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
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
|
% \iffalse meta-comment
%
%% File: l3backend-box.dtx
%
% Copyright (C) 2019-2024 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 "l3backend 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{l3backend-box} package\\ Backend box support^^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 2024-05-08}
%
% \maketitle
%
% \begin{documentation}
%
% \end{documentation}
%
% \begin{implementation}
%
% \section{\pkg{l3backend-box} implementation}
%
% \begin{macrocode}
%<*package>
%<@@=box>
% \end{macrocode}
%
% \subsection{\texttt{dvips} backend}
%
% \begin{macrocode}
%<*dvips>
% \end{macrocode}
%
% \begin{macro}{\@@_backend_clip:N}
% The \texttt{dvips} backend scales all absolute dimensions based on the
% output resolution selected and any \TeX{} magnification. Thus for any
% operation involving absolute lengths there is a correction to make. See
% \texttt{normalscale} from \texttt{special.pro} for the variables, noting
% that here everything is saved on the stack rather than as a separate
% variable. Once all of that is done, the actual clipping is trivial.
% \begin{macrocode}
\cs_new_protected:Npn \@@_backend_clip:N #1
{
\__kernel_backend_scope_begin:
\__kernel_backend_align_begin:
\__kernel_backend_literal_postscript:n { matrix~currentmatrix }
\__kernel_backend_literal_postscript:n
{ Resolution~72~div~VResolution~72~div~scale }
\__kernel_backend_literal_postscript:n { DVImag~dup~scale }
\__kernel_backend_literal_postscript:e
{
0 ~
\dim_to_decimal_in_bp:n { \box_dp:N #1 } ~
\dim_to_decimal_in_bp:n { \box_wd:N #1 } ~
\dim_to_decimal_in_bp:n { -\box_ht:N #1 - \box_dp:N #1 } ~
rectclip
}
\__kernel_backend_literal_postscript:n { setmatrix }
\__kernel_backend_align_end:
\hbox_overlap_right:n { \box_use:N #1 }
\__kernel_backend_scope_end:
\skip_horizontal:n { \box_wd:N #1 }
}
% \end{macrocode}
% \end{macro}
%
% \begin{macro}{\@@_backend_rotate:Nn}
% \begin{macro}{\@@_backend_rotate_aux:Nn}
% Rotating using \texttt{dvips} does not require that the box dimensions
% are altered and has a very convenient built-in operation. Zero rotation
% must be written as |0| not |-0| so there is a quick test.
% \begin{macrocode}
\cs_new_protected:Npn \@@_backend_rotate:Nn #1#2
{ \exp_args:NNf \@@_backend_rotate_aux:Nn #1 { \fp_eval:n {#2} } }
\cs_new_protected:Npn \@@_backend_rotate_aux:Nn #1#2
{
\__kernel_backend_scope_begin:
\__kernel_backend_align_begin:
\__kernel_backend_literal_postscript:e
{
\fp_compare:nNnTF {#2} = \c_zero_fp
{ 0 }
{ \fp_eval:n { round ( -(#2) , 5 ) } } ~
rotate
}
\__kernel_backend_align_end:
\box_use:N #1
\__kernel_backend_scope_end:
}
% \end{macrocode}
% \end{macro}
% \end{macro}
%
% \begin{macro}{\@@_backend_scale:Nnn}
% The \texttt{dvips} backend once again has a dedicated operation we can
% use here.
