summaryrefslogtreecommitdiff
path: root/Master/texmf-dist/source/latex/l3experimental/l3str/l3tl-build.dtx
blob: bbdb39586136ef0c12beacde4dfcb5870ce069b8 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
% \iffalse meta-comment
%
%% File: l3tl-build.dtx Copyright (C) 2011-2017 The LaTeX3 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
%
%    http://www.latex-project.org/lppl.txt
%
% This file is part of the "l3experimental 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|package>
% The version of expl3 required is tested as early as possible, as
% some really old versions do not define \ProvidesExplPackage.
\RequirePackage{expl3}[2017/05/13]
%<package>\@ifpackagelater{expl3}{2017/05/13}
%<package>  {}
%<package>  {%
%<package>    \PackageError{l3tl-build}{Support package l3kernel too old}
%<package>      {%
%<package>        Please install an up to date version of l3kernel\MessageBreak
%<package>        using your TeX package manager or from CTAN.\MessageBreak
%<package>        \MessageBreak
%<package>        Loading l3tl-build will abort!%
%<package>      }%
%<package>    \endinput
%<package>  }
%</driver|package>
%<*driver>
\documentclass[full]{l3doc}
\usepackage{amsmath}
\begin{document}
  \DocInput{\jobname.dtx}
\end{document}
%</driver>
% \fi
%
%
% \title{^^A
%   The \textsf{l3tl-build} package: building token lists^^A
% }
%
% \author{^^A
%  The \LaTeX3 Project\thanks
%    {^^A
%      E-mail:
%        \href{mailto:latex-team@latex-project.org}
%          {latex-team@latex-project.org}^^A
%    }^^A
% }
%
% \date{Released 2017/05/13}
%
% \maketitle
%
% \begin{documentation}
%
% \section{\pkg{l3tl-build} documentation}
%
% This module provides no user function: it is meant for kernel use
% only.
%
% There are two main ways of building token lists from individual
% tokens. Either in one go within an \texttt{x}-expanding assignment, or
% by repeatedly using \cs{tl_put_right:Nn}. The first method takes a
% linear time, but only allows expandable operations. The second method
% takes a time quadratic in the length of the token list, but allows
% expandable and non-expandable operations.
%
% The goal of this module is to provide functions to build a token list
% piece by piece in linear time, while allowing non-expandable
% operations. This is achieved by abusing \tn{toks}: adding some tokens
% to the token list is done by storing them in a free token register
% (time $O(1)$ for each such operation). Those token registers are only
% put together at the end, within an \texttt{x}-expanding assignment,
% which takes a linear time.\footnote{If we run out of token registers,
%   then the currently filled-up \tn{toks} are put together in a
%   temporary token list, and cleared, and we ultimately use
%   \cs{tl_put_right:Nx} to put those chunks together. Hence the true
%   asymptotic is quadratic, with a very small constant.}  Of course,
% all this must be done in a group: we can't go and clobber the values
% of legitimate \tn{toks} used by \LaTeXe{}.
%
% Since none of the current applications need the ability to insert
% material on the left of the token list, I have not implemented
% that. This could be done for instance by using odd-numbered \tn{toks}
% for the left part, and even-numbered \tn{toks} for the right part.
%
% \subsection{Internal functions}
%
% \begin{function}
%   {
%     \__tl_build:Nw, \__tl_gbuild:Nw,
%     \__tl_build_x:Nw, \__tl_gbuild_x:Nw
%   }
%   \begin{syntax}
%     \cs{__tl_build:Nw} \meta{tl~var} \texttt{\ldots{}}
%     \cs{__tl_build_one:n} \Arg{tokens_1} \texttt{\ldots{}}
%     \cs{__tl_build_one:n} \Arg{tokens_2} \texttt{\ldots{}}
%     \ldots{}
%     \cs{__tl_build_end:}
%   \end{syntax}
%   Defines the \meta{tl~var} to contain the contents of \meta{tokens1}
%   followed by \meta{tokens2}, \emph{etc.} This is built in such a way
%   to be more efficient than repeatedly using \cs{tl_put_right:Nn}. The
%   code in \enquote{\texttt{\ldots{}}} does not need to be
%   expandable. The commands \cs{__tl_build:Nw} and \cs{__tl_build_end:}
%   start and end a group. The assignment to the \meta{tl~var} occurs
%   just after the end of that group, using \cs{tl_set:Nn},
%   \cs{tl_gset:Nn}, \cs{tl_set:Nx}, or \cs{tl_gset:Nx}.
