Debian Codegraph @

texmf-dist/doc/luatex/base/luatex-math.tex

raw
1 % language=us engine=luatex runpath=texruns:manuals/luatex
2
3 \environment luatex-style
4
5 \startcomponent luatex-math
6
7 \startchapter[reference=math,title={Math}]
8
9 \startsection[title={Traditional alongside \OPENTYPE}]
10
11 \topicindex {math}
12
13 The handling of mathematics in \LUATEX\ differs quite a bit from how \TEX82 (and
14 therefore \PDFTEX) handles math. First, \LUATEX\ adds primitives and extends some
15 others so that \UNICODE\ input can be used easily. Second, all of \TEX82's
16 internal special values (for example for operator spacing) have been made
17 accessible and changeable via control sequences. Third, there are extensions that
18 make it easier to use \OPENTYPE\ math fonts. And finally, there are some
19 extensions that have been proposed or considered in the past that are now added
20 to the engine.
21
22 \stopsection
23
24 \startsection[title={Unicode math characters}]
25
26 \topicindex {math+\UNICODE}
27 \topicindex {\UNICODE+math}
28
29 Character handling is now extended up to the full \UNICODE\ range (the \type {\U}
30 prefix), which is compatible with \XETEX.
31
32 The math primitives from \TEX\ are kept as they are, except for the ones that
33 convert from input to math commands: \type {mathcode}, and \type {delcode}. These
34 two now allow for a 21-bit character argument on the left hand side of the equals
35 sign.
36
37 Some of the new \LUATEX\ primitives read more than one separate value. This is
38 shown in the tables below by a plus sign.
39
40 The input for such primitives would look like this:
41
42 \starttyping
43 \def\overbrace{\Umathaccent 0 1 "23DE }
44 \stoptyping
45
46 The altered \TEX82 primitives are:
47
48 \starttabulate[|l|l|r|c|l|r|]
49 \DB primitive \BC min \BC max \BC \kern 2em \BC min \BC max \NC \NR
50 \TB
51 \NC \prm {mathcode} \NC 0 \NC 10FFFF \NC = \NC 0 \NC 8000 \NC \NR
52 \NC \prm {delcode} \NC 0 \NC 10FFFF \NC = \NC 0 \NC FFFFFF \NC \NR
53 \LL
54 \stoptabulate
55
56 The unaltered ones are:
57
58 \starttabulate[|l|l|r|]
59 \DB primitive \BC min \BC max \NC \NR
60 \TB
61 \NC \prm {mathchardef} \NC 0 \NC 8000 \NC \NR
62 \NC \prm {mathchar} \NC 0 \NC 7FFF \NC \NR
63 \NC \prm {mathaccent} \NC 0 \NC 7FFF \NC \NR
64 \NC \prm {delimiter} \NC 0 \NC 7FFFFFF \NC \NR
65 \NC \prm {radical} \NC 0 \NC 7FFFFFF \NC \NR
66 \LL
67 \stoptabulate
68
69 For practical reasons \prm {mathchardef} will silently accept values larger
70 that \type {0x8000} and interpret it as \lpr {Umathcharnumdef}. This is needed
71 to satisfy older macro packages.
72
73 The following new primitives are compatible with \XETEX:
74
75 % somewhat fuzzy:
76
77 \starttabulate[|l|l|r|c|l|r|]
78 \DB primitive \BC min \BC max \BC \kern 2em \BC min \BC max \NC \NR
79 \TB
80 \NC \lpr {Umathchardef} \NC 0+0+0 \NC 7+FF+10FFFF \NC \NC \NC \NC \NR
81 \NC \lpr {Umathcharnumdef}\rlap{\high{5}} \NC -80000000 \NC 7FFFFFFF \NC \NC \NC \NC \NR
82 \NC \lpr {Umathcode} \NC 0 \NC 10FFFF \NC = \NC 0+0+0 \NC 7+FF+10FFFF \NC \NR
83 \NC \lpr {Udelcode} \NC 0 \NC 10FFFF \NC = \NC 0+0 \NC FF+10FFFF \NC \NR
84 \NC \lpr {Umathchar} \NC 0+0+0 \NC 7+FF+10FFFF \NC \NC \NC \NC \NR
85 \NC \lpr {Umathaccent} \NC 0+0+0 \NC 7+FF+10FFFF \NC \NC \NC \NC \NR
86 \NC \lpr {Udelimiter} \NC 0+0+0 \NC 7+FF+10FFFF \NC \NC \NC \NC \NR
87 \NC \lpr {Uradical} \NC 0+0 \NC FF+10FFFF \NC \NC \NC \NC \NR
88 \NC \lpr {Umathcharnum} \NC -80000000 \NC 7FFFFFFF \NC \NC \NC \NC \NR
89 \NC \lpr {Umathcodenum} \NC 0 \NC 10FFFF \NC = \NC -80000000 \NC 7FFFFFFF \NC \NR
90 \NC \lpr {Udelcodenum} \NC 0 \NC 10FFFF \NC = \NC -80000000 \NC 7FFFFFFF \NC \NR
91 \LL
92 \stoptabulate
93
94 Specifications typically look like:
95
96 \starttyping
97 \Umathchardef\xx="1"0"456
98 \Umathcode 123="1"0"789
99 \stoptyping
100
101 The new primitives that deal with delimiter|-|style objects do not set up a
102 \quote {large family}. Selecting a suitable size for display purposes is expected
103 to be dealt with by the font via the \lpr {Umathoperatorsize} parameter.
