnnNN} +% \begin{macro}[EXP]{\@@_change_case_exclude:nnnNw} +% \begin{macro}[EXP]{\@@_change_case_exclude:nnnNnn} +% \begin{macro}[EXP]{\@@_change_case_replace:nnnN} +% \begin{macro}[EXP]{\@@_change_case_replace:nnnn, \@@_change_case_replace:vnnn} +% \begin{macro}[EXP]{\@@_change_case_switch:nnnN} +% \begin{macro}[EXP] +% { +% \@@_change_case_switch_lower:nnnNnnnn , +% \@@_change_case_switch_upper:nnnNnnnn , +% \@@_change_case_switch_title:nnnNnnnn +% } +% \begin{macro}[EXP]{\@@_change_case_skip:nnw} +% \begin{macro}[EXP]{\@@_change_case_skip_N_type:nnN} +% \begin{macro}[EXP]{\@@_change_case_skip_group:nnn} +% \begin{macro}[EXP]{\@@_change_case_skip_space:nnw} +% \begin{macro}[EXP] +% { +% \@@_change_case_letterlike_lower:nnnN , +% \@@_change_case_letterlike_upper:nnnN , +% \@@_change_case_letterlike_title:nnnN +% } +% \begin{macro}[EXP]{\@@_change_case_letterlike:nnnnnN} +% \begin{macro}[EXP] +% { +% \@@_change_case_custom_lower:nnnn , +% \@@_change_case_custom_title:nnnn , +% \@@_change_case_custom_upper:nnnn +% } +% \begin{macro}[EXP]{\@@_change_case_custom:nnnnn} +% \begin{macro}[EXP] +% { +% \@@_change_case_codepoint_lower:nnnn , +% \@@_change_case_codepoint_upper:nnnn , +% \@@_change_case_codepoint_title:nnnn +% } +% \begin{macro}[EXP]{\@@_change_case_lower_sigma:nnnnn} +% \begin{macro}[EXP]{\@@_change_case_lower_sigma:nnnnw} +% \begin{macro}[EXP]{\@@_change_case_lower_sigma:nnnnN} +% \begin{macro}[EXP] +% { +% \@@_change_case_codepoint_title_auxi:nnnn , +% \@@_change_case_codepoint_title_auxii:nnnn +% } +% \begin{macro}[EXP]{\@@_change_case_codepoint_title:nnn} +% \begin{macro}[EXP]{\@@_change_case_codepoint:nnnnn} +% \begin{macro}[EXP]{\@@_change_case_codepoint:nn} +% \begin{macro}[EXP] +% { +% \@@_change_case_codepoint:nnn , +% \@@_change_case_codepoint:fnn , +% \@@_change_case_codepoint_aux:nnn +% } +% \begin{macro}[EXP]{\@@_change_case_codepoint_aux:nnn} +% \begin{macro}[EXP]{\@@_change_case_codepoint_aux:nn} +% \begin{macro}[EXP]{\@@_change_case_catcode:nn} +% \begin{macro}[EXP] +% { +% \@@_change_case_next_lower:nnn , +% \@@_change_case_next_upper:nnn , +% \@@_change_case_next_title:nnn , +% \@@_change_case_next_end:nnn +% } +% As for the expansion code, the business end of case changing is the +% handling of \texttt{N}-type tokens. First, we expand the input fully +% (so the loops here don't need to worry about awkward look-aheads and the +% like). Then we split into the different paths. +% +% The code here needs to be \texttt{f}-type expandable to deal with the +% situation where case changing is applied in running text. There, we +% might have case changing as a document command and the text containing +% other non-expandable document commands. +% \begin{verbatim} +% \cs_set_eq:NN \MakeLowercase \text_lowercase +% ... +% \MakeLowercase{\enquote*{A} text} +% \end{verbatim} +% If we use an \texttt{e}-type expansion and wrap each token in +% \cs{exp_not:n}, that would explode: the document command grabs +% \cs{exp_not:n} as an argument, and things go badly wrong. So we have to +% wrap the entire result in exactly one \cs{exp_not:n}, or rather in the +% kernel version. +% \begin{macrocode} +\cs_new:Npn \@@_change_case:nnnn #1#2#3#4 + { + \__kernel_exp_not:w \exp_after:wN + { + \exp:w + \exp_args:Ne \@@_change_case_auxi:nnnn + { \text_expand:n {#4} } + {#1} {#2} {#3} + } + } +\cs_new:Npn \@@_change_case_auxi:nnnn #1#2#3#4 + { + \exp_args:No \@@_change_case_BCP:nnnn + { \tl_to_str:n {#4} } {#1} {#2} {#3} + } +\cs_new:Npe \@@_change_case_BCP:nnnn #1#2#3#4 + { + \exp_not:N \@@_change_case_BCP:nnnw + {#2} {#3} {#4} #1 \tl_to_str:n { -x- -x- } \exp_not:N \q_@@_stop + } +\use:e + { + \cs_new:Npn \exp_not:N \@@_change_case_BCP:nnnw + #1#2#3#4 \tl_to_str:n { -x- } #5 \tl_to_str:n { -x- } #6 + \exp_not:N \q_@@_stop + } + { \@@_change_case_BCP:nnnnnw {#1} {#2} {#3} {#5} {#4} #4 - \q_@@_stop } +\cs_new:Npn \@@_change_case_BCP:nnnnnw #1#2#3#4#5#6 - #7 \q_@@_stop + { + \bool_lazy_or:nnTF + { \cs_if_exist_p:c { @@_change_case_ #2 _ #6 -x- #4 :nnnnn } } + { \tl_if_exist_p:c { l_@@_ #2 case_special_ #6 -x- #4 _tl } } + { \@@_change_case_auxii:nnnn {#1} {#2} {#3} { #6 -x- #4 } } + { + \cs_if_exist:cTF { @@_change_case_ #2 _ #6 :nnnnn } + { \@@_change_case_auxii:nnnn {#1} {#2} {#3} {#6} } + { \@@_change_case_auxii:nnnn {#1} {#2} {#3} {#5} } + } + } +\cs_new:Npn \@@_change_case_auxii:nnnn #1#2#3#4 + { + \group_align_safe_begin: + \cs_if_exist_use:c { @@_change_case_boundary_ #2 _ #4 :Nnnnw } + \@@_change_case_loop:nnnw {#2} {#3} {#4} #1 + \q_@@_recursion_tail \q_@@_recursion_stop + \@@_change_case_result:n { } + } +% \end{macrocode} +% As for expansion, collect up the tokens for future use. +% \begin{macrocode} +\cs_new:Npn \@@_change_case_store:n #1 + { \@@_change_case_store:nw {#1} } +\cs_generate_variant:Nn \@@_change_case_store:n { o , e , V , v } +\cs_new:Npn \@@_change_case_store:nw #1#2 \@@_change_case_result:n #3 + { #2 \@@_change_case_result:n { #3 #1 } } +\cs_new:Npn \@@_change_case_end:w #1 \@@_change_case_result:n #2 + { + \group_align_safe_end: + \exp_end: + #2 + } +% \end{macrocode} +% The main loop is the standard \texttt{tl action} type. +% \begin{macrocode} +\cs_new:Npn \@@_change_case_loop:nnnw #1#2#3#4 \q_@@_recursion_stop + { + \tl_if_head_is_N_type:nTF {#4} + { \@@_change_case_N_type:nnnN } + { + \tl_if_head_is_group:nTF {#4} + { \use:c { @@_change_case_group_ #1 :nnnn } } + { \@@_change_case_space:nnnw } + } + {#1} {#2} {#3} #4 \q_@@_recursion_stop + } +\cs_new:Npn \@@_change_case_break:w + { \@@_change_case_break_aux:w \prg_do_nothing: } +\cs_new:Npn \@@_change_case_break_aux:w + #1 \q_@@_recursion_tail \q_@@_recursion_stop + { + \@@_change_case_store:o {#1} + \@@_change_case_end:w + } +% \end{macrocode} +% For a group, we \emph{could} worry about whether this contains a character +% or not. However, that would make life very complex for little gain: exactly +% what a first character is is rather weakly-defined anyway. So if there is +% a group, we simply assume that a character has been seen, and for title +% case we switch to the \enquote{rest of the tokens} situation. To avoid +% having too much testing, we use a two-step process here to allow the +% titlecase functions to be separate. +% \begin{macrocode} +\cs_new:Npn \@@_change_case_group_lower:nnnn #1#2#3#4 + { + \@@_change_case_store:o + { + \exp_after:wN + { + \exp:w + \@@_change_case_auxii:nnnn {#4} {#1} {#2} {#3} + } + } + \@@_change_case_loop:nnnw {#1} {#2} {#3} + } +\cs_new_eq:NN \@@_change_case_group_upper:nnnn + \@@_change_case_group_lower:nnnn +\cs_new:Npn \@@_change_case_group_title:nnnn #1#2#3#4 + { + \@@_change_case_store:o + { + \exp_after:wN + { + \exp:w + \@@_change_case_auxii:nnnn {#4} {#1} {#2} {#3} + } + } + \@@_change_case_skip:nnw {#2} {#3} + } +\use:e + { + \cs_new:Npn \exp_not:N \@@_change_case_space:nnnw #1#2#3 \c_space_tl + } + { + \@@_change_case_store:n { ~ } + \cs_if_exist_use:cF { @@_change_case_space_ #2 :nnn } + { + \cs_if_exist_use:c { @@_change_case_boundary_ #1 _ #3 :Nnnnw } + \@@_change_case_loop:nnnw + } + {#2} {#2} {#3} + } +\cs_new:Npn \@@_change_case_space_break:nnn #1#2#3 + { \@@_change_case_break:w } +% \end{macrocode} +% The first step of handling \texttt{N}-type tokens is to filter out the +% end-of-loop. That has to be done separately from the first real step +% as otherwise we pick up the wrong delimiter. The loop here is the same +% as the \texttt{expand} one, just passing the additional data long. If no +% close-math token is found then the final clean-up is forced +% (i.e.