% \begin{macrocode}
\cs_new_protected:Npn \@@_backend_scale:Nnn #1#2#3
{
\__kernel_backend_scope_begin:
\__kernel_backend_align_begin:
\__kernel_backend_literal_postscript:e
{
\fp_eval:n { round ( #2 , 5 ) } ~
\fp_eval:n { round ( #3 , 5 ) } ~
scale
}
\__kernel_backend_align_end:
\hbox_overlap_right:n { \box_use:N #1 }
\__kernel_backend_scope_end:
}
% \end{macrocode}
% \end{macro}
%
% \begin{macrocode}
%</dvips>
% \end{macrocode}
%
% \subsection{\LuaTeX{} and \pdfTeX{} backends}
%
% \begin{macrocode}
%<*luatex|pdftex>
% \end{macrocode}
%
% \begin{macro}{\@@_backend_clip:N}
% The general method is to save the current location, define a clipping path
% equivalent to the bounding box, then insert the content at the current
% position and in a zero width box. The \enquote{real} width is then made up
% using a horizontal skip before tidying up. There are other approaches that
% can be taken (for example using XForm objects), but the logic here shares
% as much code as possible and uses the same conversions (and so same
% rounding errors) in all cases.
% \begin{macrocode}
\cs_new_protected:Npn \@@_backend_clip:N #1
{
\__kernel_backend_scope_begin:
\__kernel_backend_literal_pdf:e
{
0~
\dim_to_decimal_in_bp:n { -\box_dp:N #1 } ~
\dim_to_decimal_in_bp:n { \box_wd:N #1 } ~
\dim_to_decimal_in_bp:n { \box_ht:N #1 + \box_dp:N #1 } ~
re~W~n
}
\hbox_overlap_right:n { \box_use:N #1 }
\__kernel_backend_scope_end:
\skip_horizontal:n { \box_wd:N #1 }
}
% \end{macrocode}
% \end{macro}
%
% \begin{macro}{\@@_backend_rotate:Nn}
% \begin{macro}{\@@_backend_rotate_aux:Nn}
% \begin{variable}{\l_@@_backend_cos_fp, \l_@@_backend_sin_fp}
% Rotations are set using an affine transformation matrix which therefore
% requires sine/cosine values not the angle itself. We store the rounded
% values to avoid rounding twice. There are also a couple of comparisons to
% ensure that |-0| is not written to the output, as this avoids any issues
% with problematic display programs. Note that numbers are compared to~$0$
% after rounding.
% \begin{macrocode}
\cs_new_protected:Npn \@@_backend_rotate:Nn #1#2
{ \exp_args:NNf \@@_backend_rotate_aux:Nn #1 { \fp_eval:n {#2} } }
\cs_new_protected:Npn \@@_backend_rotate_aux:Nn #1#2
{
\__kernel_backend_scope_begin:
\box_set_wd:Nn #1 { 0pt }
\fp_set:Nn \l_@@_backend_cos_fp { round ( cosd ( #2 ) , 5 ) }
\fp_compare:nNnT \l_@@_backend_cos_fp = \c_zero_fp
{ \fp_zero:N \l_@@_backend_cos_fp }
\fp_set:Nn \l_@@_backend_sin_fp { round ( sind ( #2 ) , 5 ) }
\__kernel_backend_matrix:e
{
\fp_use:N \l_@@_backend_cos_fp \c_space_tl
\fp_compare:nNnTF \l_@@_backend_sin_fp = \c_zero_fp
{ 0~0 }
{
\fp_use:N \l_@@_backend_sin_fp
\c_space_tl
\fp_eval:n { -\l_@@_backend_sin_fp }
}
\c_space_tl
\fp_use:N \l_@@_backend_cos_fp
}
\box_use:N #1
\__kernel_backend_scope_end:
}
\fp_new:N \l_@@_backend_cos_fp
\fp_new:N \l_@@_backend_sin_fp
% \end{macrocode}
% \end{variable}
% \end{macro}
% \end{macro}
%
% \begin{macro}{\@@_backend_scale:Nnn}
% The same idea as for rotation but without the complexity of signs and
% cosines.