104
105 For some of these primitives, all information is packed into a single signed
106 integer. For the first two (\lpr {Umathcharnum} and \lpr {Umathcodenum}), the
107 lowest 21 bits are the character code, the 3 bits above that represent the math
108 class, and the family data is kept in the topmost bits. This means that the values
109 for math families 128--255 are actually negative. For \lpr {Udelcodenum} there
110 is no math class. The math family information is stored in the bits directly on
111 top of the character code. Using these three commands is not as natural as using
112 the two- and three|-|value commands, so unless you know exactly what you are
113 doing and absolutely require the speedup resulting from the faster input
114 scanning, it is better to use the verbose commands instead.
115
116 The \lpr {Umathaccent} command accepts optional keywords to control various
117 details regarding math accents. See \in {section} [mathacc] below for details.
118
119 There are more new primitives and all of these will be explained in following
120 sections:
121
122 \starttabulate[|l|l|]
123 \DB primitive \BC value range (in hex) \NC \NR
124 \TB
125 \NC \lpr {Uroot} \NC 0 + 0--FF + 10FFFF \NC \NR
126 \NC \lpr {Uoverdelimiter} \NC 0 + 0--FF + 10FFFF \NC \NR
127 \NC \lpr {Uunderdelimiter} \NC 0 + 0--FF + 10FFFF \NC \NR
128 \NC \lpr {Udelimiterover} \NC 0 + 0--FF + 10FFFF \NC \NR
129 \NC \lpr {Udelimiterunder} \NC 0 + 0--FF + 10FFFF \NC \NR
130 \LL
131 \stoptabulate
132
133 Instead of the pseudo class variable (7) you can use a family number as signal
134 for using the current family. This permits classifying characters with a class
135 and still let the family adapt. The trigger family is set with \lpr
136 {variablefam}. So:
137
138 \starttyping
139 \variablefam"24
140 \Umathchardef\foo "3 "24 123
141 \foo \fam9
142 \stoptyping
143
144 Results in a curly left brace taken from family 9 with class \quote {relation}
145 and spacing around it will be accordingly.
146
147 \stopsection
148
149 \startsection[title={Math styles}]
150
151 \subsection{\lpr {mathstyle}}
152
153 \topicindex {math+styles}
154
155 It is possible to discover the math style that will be used for a formula in an
156 expandable fashion (while the math list is still being read). To make this
157 possible, \LUATEX\ adds the new primitive: \lpr {mathstyle}. This is a \quote
158 {convert command} like e.g. \prm {romannumeral}: its value can only be read,
159 not set.
160
161 The returned value is between 0 and 7 (in math mode), or $-1$ (all other modes).
162 For easy testing, the eight math style commands have been altered so that they can
163 be used as numeric values, so you can write code like this:
164
165 \starttyping
166 \ifnum\mathstyle=\textstyle
167 \message{normal text style}
168 \else \ifnum\mathstyle=\crampedtextstyle
169 \message{cramped text style}
170 \fi \fi
171 \stoptyping
172
173 Sometimes you won't get what you expect so a bit of explanation might help to
174 understand what happens. When math is parsed and expanded it gets turned into a
175 linked list. In a second pass the formula will be build. This has to do with the
176 fact that in order to determine the automatically chosen sizes (in for instance
177 fractions) following content can influence preceding sizes. A side effect of this
178 is for instance that one cannot change the definition of a font family (and
179 thereby reusing numbers) because the number that got used is stored and used in
180 the second pass (so changing \type {\fam 12} mid|-|formula spoils over to
181 preceding use of that family).