~there is no assumption of \enquote{well-behaved} input in terms of +% math mode). +% \begin{macrocode} +\cs_new:Npn \@@_change_case_N_type:nnnN #1#2#3#4 + { + \@@_if_q_recursion_tail_stop_do:Nn #4 + { \@@_change_case_end:w } + \@@_change_case_N_type_aux:nnnN {#1} {#2} {#3} #4 + } +\cs_new:Npn \@@_change_case_N_type_aux:nnnN #1#2#3#4 + { + \exp_args:NV \@@_change_case_N_type:nnnnN + \l_text_math_delims_tl {#1} {#2} {#3} #4 + } +\cs_new:Npn \@@_change_case_N_type:nnnnN #1#2#3#4#5 + { + \@@_change_case_math_search:nnnNNN {#2} {#3} {#4} #5 #1 + \q_@@_recursion_tail \q_@@_recursion_tail + \q_@@_recursion_stop + } +\cs_new:Npn \@@_change_case_math_search:nnnNNN #1#2#3#4#5#6 + { + \@@_if_q_recursion_tail_stop_do:Nn #5 + { \@@_change_case_cs_check:nnnN {#1} {#2} {#3} #4 } + \token_if_eq_meaning:NNTF #4 #5 + { + \@@_use_i_delimit_by_q_recursion_stop:nw + { + \@@_change_case_store:n {#4} + \@@_change_case_math_loop:nnnNw {#1} {#2} {#3} #6 + } + } + { \@@_change_case_math_search:nnnNNN {#1} {#2} {#3} #4 } + } +\cs_new:Npn \@@_change_case_math_loop:nnnNw #1#2#3#4#5 \q_@@_recursion_stop + { + \tl_if_head_is_N_type:nTF {#5} + { \@@_change_case_math_N_type:nnnNN } + { + \tl_if_head_is_group:nTF {#5} + { \@@_change_case_math_group:nnnNn } + { \@@_change_case_math_space:nnnNw } + } + {#1} {#2} {#3} #4 #5 \q_@@_recursion_stop + } +\cs_new:Npn \@@_change_case_math_N_type:nnnNN #1#2#3#4#5 + { + \@@_if_q_recursion_tail_stop_do:Nn #5 + { \@@_change_case_end:w } + \@@_change_case_store:n {#5} + \token_if_eq_meaning:NNTF #5 #4 + { \@@_change_case_loop:nnnw {#1} {#2} {#3} } + { \@@_change_case_math_loop:nnnNw {#1} {#2} {#3} #4 } + } +\cs_new:Npn \@@_change_case_math_group:nnnNn #1#2#3#4#5 + { + \@@_change_case_store:n { {#5} } + \@@_change_case_math_loop:nnnNw {#1} {#2} {#3} #4 + } +\use:e + { + \cs_new:Npn \exp_not:N \@@_change_case_math_space:nnnNw #1#2#3#4 + \c_space_tl + } + { + \@@_change_case_store:n { ~ } + \@@_change_case_math_loop:nnnNw {#1} {#2} {#3} #4 + } +% \end{macrocode} +% Once potential math-mode cases are filtered out the next stage is to +% test if the token grabbed is a control sequence: the two routes the code +% may take are then very different. +% \begin{macrocode} +\cs_new:Npn \@@_change_case_cs_check:nnnN #1#2#3#4 + { + \token_if_cs:NTF #4 + { \@@_change_case_exclude:nnnN {#1} {#2} {#3} } + { + \@@_codepoint_process:nN + { \use:c { @@_change_case_custom_ #1 :nnnn } {#1} {#2} {#3} } + } + #4 + } +% \end{macrocode} +% To deal with a control sequence there is first a need to test if it is +% on the list which indicate that case changing should be skipped. That's +% done using a loop as for the other special cases. If a hit is found then +% the argument is grabbed and passed through as-is. +% \begin{macrocode} +\cs_new:Npn \@@_change_case_exclude:nnnN #1#2#3#4 + { + \exp_args:Ne \@@_change_case_exclude:nnnnN + { + \exp_not:V \l_text_math_arg_tl + \exp_not:V \l_text_case_exclude_arg_tl + } + {#1} {#2} {#3} #4 + } +\cs_new:Npn \@@_change_case_exclude:nnnnN #1#2#3#4#5 + { + \@@_change_case_exclude:nnnNN {#2} {#3} {#4} #5 #1 + \q_@@_recursion_tail \q_@@_recursion_stop + } +\cs_new:Npn \@@_change_case_exclude:nnnNN #1#2#3#4#5 + { + \@@_if_q_recursion_tail_stop_do:Nn #5 + { \@@_change_case_replace:nnnN {#1} {#2} {#3} #4 } + \str_if_eq:nnTF {#4} {#5} + { + \@@_use_i_delimit_by_q_recursion_stop:nw + { \@@_change_case_exclude:nnnNw {#1} {#2} {#3} #4 } + } + { \@@_change_case_exclude:nnnNN {#1} {#2} {#3} #4 } + } +\cs_new:Npn \@@_change_case_exclude:nnnNw #1#2#3#4#5# + { \@@_change_case_exclude:nnnNnn {#1} {#2} {#3} {#4} {#5} } +\cs_new:Npn \@@_change_case_exclude:nnnNnn #1#2#3#4#5#6 + { + \tl_if_blank:nTF {#5} + { \@@_change_case_store:n { #4 {#6} } } + { + \@@_change_case_store:o + { + \exp_after:wN #4 + \exp:w \@@_change_case_auxii:nnnn {#5} {#1} {#2} {#3} + {#6} + } + } + \@@_change_case_loop:nnnw {#1} {#2} {#3} + } +% \end{macrocode} +% Deal with any specialist replacement for case changing. +% \begin{macrocode} +\cs_new:Npn \@@_change_case_replace:nnnN #1#2#3#4 + { + \cs_if_exist:cTF { l_@@_case_ \token_to_str:N #4 _tl } + { + \@@_change_case_replace:vnnn + { l_@@_case_ \token_to_str:N #4 _tl } {#1} {#2} {#3} + } + { \@@_change_case_switch:nnnN {#1} {#2} {#3} #4 } + } +\cs_new:Npn \@@_change_case_replace:nnnn #1#2#3#4 + { \@@_change_case_loop:nnnw {#2} {#3} {#4} #1 } +\cs_generate_variant:Nn \@@_change_case_replace:nnnn { v } +% \end{macrocode} +% Allow for manually-controlled case switching. +% \begin{macrocode} +\cs_new:Npn \@@_change_case_switch:nnnN #1#2#3#4 + { + \cs_if_eq:NNTF #4 \text_case_switch:nnnn + { \use:c { @@_change_case_switch_ #1 :nnnNnnnn } } + { \use:c { @@_change_case_letterlike_ #1 :nnnN } } + {#1} {#2} {#3} #4 + } +\cs_new:Npn \@@_change_case_switch_lower:nnnNnnnn #1#2#3#4#5#6#7#8 + { + \@@_change_case_store:n {#7} + \@@_change_case_loop:nnnw {#1} {#2} {#3} + } +\cs_new:Npn \@@_change_case_switch_upper:nnnNnnnn #1#2#3#4#5#6#7#8 + { + \@@_change_case_store:n {#6} + \@@_change_case_loop:nnnw {#1} {#2} {#3} + } +\cs_new:Npn \@@_change_case_switch_title:nnnNnnnn #1#2#3#4#5#6#7#8 + { + \@@_change_case_store:n {#8} + \@@_change_case_skip:nnw {#2} {#3} + } +% \end{macrocode} +% Skip over material quickly after titlecase-first-only initials +% \begin{macrocode} +\cs_new:Npn \@@_change_case_skip:nnw #1#2#3 \q_@@_recursion_stop + { + \tl_if_head_is_N_type:nTF {#3} + { \@@_change_case_skip_N_type:nnN } + { + \tl_if_head_is_group:nTF {#3} + { \@@_change_case_skip_group:nnn } + { \@@_change_case_skip_space:nnw } + } + {#1} {#2} #3 \q_@@_recursion_stop + } +\cs_new:Npn \@@_change_case_skip_N_type:nnN #1#2#3 + { + \@@_if_q_recursion_tail_stop_do:Nn #3 + { \@@_change_case_end:w } + \@@_change_case_store:n {#3} + \@@_change_case_skip:nnw {#1} {#2} + } +\cs_new:Npn \@@_change_case_skip_group:nnn #1#2#3 + { + \@@_change_case_store:n { {#3} } + \@@_change_case_skip:nnw {#1} {#2} + } +\cs_new:Npn \@@_change_case_skip_space:nnw #1#2 + { \@@_change_case_space:nnnw {#1} {#1} {#2} } +% \end{macrocode} +% Letter-like commands may still be present: they are set up using a simple +% lookup approach, so can easily be handled with no loop. If there is no +% hit, we are at the end of the process: we loop around. Letter-like chars +% are all available only in upper- and lowercase, so titlecasing maps to the +% uppercase version. +% \begin{macrocode} +\cs_new:Npn \@@_change_case_letterlike_lower:nnnN #1#2#3#4 + { \@@_change_case_letterlike:nnnnnN {#1} {#1} {#1} {#2} {#3} #4 } +\cs_new_eq:NN \@@_change_case_letterlike_upper:nnnN + \@@_change_case_letterlike_lower:nnnN +\cs_new:Npn \@@_change_case_letterlike_title:nnnN #1#2#3#4 + { \@@_change_case_letterlike:nnnnnN { upper } { end } {#1} {#2} {#3} #4 } +\cs_new:Npn \@@_change_case_letterlike:nnnnnN #1#2#3#4#5#6 + { + \cs_if_exist:cTF { c_@@_ #1 case_ \token_to_str:N #6 _tl } + { + \@@_change_case_store:v + { c_@@_ #1 case_ \token_to_str:N #6 _tl } + \use:c { @@_change_case_next_ #2 :nnn } {#2} {#4} {#5} + } + { + \@@_change_case_store:n {#6} + \cs_if_exist:cTF + { + c_@@_ + \str_if_eq:nnTF {#1} { lower } { upper } { lower } + case_ \token_to_str:N #6 _tl + } + { \use:c { @@_change_case_next_ #2 :nnn } {#2} {#4} {#5} } + { \@@_change_case_loop:nnnw {#3} {#4} {#5} } + } + } +% \end{macrocode} +% Check for a customised codepoint result. +% \begin{macrocode} +\cs_new:Npn \@@_change_case_custom_lower:nnnn #1#2#3#4 + { + \@@_change_case_custom:nnnnnn {#1} {#1} {#2} {#3} {#4} + { \use:c { @@_change_case_codepoint_ #1 :nnnn } {#1} {#2} {#3} {#4} } + } +\cs_new_eq:NN \@@_change_case_custom_upper:nnnn + \@@_change_case_custom_lower:nnnn +\cs_new:Npn \@@_change_case_custom_title:nnnn #1#2#3#4 + { + \@@_change_case_custom:nnnnnn { title } {#1} {#2} {#3} {#4} + { + \@@_change_case_custom:nnnnnn { upper } {#1} {#2} {#3} {#4} + { \use:c { @@_change_case_codepoint_ #1 :nnnn } {#1} {#2} {#3} {#4} } + } + } +\cs_new:Npn \@@_change_case_custom:nnnnnn #1#2#3#4#5#6 + { + \tl_if_exist:cTF { l_@@_ #1 case _ \tl_to_str:n {#5} _ #4 _tl } + { + \@@_change_case_replace:vnnn + { l_@@_ #1 case _ \tl_to_str:n {#5} _ #4 _tl } {#2} {#3} {#4} + } + { + \tl_if_exist:cTF { l_@@_ #1 case _ \tl_to_str:n {#5} _tl } + { + \@@_change_case_replace:vnnn + { l_@@_ #1 case _ \tl_to_str:n {#5} _tl } {#2} {#3} {#4} + } + {#6} + } + } +% \end{macrocode} +% For upper- and lowercase changes, once we get to this stage there are only +% a couple of questions remaining: is there a language-specific mapping and +% is there the special case of a terminal sigma. If not, then we pass to +% a simple codepoint mapping. +% \begin{macrocode} +\cs_new:Npn \@@_change_case_codepoint_lower:nnnn #1#2#3#4 + { + \cs_if_exist_use:cF { @@_change_case_lower_ #3 :nnnnn } + { \@@_change_case_lower_sigma:nnnnn } + {#1} {#1} {#2} {#3} {#4} + } +\cs_new:Npn \@@_change_case_codepoint_upper:nnnn #1#2#3#4 + { + \cs_if_exist_use:cF { @@_change_case_upper_ #3 :nnnnn } + { \@@_change_case_codepoint:nnnnn } + {#1} {#1} {#2} {#3} {#4} + } +% \end{macrocode} +% If the current character is an uppercase sigma, the a check is made on the +% next item in the input. If it is \texttt{N}-type and not a control sequence +% then there is a look-ahead phase: the logic here is simply based on letters +% or actives (to cover $8$-bit engines). +% \begin{macrocode} +\cs_new:Npn \@@_change_case_lower_sigma:nnnnn #1#2#3#4#5 + { + \@@_codepoint_compare:nNnTF {#5} = { "03A3 } + { \@@_change_case_lower_sigma:nnnnw {#2} } + { \@@_change_case_codepoint:nnnnn {#1} {#2} } + {#3} {#4} {#5} + } +\cs_new:Npn \@@_change_case_lower_sigma:nnnnw #1#2#3#4#5 \q_@@_recursion_stop + { + \tl_if_head_is_N_type:nTF {#5} + { \@@_change_case_lower_sigma:nnnnN {#4} } + { + \@@_change_case_store:e + { \codepoint_generate:nn { "03C2 } { \@@_char_catcode:N #4 } } + \@@_change_case_loop:nnnw + } + {#1} {#2} {#3} #5 \q_@@_recursion_stop + } +\cs_new:Npn \@@_change_case_lower_sigma:nnnnN #1#2#3#4#5 + { + \@@_change_case_store:e + { + \bool_lazy_or:nnTF + { \token_if_letter_p:N #5 } + { + \bool_lazy_and_p:nn + { \token_if_active_p:N #5 } + { \int_compare_p:nNn {`#5} > { "80 } } + } + { \codepoint_generate:nn { "03C3 } { \@@_char_catcode:N #1 } } + { \codepoint_generate:nn { "03C2 } { \@@_char_catcode:N #1 } } + } + \@@_change_case_loop:nnnw {#2} {#3} {#4} #5 + } +% \end{macrocode} +% For titlecasing, we need to obtain the general category of the current +% codepoint. +% \begin{macrocode} +\cs_new:Npn \@@_change_case_codepoint_title:nnnn #1#2#3#4 + { + \bool_if:NTF \l_text_titlecase_check_letter_bool + { + \exp_args:Ne \@@_change_case_codepoint_title_auxi:nnnn + { + \codepoint_to_category:n + { \@@_codepoint_from_chars:Nw #4 } + } + } + { \@@_change_case_codepoint_title:nnn } + {#2} {#3} {#4} + } +\cs_new:Npn \@@_change_case_codepoint_title_auxi:nnnn #1#2#3#4 + { + \tl_if_head_eq_charcode:nNTF {#1} { L } + { \@@_change_case_codepoint_title:nnn } + { \@@_change_case_codepoint_title_auxii:nnnn { title } } + {#2} {#3} {#4} + } +\cs_new:Npn \@@_change_case_codepoint_title:nnn #1#2#3 + { \@@_change_case_codepoint_title_auxii:nnnn { end } {#1} {#2} {#3} } +\cs_new:Npn \@@_change_case_codepoint_title_auxii:nnnn #1#2#3#4 + { + \cs_if_exist_use:cF { @@_change_case_title_ #3 :nnnnn } + { + \cs_if_exist_use:cF { @@_change_case_upper_ #3 :nnnnn } + { \@@_change_case_codepoint:nnnnn } + } + { title } {#1} {#2} {#3} {#4} + } +\cs_new:Npn \@@_change_case_codepoint:nnnnn #1#2#3#4#5 + { + \bool_lazy_and:nnTF + { \tl_if_single_p:n {#5} } + { \token_if_active_p:N #5 } + { \@@_change_case_store:n {#5} } + { + \@@_change_case_store:e + { \@@_change_case_codepoint:nn {#1} {#5} } + } + \use:c { @@_change_case_next_ #2 :nnn } {#2} {#3} {#4} + } +\cs_new:Npn \@@_change_case_codepoint:nn #1#2 + { + \@@_change_case_codepoint:fnn + { \int_eval:n { \@@_codepoint_from_chars:Nw #2 } } {#1} {#2} + } +\cs_new:Npn \@@_change_case_codepoint:nnn #1#2#3 + { + \exp_args:Ne \@@_change_case_codepoint_aux:nn + { \__kernel_codepoint_case:nn { #2 case } {#1} } {#3} + } +\cs_generate_variant:Nn \@@_change_case_codepoint:nnn { f } +% \end{macrocode} +% Avoid high chars with p\TeX{}. +% \begin{macrocode} +\sys_if_engine_ptex:T + { + \cs_new_eq:NN \@@_change_case_codepoint_aux:nnn + \@@_change_case_codepoint:nnn + \cs_gset:Npn \@@_change_case_codepoint:nnn #1#2#3 + { + \int_compare:nNnTF {#1} = { -1 } + { \exp_not:n {#3} } + { \@@_change_case_codepoint_aux:nnn {#1} {#2} {#3} } + } + } +\cs_new:Npn \@@_change_case_codepoint_aux:nn #1#2 + { + \use:e { \@@_change_case_codepoint_aux:nnnn #1 {#2} } + } +\cs_new:Npn \@@_change_case_codepoint_aux:nnnn #1#2#3#4 + { + \@@_codepoint_compare:nNnTF {#4} = {#1} + { \exp_not:n {#4} } + { + \codepoint_generate:nn {#1} + { \@@_change_case_catcode:nn {#4} {#1} } + \tl_if_blank:nF {#2} + { + \codepoint_generate:nn {#2} + { \char_value_catcode:n {#2} } + \tl_if_blank:nF {#3} + { + \codepoint_generate:nn {#3} + { \char_value_catcode:n {#3} } + } + } + } + } +% \end{macrocode} +% We need to ensure that only valid catcode-extraction is attempted. That's +% fine with Unicode engines but needs a bit of work with 8-bit ones. The +% logic is that if the original codepoint was in the ASCII range, we keep +% the catcode. Otherwise, if the target is in the ASCII range, we use +% the standard catcode. If neither are true, we set as 13 on the grounds that +% this will be what is used anyway! +% \begin{macrocode} +\bool_lazy_or:nnTF + { \sys_if_engine_luatex_p: } + { \sys_if_engine_xetex_p: } + { + \cs_new:Npn \@@_change_case_catcode:nn #1#2 + { \@@_char_catcode:N #1 } + } + { + \cs_new:Npn \@@_change_case_catcode:nn #1#2 + { + \@@_codepoint_compare:nNnTF {#1} < { "80 } + { \@@_char_catcode:N #1 } + { + \int_compare:nNnTF {#2} < { "80 } + { \char_value_catcode:n {#2} } + { 13 } + } + } + } +\cs_new:Npn \@@_change_case_next_lower:nnn #1#2#3 + { \@@_change_case_loop:nnnw {#1} {#2} {#3} } +\cs_new_eq:NN \@@_change_case_next_upper:nnn + \@@_change_case_next_lower:nnn +\cs_new_eq:NN \@@_change_case_next_title:nnn + \@@_change_case_next_lower:nnn +\cs_new:Npn \@@_change_case_next_end:nnn #1#2#3 + { \@@_change_case_skip:nnw {#2} {#3} } +% \end{macrocode} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% +% \begin{macro}{\text_declare_case_equivalent:Nn} +% Create equivalents to allow replacement. +% \begin{macrocode} +\cs_new_protected:Npn \text_declare_case_equivalent:Nn #1#2 + { + \tl_clear_new:c { l_@@_case_ \token_to_str:N #1 _tl } + \tl_set:cn { l_@@_case_ \token_to_str:N #1 _tl } {#2} + } +% \end{macrocode} +% \end{macro} +% +% \begin{macro} +% { +% \text_declare_lowercase_mapping:nn , +% \text_declare_titlecase_mapping:nn , +% \text_declare_uppercase_mapping:nn +% } +% \begin{macro} +% {\@@_declare_case_mapping:nnn, \@@_declare_case_mapping_aux:nnn} +% \begin{macro} +% { +% \text_declare_lowercase_mapping:nnn , +% \text_declare_titlecase_mapping:nnn , +% \text_declare_uppercase_mapping:nnn +% } +% \begin{macro} +% {\@@_declare_case_mapping:nnnn, \@@_declare_case_mapping_aux:nnnn} +% Codepoint customisation. +% \begin{macrocode} +\cs_new_protected:Npn \text_declare_lowercase_mapping:nn #1#2 + { \@@_declare_case_mapping:nnn { lower } {#1} {#2} } +\cs_new_protected:Npn \text_declare_titlecase_mapping:nn #1#2 + { \@@_declare_case_mapping:nnn { title } {#1} {#2} } +\cs_new_protected:Npn \text_declare_uppercase_mapping:nn #1#2 + { \@@_declare_case_mapping:nnn { upper } {#1} {#2} } +\cs_new_protected:Npn \@@_declare_case_mapping:nnn #1#2#3 + { + \exp_args:Ne \@@_declare_case_mapping_aux:nnn + { \codepoint_str_generate:n {#2} } {#1} {#3} + } +\cs_new_protected:Npn \@@_declare_case_mapping_aux:nnn #1#2#3 + { + \tl_clear_new:c { l_@@_ #2 case _ #1 _tl } + \tl_set:cn { l_@@_ #2 case _ #1 _ tl } {#3} + } +\cs_new_protected:Npn \text_declare_lowercase_mapping:nnn #1#2#3 + { \@@_declare_case_mapping:nnnn { lower } {#1} {#2} {#3} } +\cs_new_protected:Npn \text_declare_titlecase_mapping:nnn #1#2#3 + { \@@_declare_case_mapping:nnnn { title } {#1} {#2} {#3} } +\cs_new_protected:Npn \text_declare_uppercase_mapping:nnn #1#2#3 + { \@@_declare_case_mapping:nnnn { upper } {#1} {#2} {#3} } +\cs_new_protected:Npn \@@_declare_case_mapping:nnnn #1#2#3#4 + { + \exp_args:Ne \@@_declare_case_mapping_aux:nnnn + { \codepoint_str_generate:n {#3} } {#1} {#2} {#4} + } +\cs_new_protected:Npn \@@_declare_case_mapping_aux:nnnn #1#2#3#4 + { + \tl_clear_new:c { l_@@_ #2 case _ #1 _ #3 _tl } + \tl_set:cn { l_@@_ #2 case _ #1 _ #3 _ tl } {#4} + \tl_clear_new:c { l_@@_ #2 case_special_ #3 _tl } + } +% \end{macrocode} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% +% \begin{macro}{\text_case_switch:nnnn} +% \begin{macro}{\@@_case_switch_marker:} +% Set up the mechanism for manual case switching. +% \begin{macrocode} +\cs_new:Npn \text_case_switch:nnnn #1#2#3#4 + { + \@@_case_switch_marker: + #1 + } +\cs_new:Npn \@@_case_switch_marker: { } +% \end{macrocode} +% \end{macro} +% \end{macro} +% +% \begin{macro}[EXP]{\@@_change_case_generate:n} +% A utility. +% \begin{macrocode} +\cs_new:Npn \@@_change_case_generate:n #1 + { \codepoint_generate:nn {#1} { \char_value_catcode:n {#1} } } +% \end{macrocode} +% \end{macro} +% +% \begin{macro}[EXP] +% { +% \@@_change_case_upper_de-x-eszett:nnnnn, +% \@@_change_case_upper_de-alt:nnnnn +% } +% A simple alternative version for German. +% \begin{macrocode} +\cs_new:cpn { @@_change_case_upper_de-x-eszett:nnnnn } #1#2#3#4#5 + { + \@@_codepoint_compare:nNnTF {#5} = { "00DF } + { + \@@_change_case_store:e + { + \codepoint_generate:nn { "1E9E } + { \@@_change_case_catcode:nn {#5} { "1E9E } } + } + \use:c { @@_change_case_next_ #2 :nnn } + {#2} {#3} {#4} + } + { \@@_change_case_codepoint:nnnnn {#1} {#2} {#3} {#4} {#5} } + } +\cs_new_eq:cc { @@_change_case_upper_de-alt:nnnnn } + { @@_change_case_upper_de-x-eszett:nnnnn } +% \end{macrocode} +% \end{macro} +% +% \begin{macro}[EXP] +% { +% \@@_change_case_upper_el:nnnnn , +% \@@_change_case_upper_el-x-iota:nnnnn , +% \@@_change_case_upper_el_aux:nnnnn +% } +% \begin{macro}[EXP]{\@@_change_case_upper_el:nnnn} +% \begin{macro}[EXP]{\@@_change_case_upper_el:nnnnw} +% \begin{macro}[EXP] +% {\@@_change_case_upper_el:nnnnN, \@@_change_case_upper_el_aux:nnnnN} +% \begin{macro}[EXP]{\@@_change_case_upper_el_ypogegrammeni:nnnnnnw} +% \begin{macro}[EXP]{\@@_change_case_upper_el_ypogegrammeni:nnnnnnN} +% \begin{macro}[EXP]{\@@_change_case_upper_el_ypogegrammeni:nnnnnnn} +% \begin{macro}[EXP]{\@@_change_case_upper_el_dialytika:nnnn} +% \begin{macro}[EXP]{\@@_change_case_upper_el_dialytika:n} +% \begin{macro}[EXP]{\@@_change_case_upper_el_hiatus:nnnnw} +% \begin{macro}[EXP]{\@@_change_case_upper_el_hiatus:nnnnN} +% \begin{macro}[EXP]{\@@_change_case_upper_el_hiatus:nnnnn} +% \begin{macro}[EXP] +% { +% \@@_change_case_upper_el_ypogegrammeni:n , +% \@@_change_case_upper_el-x-iota_ypogegrammeni:n +% } +% \begin{macro}[EXP]{\@@_change_case_upper_el_stress:nn} +% \begin{macro}[EXP]{\@@_change_case_upper_el_gobble:nnnw} +% \begin{macro}[EXP]{\@@_change_case_upper_el_gobble:nnnN} +% \begin{macro}[EXP]{\@@_change_case_upper_el_gobble:nnnn} +% \begin{macro}[EXP,noTF] +% { +% \@@_change_case_if_greek:n , +% \@@_change_case_if_greek_spacing_diacritic:n , +% \@@_change_case_if_greek_accent:n , +% \@@_change_case_if_greek_breathing:n , +% \@@_change_case_if_greek_stress:n , +% \@@_change_case_if_takes_dialytika:n , +% \@@_change_case_if_takes_ypogegrammeni:n +% } +% For Greek uppercasing, we need to know if characters \emph{in the Greek +% range} have accents. That means doing a \textsc{nfd} conversion first, then +% starting a search. As described by the Unicode \textsc{cldr}, Greek accents +% need to be found \emph{after} any U+0308 (diaeresis) and are done in two +% groups to allow for the canonical ordering. The implementation here follows +% the data and examples from \textsc{icu} +% (\url{https://icu.unicode.org/design/case/greek-upper}), +% although necessarily the implementation is somewhat different. The +% \emph{ypogegrammeni} is filtered out here as it is not actually in the +% Greek range, so gets lost if we leave until later. The one Greek codepoint +% we skip is the numeral sign and question mark: the first has an awkward NFD +% for \pdfTeX{} so is best left unchanged, and the latter has issues concerning +% how \texttt{LGR} outputs the input and output (differently!). +% \begin{macrocode} +\cs_new:Npn \@@_change_case_upper_el:nnnnn #1#2#3#4#5 + { + \bool_lazy_and:nnTF + { \@@_change_case_if_greek_p:n {#5} } + { + ! \bool_lazy_or_p:nn + { \@@_codepoint_compare_p:nNn {#5} = { "0374 } } + { \@@_codepoint_compare_p:nNn {#5} = { "037E } } + } + { + \@@_change_case_if_greek_spacing_diacritic:nTF {#5} + { + \@@_change_case_store:n {#5} + \@@_change_case_loop:nnnw + } + { + \exp_args:Ne \@@_change_case_upper_el:nnnn + { + \codepoint_to_nfd:n + { \@@_codepoint_from_chars:Nw #5 } + } + } + {#2} {#3} {#4} + } + { + \@@_codepoint_compare:nNnTF {#5} = { "0345 } + { + \@@_change_case_store:e + { + \codepoint_generate:nn { "0399 } + { \char_value_catcode:n { "0399 } } + } + \@@_change_case_loop:nnnw {#2} {#3} {#4} + } + { \@@_change_case_codepoint:nnnnn {#1} {#2} {#3} {#4} {#5} } + } + } +\cs_new_eq:cN { @@_change_case_upper_el-x-iota:nnnnn } + \@@_change_case_upper_el:nnnnn +\cs_new:Npn \@@_change_case_upper_el:nnnn #1#2#3#4 + { + \@@_codepoint_process:nN + { \@@_change_case_upper_el:nnnnw {#2} {#3} {#4} } #1 + } +% \end{macrocode} +% At this stage we have the first NFD codepoint as |#3|. What we need to know +% is whether after that we have another character, either from the NFD or +% directly in the input. If not, we store the changed character at this stage. +% \begin{macrocode} +\cs_new:Npn \@@_change_case_upper_el:nnnnw #1#2#3#4#5 \q_@@_recursion_stop + { + \tl_if_head_is_N_type:nTF {#5} + { \@@_change_case_upper_el:nnnnN {#4} } + { + \@@_change_case_store:e + { \@@_change_case_codepoint:nn { upper } {#4} } + \@@_change_case_loop:nnnw + } + {#1} {#2} {#3} #5 \q_@@_recursion_stop + } +% \end{macrocode} +% Now, we check the detail of the next codepoint: again we filter out the +% not-a-char cases, before checking if it's an dialytika, accent or diacritic. +% (The latter do not have the same hiatus behavior as accents.) There is +% additional work if the codepoint can take a ypogegrammeni: there, we need +% to move any ypogegrammeni to after accents (in case the input is not +% normalised). The ypogegrammeni itself is handled separately. +% \begin{macrocode} +\cs_new:Npn \@@_change_case_upper_el:nnnnN #1#2#3#4#5 + { + \token_if_cs:NTF #5 + { + \@@_change_case_store:e + { \@@_change_case_codepoint:nn { upper } {#1} } + \@@_change_case_loop:nnnw {#2} {#3} {#4} #5 + } + { + \@@_change_case_if_takes_ypogegrammeni:nTF {#1} + { + \@@_change_case_upper_el_ypogegrammeni:nnnnnnw + {#1} {#2} {#3} {#4} { } { } #5 + } + { \@@_change_case_upper_el_aux:nnnnN {#1} {#2} {#3} {#4} #5 } + } + } +\cs_new:Npn \@@_change_case_upper_el_ypogegrammeni:nnnnnnw + #1#2#3#4#5#6#7 \q_@@_recursion_stop + { + \tl_if_head_is_N_type:nTF {#7} + { + \@@_change_case_upper_el_ypogegrammeni:nnnnnnN + {#1} {#2} {#3} {#4} {#5} {#6} + } + { \@@_change_case_upper_el_aux:nnnnN {#1} {#2} {#3} {#4} #5#6 } + #7 \q_@@_recursion_stop + } +\cs_new:Npn \@@_change_case_upper_el_ypogegrammeni:nnnnnnN #1#2#3#4#5#6#7 + { + \token_if_cs:NTF #7 + { \@@_change_case_upper_el_aux:nnnnN {#1} {#2} {#3} {#4} #5#6 } + { + \@@_codepoint_process:nN + { + \@@_change_case_upper_el_ypogegrammeni:nnnnnnn + {#1} {#2} {#3} {#4} {#5} {#6} + } + } + #7 + } +\cs_new:Npn \@@_change_case_upper_el_ypogegrammeni:nnnnnnn #1#2#3#4#5#6#7 + { + \@@_codepoint_compare:nNnTF {#7} = { "0345 } + { + \@@_change_case_upper_el_ypogegrammeni:nnnnnnw + {#1} {#2} {#3} {#4} {#5} {#7} + } + { + \bool_lazy_or:nnTF + { \@@_change_case_if_greek_accent_p:n {#7} } + { \@@_change_case_if_greek_breathing_p:n {#7} } + { + \@@_change_case_upper_el_ypogegrammeni:nnnnnnw + {#1} {#2} {#3} {#4} {#5#7} {#6} + } + { \@@_change_case_upper_el_aux:nnnnN {#1} {#2} {#3} {#4} #5#6 #7 } + } + } +\cs_new:Npn \@@_change_case_upper_el_aux:nnnnN #1#2#3#4#5 + { + \@@_codepoint_process:nN + { \@@_change_case_upper_el_aux:nnnnn {#1} {#2} {#3} {#4} } #5 + } +\cs_new:Npn \@@_change_case_upper_el_aux:nnnnn #1#2#3#4#5 + { + \@@_codepoint_compare:nNnTF {#5} = { "0308 } + { \@@_change_case_upper_el_dialytika:nnnn {#2} {#3} {#4} {#1} } + { + \@@_change_case_if_greek_accent:nTF {#5} + { \@@_change_case_upper_el_hiatus:nnnnw {#2} {#3} {#4} {#1} } + { + \@@_change_case_if_greek_breathing:nTF {#5} + { \@@_change_case_upper_el:nnnn {#1} {#2} {#3} {#4} } + { + \@@_codepoint_compare:nNnTF {#5} = { "0345 } + { + \@@_change_case_store:e + { \use:c { @@_change_case_upper_ #4 _ypogegrammeni:n } {#1} } + \@@_change_case_loop:nnnw {#2} {#3} {#4} + } + { + \@@_change_case_if_greek_stress:nTF {#5} + { + \@@_change_case_store:e + { \@@_change_case_upper_el_stress:nn {#1} {#5} } + \@@_change_case_loop:nnnw {#2} {#3} {#4} + + } + { + \@@_change_case_store:e + { \@@_change_case_codepoint:nn { upper } {#1} } + \@@_change_case_loop:nnnw {#2} {#3} {#4} #5 + } + } + } + } + } + } +% \end{macrocode} +% We handle \emph{dialytika} in parts as it's also needed for the hiatus. +% We know only two letters take it, so we can shortcut here on the second +% part of the tests. +% \begin{macrocode} +\cs_new:Npn \@@_change_case_upper_el_dialytika:nnnn #1#2#3#4 + { + \@@_change_case_if_takes_dialytika:nTF {#4} + { \@@_change_case_upper_el_dialytika:n {#4} } + { + \@@_change_case_store:e + { \@@_change_case_codepoint:nn { upper } {#4} } + } + \@@_change_case_upper_el_gobble:nnnw {#1} {#2} {#3} + } +\cs_new:Npn \@@_change_case_upper_el_dialytika:n #1 + { + \@@_change_case_store:e + { + \bool_lazy_or:nnTF + { \@@_codepoint_compare_p:nNn {#1} = { "0399 } } + { \@@_codepoint_compare_p:nNn {#1} = { "03B9 } } + { + \codepoint_generate:nn { "03AA } + { \@@_change_case_catcode:nn {#1} { "03AA } } + } + { + \codepoint_generate:nn { "03AB } + { \@@_change_case_catcode:nn {#1} { "03AB } } + } + } + } +% \end{macrocode} +% Adding a hiatus needs some of the same ideas, but if there is not one we +% skip this code point, hence needing a separate function. +% \begin{macrocode} +\cs_new:Npn \@@_change_case_upper_el_hiatus:nnnnw + #1#2#3#4#5 \q_@@_recursion_stop + { + \tl_if_head_is_N_type:nTF {#5} + { \@@_change_case_upper_el_hiatus:nnnnN {#4} } + { + \@@_change_case_store:e + { \@@_change_case_codepoint:nn { upper } {#4} } + \@@_change_case_loop:nnnw + } + {#1} {#2} {#3} #5 \q_@@_recursion_stop + } +\cs_new:Npn \@@_change_case_upper_el_hiatus:nnnnN #1#2#3#4#5 + { + \token_if_cs:NTF #5 + { + \@@_change_case_store:e + { \@@_change_case_codepoint:nn { upper } {#1} } + \@@_change_case_loop:nnnw {#2} {#3} {#4} #5 + } + { + \@@_codepoint_process:nN + { \@@_change_case_upper_el_hiatus:nnnnn {#1} {#2} {#3} {#4} } #5 + } + } +\cs_new:Npn \@@_change_case_upper_el_hiatus:nnnnn #1#2#3#4#5 + { + \@@_change_case_if_takes_dialytika:nTF {#5} + { + \@@_change_case_store:e + { \@@_change_case_codepoint:nn { upper } {#1} } + \@@_change_case_upper_el_dialytika:n {#5} + \@@_change_case_upper_el_gobble:nnnw {#2} {#3} {#4} + } + { \@@_change_case_upper_el:nnnn {#1} {#2} {#3} {#4} #5 } + } +% \end{macrocode} +% Handling the \emph{ypogegrammeni} output depends on the selected approach +% \begin{macrocode} +\cs_new:Npn \@@_change_case_upper_el_ypogegrammeni:n #1 + { + \exp_args:Ne \@@_change_case_generate:n + { + \int_case:nn + { \@@_codepoint_from_chars:Nw #1 } + { + { "0391 } { "1FBC } + { "03B1 } { "1FBC } + { "0397 } { "1FCC } + { "03B7 } { "1FCC } + { "03A9 } { "1FFC } + { "03C9 } { "1FFC } + } + } + } +\cs_new:cpn { @@_change_case_upper_el-x-iota_ypogegrammeni:n } #1 + { + \@@_change_case_codepoint:nn { upper } {#1} + \codepoint_generate:nn { "0399 } + { \char_value_catcode:n { "0399 } } + } +% \end{macrocode} +% We choose to retain stress diacritics, but we also need to recombine +% them for pdf\TeX{}. That is handled here. +% \begin{macrocode} +\cs_new:Npn \@@_change_case_upper_el_stress:nn #1#2 + { + \exp_args:Ne \@@_change_case_generate:n + { + \int_case:nn + { \@@_codepoint_from_chars:Nw #2 } + { + { "0304 } + { + \int_case:nn { \@@_codepoint_from_chars:Nw #1 } + { + { "0391 } { "1FB9 } + { "03B1 } { "1FB9 } + { "0399 } { "1FD9 } + { "03B9 } { "1FD9 } + { "03A5 } { "1FE9 } + { "03C5 } { "1FE9 } + } + } + { "0306 } + { + \int_case:nn { \@@_codepoint_from_chars:Nw #1 } + { + { "0391 } { "1FB8 } + { "03B1 } { "1FB8 } + { "0399 } { "1FD8 } + { "03B9 } { "1FD8 } + { "03A5 } { "1FE8 } + { "03C5 } { "1FE8 } + } + } + } + } + } +% \end{macrocode} +% For clearing out trailing combining marks after we have dealt with +% the first one. +% \begin{macrocode} +\cs_new:Npn \@@_change_case_upper_el_gobble:nnnw + #1#2#3#4 \q_@@_recursion_stop + { + \tl_if_head_is_N_type:nTF {#4} + { \@@_change_case_upper_el_gobble:nnnN } + { \@@_change_case_loop:nnnw } + {#1} {#2} {#3} #4 \q_@@_recursion_stop + } +\cs_new:Npn \@@_change_case_upper_el_gobble:nnnN #1#2#3#4 + { + \token_if_cs:NTF #4 + { \@@_change_case_loop:nnnw {#1} {#2} {#3} } + { + \@@_codepoint_process:nN + { \@@_change_case_upper_el_gobble:nnnn {#1} {#2} {#3} } + } + #4 + } +\cs_new:Npn \@@_change_case_upper_el_gobble:nnnn #1#2#3#4 + { + \bool_lazy_or:nnTF + { \@@_change_case_if_greek_accent_p:n {#4} } + { \@@_change_case_if_greek_breathing_p:n {#4} } + { \@@_change_case_upper_el_gobble:nnnw {#1} {#2} {#3} } + { \@@_change_case_loop:nnnw {#1} {#2} {#3} #4 } + } +% \end{macrocode} +% Luckily the Greek range is limited and clear. +% \begin{macrocode} +\prg_new_conditional:Npnn \@@_change_case_if_greek:n #1 { p , TF } + { + \exp_args:Nf \@@_change_case_if_greek:n + { \int_eval:n { \@@_codepoint_from_chars:Nw #1 } } + } +\cs_new:Npn \@@_change_case_if_greek:n #1 + { + \if_int_compare:w #1 < "0370 \exp_stop_f: + \prg_return_false: + \else: + \if_int_compare:w #1 > "03FF \exp_stop_f: + \if_int_compare:w #1 < "1F00 \exp_stop_f: + \prg_return_false: + \else: + \if_int_compare:w #1 > "1FFF \exp_stop_f: + \if_int_compare:w #1 = "2126 \exp_stop_f: + \prg_return_true: + \else: + \prg_return_false: + \fi: + \else: + \prg_return_true: + \fi: + \fi: + \else: + \prg_return_true: + \fi: + \fi: + } +% \end{macrocode} +% We follow ICU in adding a few extras to the accent list here. +% \begin{macrocode} +\prg_new_conditional:Npnn \@@_change_case_if_greek_accent:n #1 { TF , p } + { + \exp_args:Nf \@@_change_case_if_greek_accent:n + { \int_eval:n { \@@_codepoint_from_chars:Nw #1 } } + } +\cs_new:Npn \@@_change_case_if_greek_accent:n #1 + { + \if_int_compare:w #1 = "0300 \exp_stop_f: + \prg_return_true: + \else: + \if_int_compare:w #1 = "0301 \exp_stop_f: + \prg_return_true: + \else: + \if_int_compare:w #1 = "0342 \exp_stop_f: + \prg_return_true: + \else: + \if_int_compare:w #1 = "0302 \exp_stop_f: + \prg_return_true: + \else: + \if_int_compare:w #1 = "0303 \exp_stop_f: + \prg_return_true: + \else: + \if_int_compare:w #1 = "0311 \exp_stop_f: + \prg_return_true: + \else: + \prg_return_false: + \fi: + \fi: + \fi: + \fi: + \fi: + \fi: + } +\prg_new_conditional:Npnn \@@_change_case_if_greek_spacing_diacritic:n + #1 { TF } + { + \exp_args:Nf \@@_change_case_if_greek_spacing_diacritic:n + { \int_eval:n { \@@_codepoint_from_chars:Nw #1 } } + } +\cs_new:Npn \@@_change_case_if_greek_spacing_diacritic:n #1 + { + \if_int_compare:w #1 < "1FBD \exp_stop_f: + \if_int_compare:w #1 = "037A \exp_stop_f: + \prg_return_true: + \else: + \prg_return_false: + \fi: + \else: + \if_int_compare:w #1 = "1FBD \exp_stop_f: + \prg_return_true: + \else: + \if_int_compare:w #1 = "1FBF \exp_stop_f: + \prg_return_true: + \else: + \if_int_compare:w #1 = "1FC0 \exp_stop_f: + \prg_return_true: + \else: + \if_int_compare:w #1 = "1FC1 \exp_stop_f: + \prg_return_true: + \else: + \if_int_compare:w #1 = "1FCD \exp_stop_f: + \prg_return_true: + \else: + \if_int_compare:w #1 = "1FCE \exp_stop_f: + \prg_return_true: + \else: + \if_int_compare:w #1 = "1FCF \exp_stop_f: + \prg_return_true: + \else: + \if_int_compare:w #1 = "1FDD \exp_stop_f: + \prg_return_true: + \else: + \if_int_compare:w #1 = "1FDE \exp_stop_f: + \prg_return_true: + \else: + \if_int_compare:w #1 = "1FDF \exp_stop_f: + \prg_return_true: + \else: + \if_int_compare:w #1 = "1FED \exp_stop_f: + \prg_return_true: + \else: + \if_int_compare:w #1 = "1FEE \exp_stop_f: + \prg_return_true: + \else: + \if_int_compare:w #1 = "1FEF \exp_stop_f: + \prg_return_true: + \else: + \if_int_compare:w #1 = "1FFD \exp_stop_f: + \prg_return_true: + \else: + \if_int_compare:w #1 = "1FFE \exp_stop_f: + \prg_return_true: + \else: + \prg_return_false: + \fi: + \fi: + \fi: + \fi: + \fi: + \fi: + \fi: + \fi: + \fi: + \fi: + \fi: + \fi: + \fi: + \fi: + \fi: + \fi: + } +\prg_new_conditional:Npnn \@@_change_case_if_greek_breathing:n + #1 { TF , p } + { + \exp_args:Nf \@@_change_case_if_greek_breathing:n + { \int_eval:n { \@@_codepoint_from_chars:Nw #1 } } + } +\cs_new:Npn \@@_change_case_if_greek_breathing:n #1 + { + \if_int_compare:w #1 = "0313 \exp_stop_f: + \prg_return_true: + \else: + \if_int_compare:w #1 = "0314 \exp_stop_f: + \prg_return_true: + \else: + \prg_return_false: + \fi: + \fi: + } +\prg_new_conditional:Npnn \@@_change_case_if_greek_stress:n + #1 { TF , p } + { + \exp_args:Nf \@@_change_case_if_greek_stress:n + { \int_eval:n { \@@_codepoint_from_chars:Nw #1 } } + } +\cs_new:Npn \@@_change_case_if_greek_stress:n #1 + { + \if_int_compare:w #1 = "0304 \exp_stop_f: + \prg_return_true: + \else: + \if_int_compare:w #1 = "0306 \exp_stop_f: + \prg_return_true: + \else: + \prg_return_false: + \fi: + \fi: + } +\prg_new_conditional:Npnn \@@_change_case_if_takes_dialytika:n #1 { TF } + { + \exp_args:Nf \@@_change_case_if_takes_dialytika:n + { \int_eval:n { \@@_codepoint_from_chars:Nw #1 } } + } +\cs_new:Npn \@@_change_case_if_takes_dialytika:n #1 + { + \if_int_compare:w #1 = "0399 \exp_stop_f: + \prg_return_true: + \else: + \if_int_compare:w #1 = "03B9 \exp_stop_f: + \prg_return_true: + \else: + \if_int_compare:w #1 = "03A5 \exp_stop_f: + \prg_return_true: + \else: + \if_int_compare:w #1 = "03C5 \exp_stop_f: + \prg_return_true: + \else: + \prg_return_false: + \fi: + \fi: + \fi: + \fi: + } +\prg_new_conditional:Npnn \@@_change_case_if_takes_ypogegrammeni:n #1 { TF } + { + \exp_args:Nf \@@_change_case_if_takes_ypogegrammeni:n + { \int_eval:n { \@@_codepoint_from_chars:Nw #1 } } + } +\cs_new:Npn \@@_change_case_if_takes_ypogegrammeni:n #1 + { + \if_int_compare:w #1 = "03B1 \exp_stop_f: + \prg_return_true: + \else: + \if_int_compare:w #1 = "03B7 \exp_stop_f: + \prg_return_true: + \else: + \if_int_compare:w #1 = "03C9 \exp_stop_f: + \prg_return_true: + \else: + \prg_return_false: + \fi: + \fi: + \fi: + } +% \end{macrocode} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \begin{macro}[EXP] +% { +% \@@_change_case_boundary_upper_el:Nnnnw, +% \@@_change_case_boundary_upper_el-x-iota:Nnnnw +% } +% \begin{macro}[EXP]{\@@_change_case_boundary_upper_el:nnnN} +% \begin{macro}[EXP]{\@@_change_case_boundary_upper_el:nnnn} +% \begin{macro}[EXP]{\@@_change_case_boundary_upper_el:nnnnw} +% There is one things that need special treatment at the start of +% words in Greek. For an isolated accent \emph{eta}, +% which is handled by seeing if we have exactly one of the affected +% codepoints followed by a space or brace group. +% \begin{macrocode} +\cs_new:Npn \@@_change_case_boundary_upper_el:Nnnnw + #1#2#3#4#5 \q_@@_recursion_stop + { + \tl_if_head_is_N_type:nTF {#5} + { \@@_change_case_boundary_upper_el:nnnN } + { \@@_change_case_loop:nnnw } + {#2} {#3} {#4} #5 \q_@@_recursion_stop + } +\cs_new_eq:cN { @@_change_case_boundary_upper_el-x-iota:Nnnnw } + \@@_change_case_boundary_upper_el:Nnnnw +\cs_new:Npn \@@_change_case_boundary_upper_el:nnnN #1#2#3#4 + { + \token_if_cs:NTF #4 + { \@@_change_case_loop:nnnw {#1} {#2} {#3} } + { + \@@_codepoint_process:nN + { \@@_change_case_boundary_upper_el:nnnn {#1} {#2} {#3} } + } + #4 + } +\cs_new:Npn \@@_change_case_boundary_upper_el:nnnn #1#2#3#4 + { + \bool_lazy_any:nTF + { + { \@@_codepoint_compare_p:nNn {#4} = { "0389 } } + { \@@_codepoint_compare_p:nNn {#4} = { "03AE } } + { \@@_codepoint_compare_p:nNn {#4} = { "1F22 } } + { \@@_codepoint_compare_p:nNn {#4} = { "1F2A } } + } + { \@@_change_case_boundary_upper_el:nnnnw {#1} {#2} {#3} {#4} } + { \@@_change_case_breathing:nnnn {#1} {#2} {#3} {#4} } + } +\cs_new:Npn \@@_change_case_boundary_upper_el:nnnnw + #1#2#3#4#5 \q_@@_recursion_stop + { + \tl_if_head_is_N_type:nTF {#5} + { \@@_change_case_loop:nnnw {#1} {#2} {#3} #4 } + { + \@@_change_case_store:e + { + \codepoint_generate:nn { "0389 } + { \@@_change_case_catcode:nn {#4} { "0389 } } + } + \@@_change_case_loop:nnnw {#1} {#2} {#3} + } + #5 \q_@@_recursion_stop + } +% \end{macrocode} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \begin{macro}[EXP]{\@@_change_case_breathing:nnnn} +% \begin{macro}[EXP]{\@@_change_case_breathing:nnnnn} +% \begin{macro}[EXP]{\@@_change_case_breathing:nnnnnw} +% \begin{macro}[EXP]{\@@_change_case_breathing:nnnnnnw} +% \begin{macro}[EXP]{\@@_change_case_breathing_aux:nnnnnn} +% \begin{macro}[EXP]{\@@_change_case_breathing_aux:nnnnw} +% \begin{macro}[EXP]{\@@_change_case_breathing_aux:nnnN} +% \begin{macro}[EXP]{\@@_change_case_breathing_dialytika:nnnn} +% In Greek, breathing diacritics are normally dropped when uppercasing: +% see the code for the general case. However, for the first character +% of a word, if there is a breather \emph{and} the next character takes +% a \emph{dialytika}, it needs to be added. We start by checking if +% the current codepoint is in the Greek range, then decomposing. +% \begin{macrocode} +\cs_new:Npn \@@_change_case_breathing:nnnn #1#2#3#4 + { + \@@_change_case_if_greek:nTF {#4} + { + \exp_args:Ne \@@_change_case_breathing:nnnnn + { + \codepoint_to_nfd:n + { \@@_codepoint_from_chars:Nw #4 } + } + {#1} {#2} {#3} {#4} + } + { \@@_change_case_loop:nnnw {#1} {#2} {#3} #4 } + } +\cs_new:Npn \@@_change_case_breathing:nnnnn #1#2#3#4#5 + { + \@@_codepoint_process:nN + { \@@_change_case_breathing:nnnnnw {#2} {#3} {#4} {#5} } + #1 \q_mark + } +% \end{macrocode} +% Normal form decomposition will always give between one and three +% codepoints. Luckily, the two breathing marks (\emph{psili} and +% \emph{dasia}) will be in a predictable position: last. So we can +% quickly establish first that there was a change on decomposition, +% and second if the final resulting codepoint is one of the two we +% care about. +% \begin{macrocode} +\cs_new:Npn \@@_change_case_breathing:nnnnnw #1#2#3#4#5#6 \q_mark + { + \tl_if_blank:nTF {#6} + { \@@_change_case_loop:nnnw {#1} {#2} {#3} #4 } + { + \@@_codepoint_process:nN + { \@@_change_case_breathing:nnnnnnw {#1} {#2} {#3} {#4} {#5} } + #6 \q_mark + } + } +\cs_new:Npn \@@_change_case_breathing:nnnnnnw #1#2#3#4#5#6#7 \q_mark + { + \tl_if_blank:nTF {#7} + { + \@@_change_case_breathing_aux:nnnnnn + {#1} {#2} {#3} {#4} {#5} {#6} + } + { + \@@_codepoint_process:nN + { \@@_change_case_breathing:nnnnnnw {#1} {#2} {#3} {#4} {#5} } + #7 \q_mark + } + } +\cs_new:Npn \@@_change_case_breathing_aux:nnnnnn #1#2#3#4#5#6 + { + \bool_lazy_or:nnTF + { \@@_codepoint_compare_p:nNn {#6} = { "0313 } } + { \@@_codepoint_compare_p:nNn {#6} = { "0314 } } + { \@@_change_case_breathing_aux:nnnnw {#1} {#2} {#3} {#5} } + { \@@_change_case_loop:nnnw {#1} {#2} {#3} #4 } + } +% \end{macrocode} +% Now the lookahead can be fired: check the next codepoint and assess +% whether it takes a \emph{dialytika}. Drop the +% breathing mark or generate the \emph{dialytika}: the +% latter is code shared with the general mechanism. +% \begin{macrocode} +\cs_new:Npn \@@_change_case_breathing_aux:nnnnw #1#2#3#4#5 + \q_@@_recursion_stop + { + \@@_change_case_store:e + { \@@_change_case_codepoint:nn { upper } {#4} } + \tl_if_head_is_N_type:nTF {#5} + { \@@_change_case_breathing_aux:nnnN } + { \@@_change_case_loop:nnnw } + {#1} {#2} {#3} #5 \q_@@_recursion_stop + } +\cs_new:Npn \@@_change_case_breathing_aux:nnnN #1#2#3#4 + { + \@@_codepoint_process:nN + { \@@_change_case_breathing_dialytika:nnnn {#1} {#2} {#3} } #4 + } +\cs_new:Npn \@@_change_case_breathing_dialytika:nnnn #1#2#3#4 + { + \@@_change_case_if_takes_dialytika:nTF {#4} + { + \@@_change_case_upper_el_dialytika:n {#4} + \@@_change_case_loop:nnnw {#1} {#2} {#3} + } + { \@@_change_case_loop:nnnw {#1} {#2} {#3} #4 } + } +% \end{macrocode} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \begin{macro}[EXP]{\@@_change_case_title_el:nnnnn} +% Titlecasing retains accents, but to prevent the uppercasing code +% from kicking in, there has to be an explicit function here. +% \begin{macrocode} +\cs_new:Npn \@@_change_case_title_el:nnnnn #1#2#3#4#5 + { \@@_change_case_codepoint:nnnnn {#1} {#2} {#3} {#4} {#5} } +% \end{macrocode} +% \end{macro} +% +% \begin{macro}[EXP] +% { +% \@@_change_case_upper_hy:nnnnn , +% \@@_change_case_title_hy:nnnnn , +% \@@_change_case_upper_hy-x-yiwn:nnnnn , +% \@@_change_case_title_hy-x-yiwn:nnnnn +% } +% See \url{https://www.unicode.org/L2/L2020/20143-armenian-ech-yiwn.pdf}. +% \begin{macrocode} +\cs_new:Npn \@@_change_case_upper_hy:nnnnn #1#2#3#4#5 + { + \@@_codepoint_compare:nNnTF {#5} = { "0587 } + { + \@@_change_case_store:e + { + \codepoint_generate:nn { "0535 } + { \@@_change_case_catcode:nn {#5} { "0535 } } + \codepoint_generate:nn { "054E } + { \@@_change_case_catcode:nn {#5} { "054E } } + } + \use:c { @@_change_case_next_ #2 :nnn } + {#2} {#3} {#4} + } + { \@@_change_case_codepoint:nnnnn {#1} {#2} {#3} {#4} {#5} } + } +\cs_new:Npn \@@_change_case_title_hy:nnnnn #1#2#3#4#5 + { + \@@_codepoint_compare:nNnTF {#5} = { "0587 } + { + \@@_change_case_store:e + { + \codepoint_generate:nn { "0535 } + { \@@_change_case_catcode:nn {#5} { "0535 } } + \codepoint_generate:nn { "057E } + { \@@_change_case_catcode:nn {#5} { "057E } } + } + \use:c { @@_change_case_next_ #2 :nnn } + {#2} {#3} {#4} + } + { \@@_change_case_codepoint:nnnnn {#1} {#2} {#3} {#4} {#5} } + } +\cs_new:cpn { @@_change_case_upper_hy-x-yiwn:nnnnn } #1#2#3#4#5 + { \@@_change_case_codepoint:nnnnn {#1} {#2} {#3} {#4} {#5} } +\cs_new_eq:cc { @@_change_case_title_hy-x-yiwn:nnnnn } + { @@_change_case_upper_hy-x-yiwn:nnnnn } +% \end{macrocode} +% \end{macro} +% +% \begin{macro}[EXP]{\@@_change_case_lower_la-x-medieval:nnnnn} +% \begin{macro}[EXP]{\@@_change_case_upper_la-x-medieval:nnnnn} +% Simply swaps of characters. +% \begin{macrocode} +\cs_new:cpn { @@_change_case_lower_la-x-medieval:nnnnn } #1#2#3#4#5 + { + \@@_codepoint_compare:nNnTF {#5} = { "0056 } + { + \@@_change_case_store:e + { \char_generate:nn { "0075 } { \@@_char_catcode:N #5 } } + \use:c { @@_change_case_next_ #2 :nnn } + {#2} {#3} {#4} + } + { \@@_change_case_codepoint:nnnnn {#1} {#2} {#3} {#4} {#5} } + } +\cs_new:cpn { @@_change_case_upper_la-x-medieval:nnnnn } #1#2#3#4#5 + { + \@@_codepoint_compare:nNnTF {#5} = { "0075 } + { + \@@_change_case_store:e + { \char_generate:nn { "0056 } { \@@_char_catcode:N #5 } } + \use:c { @@_change_case_next_ #2 :nnn } + {#2} {#3} {#4} + } + { \@@_change_case_codepoint:nnnnn {#1} {#2} {#3} {#4} {#5} } + } +% \end{macrocode} +% \end{macro} +% \end{macro} +% +% \begin{macro}[EXP] +% { +% \@@_change_cases_lower_lt:nnnnn , +% \@@_change_cases_lower_lt_auxi:nnnnn , +% \@@_change_cases_lower_lt_auxii:nnnnn +% } +% \begin{macro}[rEXP]{\@@_change_case_lower_lt:nnnw} +% \begin{macro}[rEXP]{\@@_change_case_lower_lt:nnnN} +% \begin{macro}[rEXP]{\@@_change_case_lower_lt:nnnn} +% For Lithuanian, the issue to be dealt with is dots over lower case +% letters: these should be present if there is another accent. The first step +% is a simple match attempt: look for the three uppercase accented letters +% which should gain a dot-above char in their lowercase form. +% \begin{macrocode} +\cs_new:Npn \@@_change_case_lower_lt:nnnnn #1#2#3#4#5 + { + \exp_args:Ne \@@_change_case_lower_lt_auxi:nnnnn + { + \int_case:nn { \@@_codepoint_from_chars:Nw #5 } + { + { "00CC } { "0300 } + { "00CD } { "0301 } + { "0128 } { "0303 } + } + } + {#2} {#3} {#4} {#5} + } +% \end{macrocode} +% If there was a hit, output the result with the dot-above and move on. +% Otherwise, look for one of the three letters that can take a combining +% accent: I, J nd I-ogonek. +% \begin{macrocode} +\cs_new:Npn \@@_change_case_lower_lt_auxi:nnnnn #1#2#3#4#5 + { + \tl_if_blank:nTF {#1} + { + \exp_args:Ne \@@_change_case_lower_lt_auxii:nnnnn + { + \int_case:nn { \@@_codepoint_from_chars:Nw #5 } + { + { "0049 } { "0069 } + { "004A } { "006A } + { "012E } { "012F } + } + } + {#2} {#3} {#4} {#5} + } + { + \@@_change_case_store:e + { + \codepoint_generate:nn { "0069 } + { \@@_change_case_catcode:nn {#5} { "0069 } } + \codepoint_generate:nn { "0307 } + { \@@_change_case_catcode:nn {#5} { "0307 } } + \codepoint_generate:nn {#1} + { \@@_change_case_catcode:nn {#5} {#1} } + } + \@@_change_case_loop:nnnw {#2} {#3} {#4} + } + } +% \end{macrocode} +% Again, branch depending on a hit. If there is one, we output the character +% then need to look for a combining accent: as usual, we need to be aware of +% the loop situation. +% \begin{macrocode} +\cs_new:Npn \@@_change_case_lower_lt_auxii:nnnnn #1#2#3#4#5 + { + \tl_if_blank:nTF {#1} + { \@@_change_case_codepoint:nnnnn {#2} {#2} {#3} {#4} {#5} } + { + \@@_change_case_store:e + { + \codepoint_generate:nn {#1} + { \@@_change_case_catcode:nn {#5} {#1} } + } + \@@_change_case_lower_lt:nnnw {#2} {#3} {#4} + } + } +\cs_new:Npn \@@_change_case_lower_lt:nnnw #1#2#3#4 \q_@@_recursion_stop + { + \tl_if_head_is_N_type:nTF {#4} + { \@@_change_case_lower_lt:nnnN } + { \@@_change_case_loop:nnnw } + {#1} {#2} {#3} #4 \q_@@_recursion_stop + } +\cs_new:Npn \@@_change_case_lower_lt:nnnN #1#2#3#4 + { + \@@_codepoint_process:nN + { \@@_change_case_lower_lt:nnnn {#1} {#2} {#3} } #4 + } +\cs_new:Npn \@@_change_case_lower_lt:nnnn #1#2#3#4 + { + \bool_lazy_and:nnT + { + \bool_lazy_or_p:nn + { ! \tl_if_single_p:n {#4} } + { ! \token_if_cs_p:N #4 } + } + { + \bool_lazy_any_p:n + { + { \@@_codepoint_compare_p:nNn {#4} = { "0300 } } + { \@@_codepoint_compare_p:nNn {#4} = { "0301 } } + { \@@_codepoint_compare_p:nNn {#4} = { "0303 } } + } + } + { + \@@_change_case_store:e + { + \codepoint_generate:nn { "0307 } + { \@@_change_case_catcode:nn {#4} { "0307 } } + } + } + \@@_change_case_loop:nnnw {#1} {#2} {#3} #4 + } +% \end{macrocode} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \begin{macro}[EXP] +% { +% \@@_change_cases_upper_lt:nnnnn , +% \@@_change_cases_upper_lt_aux:nnnnn +% } +% \begin{macro}[rEXP]{\@@_change_case_upper_lt:nnnw} +% \begin{macro}[rEXP]{\@@_change_case_upper_lt:nnnN} +% \begin{macro}[rEXP]{\@@_change_case_upper_lt:nnnn} +% The uppercasing version: first find i/j/i-ogonek, then look for the +% combining char: drop it if present. +% \begin{macrocode} +\cs_new:Npn \@@_change_case_upper_lt:nnnnn #1#2#3#4#5 + { + \exp_args:Ne \@@_change_case_upper_lt_aux:nnnnn + { + \int_case:nn { \@@_codepoint_from_chars:Nw #5 } + { + { "0069 } { "0049 } + { "006A } { "004A } + { "012F } { "012E } + } + } + {#2} {#3} {#4} {#5} + } +\cs_new:Npn \@@_change_case_upper_lt_aux:nnnnn #1#2#3#4#5 + { + \tl_if_blank:nTF {#1} + { \@@_change_case_codepoint:nnnnn { upper } {#2} {#3} {#4} {#5} } + { + \@@_change_case_store:e + { + \codepoint_generate:nn {#1} + { \@@_change_case_catcode:nn {#5} {#1} } + } + \@@_change_case_upper_lt:nnnw {#2} {#3} {#4} + } + } +\cs_new:Npn \@@_change_case_upper_lt:nnnw #1#2#3#4 \q_@@_recursion_stop + { + \tl_if_head_is_N_type:nTF {#4} + { \@@_change_case_upper_lt:nnnN } + { \use:c { @@_change_case_next_ #1 :nnn } } + {#1} {#2} {#3} #4 \q_@@_recursion_stop + } +\cs_new:Npn \@@_change_case_upper_lt:nnnN #1#2#3#4 + { + \@@_codepoint_process:nN + { \@@_change_case_upper_lt:nnnn {#1} {#2} {#3} } #4 + } +\cs_new:Npn \@@_change_case_upper_lt:nnnn #1#2#3#4 + { + \bool_lazy_and:nnTF + { + \bool_lazy_or_p:nn + { ! \tl_if_single_p:n {#4} } + { ! \token_if_cs_p:N #4 } + } + { \@@_codepoint_compare_p:nNn {#4} = { "0307 } } + { \use:c { @@_change_case_next_ #1 :nnn } {#1} {#2} {#3} } + { \use:c { @@_change_case_next_ #1 :nnn } {#1} {#2} {#3} #4 } + } +% \end{macrocode} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% +% \begin{macro}[EXP] +% {\@@_change_case_title_nl:nnnnn, \@@_change_case_title_nl_aux:nnnnn} +% \begin{macro}[EXP]{\@@_change_case_title_nl:nnnw} +% \begin{macro}[EXP]{\@@_change_case_title_nl:nnnN} +% For Dutch, there is a single look-ahead test for \texttt{ij} when +% title casing. If the appropriate letters are found, produce \texttt{IJ} +% and gobble the \texttt{j}/\texttt{J}. +% \begin{macrocode} +\cs_new:Npn \@@_change_case_title_nl:nnnnn #1#2#3#4#5 + { + \tl_if_single:nTF {#5} + { \@@_change_case_title_nl_aux:nnnnn } + { \@@_change_case_codepoint:nnnnn } + {#1} {#2} {#3} {#4} {#5} + } +\cs_new:Npn \@@_change_case_title_nl_aux:nnnnn #1#2#3#4#5 + { + \bool_lazy_or:nnTF + { \int_compare_p:nNn {`#5} = { "0049 } } + { \int_compare_p:nNn {`#5} = { "0069 } } + { + \@@_change_case_store:e + { \char_generate:nn { "0049 } { \@@_char_catcode:N #5 } } + \@@_change_case_title_nl:nnnw {#2} {#3} {#4} + } + { \@@_change_case_codepoint:nnnnn {#1} {#2} {#3} {#4} {#5} } + } +\cs_new:Npn \@@_change_case_title_nl:nnnw #1#2#3#4 \q_@@_recursion_stop + { + \tl_if_head_is_N_type:nTF {#4} + { \@@_change_case_title_nl:nnnN } + { \use:c { @@_change_case_next_ #1 :nnn } } + {#1} {#2} {#3} #4 \q_@@_recursion_stop + } +\cs_new:Npn \@@_change_case_title_nl:nnnN #1#2#3#4 + { + \bool_lazy_and:nnTF + { ! \token_if_cs_p:N #4 } + { + \bool_lazy_or_p:nn + { \int_compare_p:nNn {`#4} = { "004A } } + { \int_compare_p:nNn {`#4} = { "006A } } + } + { + \@@_change_case_store:e + { \char_generate:nn { "004A } { \@@_char_catcode:N #4 } } + \use:c { @@_change_case_next_ #1 :nnn } {#1} {#2} {#3} + } + { \use:c { @@_change_case_next_ #1 :nnn } {#1} {#2} {#3} #4 } + } +% \end{macrocode} +% \end{macro} +% \end{macro} +% \end{macro} +% +% \begin{macro}[EXP]{\@@_change_case_lower_tr:nnnnn} +% \begin{macro}[EXP]{\@@_change_case_lower_tr:nnnNw} +% \begin{macro}[EXP]{\@@_change_case_lower_tr:NnnnN} +% \begin{macro}[EXP]{\@@_change_case_lower_tr:Nnnnn} +% The Turkic languages need special treatment for dotted-i and dotless-i. +% The lower casing rule can be expressed in terms of searching first for +% either a dotless-I or a dotted-I. In the latter case the mapping is +% easy, but in the former there is a second stage search. +% \begin{macrocode} +\cs_new:Npn \@@_change_case_lower_tr:nnnnn #1#2#3#4#5 + { + \@@_codepoint_compare:nNnTF {#5} = { "0049 } + { \@@_change_case_lower_tr:nnnNw {#1} {#3} {#4} #5 } + { + \@@_codepoint_compare:nNnTF {#5} = { "0130 } + { + \@@_change_case_store:e + { + \codepoint_generate:nn { "0069 } + { \@@_change_case_catcode:nn {#5} { "0069 } } + } + \@@_change_case_loop:nnnw {#1} {#3} {#4} + } + { \@@_change_case_codepoint:nnnnn {#1} {#2} {#3} {#4} {#5} } + } + } +% \end{macrocode} +% After a dotless-I there may be a dot-above character. If there is then +% a dotted-i should be produced, otherwise output a dotless-i. When the +% combination is found both the dotless-I and the dot-above char have to +% be removed from the input. +% \begin{macrocode} +\cs_new:Npn \@@_change_case_lower_tr:nnnNw #1#2#3#4#5 \q_@@_recursion_stop + { + \tl_if_head_is_N_type:nTF {#5} + { \@@_change_case_lower_tr:NnnnN #4 {#1} {#2} {#3} } + { + \@@_change_case_store:e + { + \codepoint_generate:nn { "0131 } + { \@@_change_case_catcode:nn {#4} { "0131 } } + } + \@@_change_case_loop:nnnw {#1} {#2} {#3} + } + #5 \q_@@_recursion_stop + } +\cs_new:Npn \@@_change_case_lower_tr:NnnnN #1#2#3#4#5 + { + \@@_codepoint_process:nN + { \@@_change_case_lower_tr:Nnnnn #1 {#2} {#3} {#4} } #5 + } +\cs_new:Npn \@@_change_case_lower_tr:Nnnnn #1#2#3#4#5 + { + \bool_lazy_or:nnTF + { + \bool_lazy_and_p:nn + { \tl_if_single_p:n {#5} } + { \token_if_cs_p:N #5 } + } + { ! \@@_codepoint_compare_p:nNn {#5} = { "0307 } } + { + \@@_change_case_store:e + { + \codepoint_generate:nn { "0131 } + { \@@_change_case_catcode:nn {#1} { "0131 } } + } + \@@_change_case_loop:nnnw {#2} {#3} {#4} #5 + } + { + \@@_change_case_store:e + { + \codepoint_generate:nn { "0069 } + { \@@_change_case_catcode:nn {#1} { "0069 } } + } + \@@_change_case_loop:nnnw {#2} {#3} {#4} + } + } +% \end{macrocode} +% \end{macro} +% \end{macro} +% \end{macro} +% \end{macro} +% \begin{macro}[EXP]{\@@_change_case_upper_tr:nnnnn} +% Uppercasing is easier: just one exception with no context. +% \begin{macrocode} +\cs_new:Npn \@@_change_case_upper_tr:nnnnn #1#2#3#4#5 + { + \@@_codepoint_compare:nNnTF {#5} = { "0069 } + { + \@@_change_case_store:e + { + \codepoint_generate:nn { "0130 } + { \@@_change_case_catcode:nn {#5} { "0130 } } + } + \use:c { @@_change_case_next_ #2 :nnn } {#2} {#3} {#4} + } + { \@@_change_case_codepoint:nnnnn {#1} {#2} {#3} {#4} {#5} } + } +% \end{macrocode} +% \end{macro} +% +% \begin{macro}[EXP] +% {\@@_change_case_lower_az:nnnnn, \@@_change_case_upper_az:nnnnn} +% Straight copies. +% \begin{macrocode} +\cs_new_eq:NN \@@_change_case_lower_az:nnnnn + \@@_change_case_lower_tr:nnnnn +\cs_new_eq:NN \@@_change_case_upper_az:nnnnn + \@@_change_case_upper_tr:nnnnn +% \end{macrocode} +% \end{macro} +% +% The (fixed) look-up mappings for letter-like control sequences. +% \begin{macrocode} +\group_begin: + \cs_set_protected:Npn \@@_change_case_setup:NN #1#2 + { + \quark_if_recursion_tail_stop:N #1 + \tl_const:cn { c_@@_lowercase_ \token_to_str:N #1 _tl } + { #2 } + \tl_const:cn { c_@@_uppercase_ \token_to_str:N #2 _tl } + { #1 } + \@@_change_case_setup:NN + } + \@@_change_case_setup:NN + \AA \aa + \AE \ae + \DH \dh + \DJ \dj + \IJ \ij + \L \l + \NG \ng + \O \o + \OE \oe + \SS \ss + \TH \th + \q_recursion_tail ? + \q_recursion_stop + \tl_const:cn { c_@@_uppercase_ \token_to_str:N \i _tl } { I } + \tl_const:cn { c_@@_uppercase_ \token_to_str:N \j _tl } { J } +\group_end: +% \end{macrocode} +% +% To deal with possible encoding-specific extensions to \tn{@uclclist}, +% we check at the end of the preamble. This will therefore only apply +% to \LaTeXe{} package mode. +% \begin{macrocode} +\tl_if_exist:NT \@expl@finalise@setup@@@@ + { + \tl_gput_right:Nn \@expl@finalise@setup@@@@ + { + \tl_gput_right:Nn \@kernel@after@begindocument + { + \group_begin: + \cs_set_protected:Npn \@@_change_case_setup:Nn #1#2 + { + \quark_if_recursion_tail_stop:N #1 + \tl_if_single_token:nT {#2} + { + \cs_if_exist:cF + { c_@@_uppercase_ \token_to_str:N #1 _tl } + { + \tl_const:cn + { c_@@_uppercase_ \token_to_str:N #1 _tl } + { #2 } + } + \cs_if_exist:cF + { c_@@_lowercase_ \token_to_str:N #2 _tl } + { + \tl_const:cn + { c_@@_lowercase_ \token_to_str:N #2 _tl } + { #1 } + } + } + \@@_change_case_setup:Nn + } + \exp_after:wN \@@_change_case_setup:Nn \@uclclist + \q_recursion_tail ? + \q_recursion_stop + \group_end: + } + } + } +% \end{macrocode} +% +% A few adjustments to case mapping for combining chars: these are not needed +% for the Unicode engines +% \begin{macrocode} +\bool_lazy_or:nnF + { \sys_if_engine_luatex_p: } + { \sys_if_engine_xetex_p: } + { + \text_declare_uppercase_mapping:nn { "01F0 } { \v { J } } + } +% \end{macrocode} +% +% \begin{macrocode} +%</package> +% \end{macrocode} +% +% \end{implementation} +% +% \PrintIndex |