% \begin{macrocode}
\cs_new_protected:Npn \@@_backend_scale:Nnn #1#2#3
{
\__kernel_backend_scope_begin:
\__kernel_backend_matrix:e
{
\fp_eval:n { round ( #2 , 5 ) } ~
0~0~
\fp_eval:n { round ( #3 , 5 ) }
}
\hbox_overlap_right:n { \box_use:N #1 }
\__kernel_backend_scope_end:
}
% \end{macrocode}
% \end{macro}
%
% \begin{macrocode}
%</luatex|pdftex>
% \end{macrocode}
%
% \subsection{\texttt{dvipdfmx}/\XeTeX{} backend}
%
% \begin{macrocode}
%<*dvipdfmx|xetex>
% \end{macrocode}
%
% \begin{macro}{\@@_backend_clip:N}
% The code here is identical to that for \LuaTeX{}/\pdfTeX{}: unlike rotation and
% scaling, there is no higher-level support in the backend for clipping.
% \begin{macrocode}
\cs_new_protected:Npn \@@_backend_clip:N #1
{
\__kernel_backend_scope_begin:
\__kernel_backend_literal_pdf:e
{
0~
\dim_to_decimal_in_bp:n { -\box_dp:N #1 } ~
\dim_to_decimal_in_bp:n { \box_wd:N #1 } ~
\dim_to_decimal_in_bp:n { \box_ht:N #1 + \box_dp:N #1 } ~
re~W~n
}
\hbox_overlap_right:n { \box_use:N #1 }
\__kernel_backend_scope_end:
\skip_horizontal:n { \box_wd:N #1 }
}
% \end{macrocode}
% \end{macro}
%
% \begin{macro}{\@@_backend_rotate:Nn}
% \begin{macro}{\@@_backend_rotate_aux:Nn}
% Rotating in \texttt{dvipdmfx}/\XeTeX{} can be implemented using either PDF or
% backend-specific code. The former approach however is not \enquote{aware}
% of the content of boxes: this means that any embedded links would not be
% adjusted by the rotation. As such, the backend-native approach is preferred:
% the code therefore is similar (though not identical) to the \texttt{dvips}
% version (notice the rotation angle here is positive). As for
% \texttt{dvips}, zero rotation is written as |0| not |-0|.
% \begin{macrocode}
\cs_new_protected:Npn \@@_backend_rotate:Nn #1#2
{ \exp_args:NNf \@@_backend_rotate_aux:Nn #1 { \fp_eval:n {#2} } }
\cs_new_protected:Npn \@@_backend_rotate_aux:Nn #1#2
{
\__kernel_backend_scope_begin:
\__kernel_backend_literal:e
{
x:rotate~
\fp_compare:nNnTF {#2} = \c_zero_fp
{ 0 }
{ \fp_eval:n { round ( #2 , 5 ) } }
}
\box_use:N #1
\__kernel_backend_scope_end:
}
% \end{macrocode}
% \end{macro}
% \end{macro}
%
% \begin{macro}{\@@_backend_scale:Nnn}
% Much the same idea for scaling: use the higher-level backend operation to allow
% for box content.
% \begin{macrocode}
\cs_new_protected:Npn \@@_backend_scale:Nnn #1#2#3
{
\__kernel_backend_scope_begin:
\__kernel_backend_literal:e
{
x:scale~
\fp_eval:n { round ( #2 , 5 ) } ~
\fp_eval:n { round ( #3 , 5 ) }
}
\hbox_overlap_right:n { \box_use:N #1 }
\__kernel_backend_scope_end:
}
% \end{macrocode}
% \end{macro}
%
% \begin{macrocode}
%</dvipdfmx|xetex>
% \end{macrocode}
%
% \subsection{\texttt{dvisvgm} backend}
%
% \begin{macrocode}
%<*dvisvgm>
% \end{macrocode}
%
% \begin{macro}{\@@_backend_clip:N}
% \begin{variable}{\g__kernel_clip_path_int}
% Clipping in SVG is more involved than with other backends. The first issue
% is that the clipping path must be defined separately from where it is used,
% so we need to track how many paths have applied. The naming here uses
% \texttt{l3cp} as the namespace with a number following. Rather than use
% a rectangular operation, we define the path manually as this allows it to
% have a depth: easier than the alternative approach of shifting content
% up and down using scopes to allow for the depth of the \TeX{} box and
% keep the reference point the same!