% \end{function}
%
% \begin{function}{\__tl_build_one:n, \__tl_build_one:o, \__tl_build_one:x}
%   \begin{syntax}
%     \cs{__tl_build_one:n} \Arg{tokens}
%   \end{syntax}
%   This function may only be used within the scope of a
%   \cs{__tl_build:Nw} function. It adds the \meta{tokens} on the
%   right of the current token list.
% \end{function}
%
% \begin{function}{\__tl_build_end:}
%   Ends the scope started by \cs{__tl_build:Nw}, and performs the
%   relevant assignment.
% \end{function}
%
% \end{documentation}
%
% \begin{implementation}
%
% \section{\pkg{l3tl-build} implementation}
%
%    \begin{macrocode}
%<*initex|package>
%    \end{macrocode}
%
%    \begin{macrocode}
%<@@=tl_build>
%    \end{macrocode}
%
%    \begin{macrocode}
\ProvidesExplPackage{l3tl-build}{2017/05/13}{}
  {L3 Experimental token list construction}
%    \end{macrocode}
%
% \subsection{Variables and helper functions}
%
% \begin{variable}{\l_@@_start_index_int, \l_@@_index_int}
%   Integers pointing to the starting index (currently always starts at
%   zero), and the current index. The corresponding \tn{toks} are
%   accessed directly by number.
%    \begin{macrocode}
\int_new:N \l_@@_start_index_int
\int_new:N \l_@@_index_int
%    \end{macrocode}
% \end{variable}
%
% \begin{variable}{\l_@@_result_tl}
%   The resulting token list is normally built in one go by unpacking
%   all \tn{toks} in some range. In the rare cases where there are too
%   many \cs{@@_one:n} commands, leading to the depletion of
%   registers, the contents of the current set of \tn{toks} is unpacked
%   into \cs{l_@@_result_tl}. This prevents overflow from
%   affecting the end-user (beyond an obvious performance hit).
%    \begin{macrocode}
\tl_new:N \l_@@_result_tl
%    \end{macrocode}
% \end{variable}
%
% \begin{macro}{\@@_unpack:}
% \begin{macro}[aux, EXP]{\@@_unpack_loop:w}
%   The various pieces of the token list are built in \tn{toks} from the
%   \texttt{start_index} (inclusive) to the (current) \texttt{index}
%   (excluded). Those \tn{toks} are unpacked and stored in order in the
%   \texttt{result} token list.  Optimizations would be possible here,
%   for instance, unpacking $10$ \tn{toks} at a time with a macro
%   expanding to |\the\toks#10...\the\toks#19|, but this should be kept
%   for much later.
%    \begin{macrocode}
\cs_new_protected:Npn \@@_unpack:
  {
    \tl_put_right:Nx \l_@@_result_tl
      {
        \exp_after:wN \@@_unpack_loop:w
          \int_use:N \l_@@_start_index_int ;
        \__prg_break_point:
      }
  }
\cs_new:Npn \@@_unpack_loop:w #1 ;
  {
    \if_int_compare:w #1 = \l_@@_index_int
      \exp_after:wN \__prg_break:
    \fi:
    \tex_the:D \tex_toks:D #1 \exp_stop_f:
    \exp_after:wN \@@_unpack_loop:w
      \int_use:N \__int_eval:w #1 + 1 ;
  }
%    \end{macrocode}
% \end{macro}
% \end{macro}
%
% \subsection{Building the token list}
%
% \begin{macro}
%   {
%     \@@:Nw  , \@@_x:Nw  ,
%     \__tl_gbuild:Nw , \__tl_gbuild_x:Nw
%   }
% \begin{macro}[aux]{\@@_aux:NNw}
%   Similar to what is done for coffins: redefine some command, here
%   \cs{@@_end_aux:n} to hold the relevant assignment (see
%   \cs{@@_end:} for details). Then initialize the start index and
%   the current index at zero, and empty the \texttt{result} token list.
%    \begin{macrocode}
\cs_new_protected:Npn \@@:Nw
  { \@@_aux:NNw \tl_set:Nn }
\cs_new_protected:Npn \@@_x:Nw
  { \@@_aux:NNw \tl_set:Nx }
\cs_new_protected:Npn \__tl_gbuild:Nw
  { \@@_aux:NNw \tl_gset:Nn }
\cs_new_protected:Npn \__tl_gbuild_x:Nw
  { \@@_aux:NNw \tl_gset:Nx }
\cs_new_protected:Npn \@@_aux:NNw #1#2
  {
    \group_begin:
      \cs_set:Npn \@@_end_assignment:n
        { \group_end: #1 #2 }
      \int_zero:N \l_@@_start_index_int
      \int_zero:N \l_@@_index_int
      \tl_clear:N \l_@@_result_tl
  }
%    \end{macrocode}
% \end{macro}
% \end{macro}
%
% \begin{macro}{\@@_end:}
% \begin{macro}[aux]{\@@_end_assignment:n}
%   When we are done building a token list, unpack all \tn{toks} into
%   the \texttt{result} token list, and expand this list before closing
%   the group. The \cs{@@_end_assignment:n} function is defined by
%   \cs{@@_aux:NNw} to end the group and hold the relevant
%   assignment. Its value outside is irrelevant, but just in case, we
%   set it to a function which would clean up the contents of
%   \cs{l_@@_result_tl}.
%    \begin{macrocode}
\cs_new_protected:Npn \@@_end:
  {
      \@@_unpack:
      \exp_args:No
    \@@_end_assignment:n \l_@@_result_tl
  }
\cs_new_eq:NN \@@_end_assignment:n \use_none:n
%    \end{macrocode}
% \end{macro}
% \end{macro}
%
% \begin{macro}{\@@_one:n, \@@_one:o, \@@_one:x}
%   Store the tokens in a free \tn{toks}, then move the pointer to the
%   next one. If we overflow, unpack the current \tn{toks}, and reset
%   the current index, preparing to fill more \tn{toks}.  This could be
%   optimized by avoiding to read |#1|, using \tn{afterassignment}.
%    \begin{macrocode}
\cs_new_protected:Npn \@@_one:n #1
  {
    \tex_toks:D \l_@@_index_int {#1}
    \int_incr:N \l_@@_index_int
    \if_int_compare:w \l_@@_index_int > \c_max_register_int
      \@@_unpack:
      \l_@@_index_int \l_@@_start_index_int
    \fi:
  }
\cs_new_protected:Npn \@@_one:o #1
  {
    \tex_toks:D \l_@@_index_int \exp_after:wN {#1}
    \int_incr:N \l_@@_index_int
    \if_int_compare:w \l_@@_index_int > \c_max_register_int
      \@@_unpack:
      \l_@@_index_int \l_@@_start_index_int
    \fi:
  }
\cs_new_protected:Npn \@@_one:x #1
  { \use:x { \@@_one:n {#1} } }
%    \end{macrocode}
% \end{macro}
%
%    \begin{macrocode}
%</initex|package>
%    \end{macrocode}
%
% \end{implementation}
%
% \PrintIndex