% \begin{macrocode}
\cs_new_protected:Npn \@@_backend_clip:N #1
{
\int_gincr:N \g__kernel_clip_path_int
\__kernel_backend_literal_svg:e
{ < clipPath~id = " l3cp \int_use:N \g__kernel_clip_path_int " > }
\__kernel_backend_literal_svg:e
{
<
path ~ d =
"
M ~ 0 ~
\dim_to_decimal:n { -\box_dp:N #1 } ~
L ~ \dim_to_decimal:n { \box_wd:N #1 } ~
\dim_to_decimal:n { -\box_dp:N #1 } ~
L ~ \dim_to_decimal:n { \box_wd:N #1 } ~
\dim_to_decimal:n { \box_ht:N #1 + \box_dp:N #1 } ~
L ~ 0 ~
\dim_to_decimal:n { \box_ht:N #1 + \box_dp:N #1 } ~
Z
"
/>
}
\__kernel_backend_literal_svg:n
{ < /clipPath > }
% \end{macrocode}
% In general the SVG set up does not try to transform coordinates to the
% current point. For clipping we need to do that, so have a transformation
% here to get us to the right place, and a matching one just before the
% \TeX{} box is inserted to get things back on track. The clip path needs to
% come between those two such that if lines up with the current point, as
% does the \TeX{} box.
% \begin{macrocode}
\__kernel_backend_scope_begin:n
{
transform =
"
translate ( { ?x } , { ?y } ) ~
scale ( 1 , -1 )
"
}
\__kernel_backend_scope:e
{
clip-path =
"url ( \c_hash_str l3cp \int_use:N \g__kernel_clip_path_int ) "
}
\__kernel_backend_scope:n
{
transform =
"
scale ( -1 , 1 ) ~
translate ( { ?x } , { ?y } ) ~
scale ( -1 , -1 )
"
}
\box_use:N #1
\__kernel_backend_scope_end:
}
\int_new:N \g__kernel_clip_path_int
% \end{macrocode}
% \end{variable}
% \end{macro}
%
% \begin{macro}{\@@_backend_rotate:Nn}
% Rotation has a dedicated operation which includes a centre-of-rotation
% optional pair. That can be picked up from the backend syntax, so there is
% no need to worry about the transformation matrix.
% \begin{macrocode}
\cs_new_protected:Npn \@@_backend_rotate:Nn #1#2
{
\__kernel_backend_scope_begin:e
{
transform =
"
rotate
( \fp_eval:n { round ( -(#2) , 5 ) } , ~ { ?x } , ~ { ?y } )
"
}
\box_use:N #1
\__kernel_backend_scope_end:
}
% \end{macrocode}
% \end{macro}
%
% \begin{macro}{\@@_backend_scale:Nnn}
% In contrast to rotation, we have to account for the current position in this
% case. That is done using a couple of translations in addition to the scaling
% (which is therefore done backward with a flip).
% \begin{macrocode}
\cs_new_protected:Npn \@@_backend_scale:Nnn #1#2#3
{
\__kernel_backend_scope_begin:e
{
transform =
"
translate ( { ?x } , { ?y } ) ~
scale
(
\fp_eval:n { round ( -#2 , 5 ) } ,
\fp_eval:n { round ( -#3 , 5 ) }
) ~
translate ( { ?x } , { ?y } ) ~
scale ( -1 )
"
}
\hbox_overlap_right:n { \box_use:N #1 }
\__kernel_backend_scope_end:
}
% \end{macrocode}
% \end{macro}
%
% \begin{macrocode}
%</dvisvgm>
% \end{macrocode}
%
% \begin{macrocode}
%</package>
% \end{macrocode}
%
% \end{implementation}
%
% \PrintIndex
|