182
183 The style switching primitives like \prm {textstyle} are turned into nodes so the
184 styles set there are frozen. The \prm {mathchoice} primitive results in four
185 lists being constructed of which one is used in the second pass. The fact that
186 some automatic styles are not yet known also means that the \lpr {mathstyle}
187 primitive expands to the current style which can of course be different from the
188 one really used. It's a snapshot of the first pass state. As a consequence in the
189 following example you get a style number (first pass) typeset that can actually
190 differ from the used style (second pass). In the case of a math choice used
191 ungrouped, the chosen style is used after the choice too, unless you group.
192
193 \startbuffer[1]
194 [a:\mathstyle]\quad
195 \bgroup
196 \mathchoice
197 {\bf \scriptstyle (x:d :\mathstyle)}
198 {\bf \scriptscriptstyle (x:t :\mathstyle)}
199 {\bf \scriptscriptstyle (x:s :\mathstyle)}
200 {\bf \scriptscriptstyle (x:ss:\mathstyle)}
201 \egroup
202 \quad[b:\mathstyle]\quad
203 \mathchoice
204 {\bf \scriptstyle (y:d :\mathstyle)}
205 {\bf \scriptscriptstyle (y:t :\mathstyle)}
206 {\bf \scriptscriptstyle (y:s :\mathstyle)}
207 {\bf \scriptscriptstyle (y:ss:\mathstyle)}
208 \quad[c:\mathstyle]\quad
209 \bgroup
210 \mathchoice
211 {\bf \scriptstyle (z:d :\mathstyle)}
212 {\bf \scriptscriptstyle (z:t :\mathstyle)}
213 {\bf \scriptscriptstyle (z:s :\mathstyle)}
214 {\bf \scriptscriptstyle (z:ss:\mathstyle)}
215 \egroup
216 \quad[d:\mathstyle]
217 \stopbuffer
218
219 \startbuffer[2]
220 [a:\mathstyle]\quad
221 \begingroup
222 \mathchoice
223 {\bf \scriptstyle (x:d :\mathstyle)}
224 {\bf \scriptscriptstyle (x:t :\mathstyle)}
225 {\bf \scriptscriptstyle (x:s :\mathstyle)}
226 {\bf \scriptscriptstyle (x:ss:\mathstyle)}
227 \endgroup
228 \quad[b:\mathstyle]\quad
229 \mathchoice
230 {\bf \scriptstyle (y:d :\mathstyle)}
231 {\bf \scriptscriptstyle (y:t :\mathstyle)}
232 {\bf \scriptscriptstyle (y:s :\mathstyle)}
233 {\bf \scriptscriptstyle (y:ss:\mathstyle)}
234 \quad[c:\mathstyle]\quad
235 \begingroup
236 \mathchoice
237 {\bf \scriptstyle (z:d :\mathstyle)}
238 {\bf \scriptscriptstyle (z:t :\mathstyle)}
239 {\bf \scriptscriptstyle (z:s :\mathstyle)}
240 {\bf \scriptscriptstyle (z:ss:\mathstyle)}
241 \endgroup
242 \quad[d:\mathstyle]
243 \stopbuffer
244
245 \typebuffer[1]
246
247 % \typebuffer[2]
248
249 This gives:
250
251 \blank $\displaystyle \getbuffer[1]$ \blank
252 \blank $\textstyle \getbuffer[1]$ \blank
253
254 Using \prm {begingroup} \unknown\ \prm {endgroup} instead gives:
255
256 \blank $\displaystyle \getbuffer[2]$ \blank
257 \blank $\textstyle \getbuffer[2]$ \blank
258
259 This might look wrong but it's just a side effect of \lpr {mathstyle} expanding
260 to the current (first pass) style and the number being injected in the list that
261 gets converted in the second pass. It all makes sense and it illustrates the
262 importance of grouping. In fact, the math choice style being effective afterwards
263 has advantages. It would be hard to get it otherwise.
264
265 \subsection{\lpr {Ustack}}
266
267 \topicindex {math+stacks}
268
269 There are a few math commands in \TEX\ where the style that will be used is not
270 known straight from the start. These commands (\prm {over}, \prm {atop},
271 \prm {overwithdelims}, \prm {atopwithdelims}) would therefore normally return
272 wrong values for \lpr {mathstyle}. To fix this, \LUATEX\ introduces a special
273 prefix command: \lpr {Ustack}:
274
275 \starttyping
276 $\Ustack {a \over b}$
277 \stoptyping
278
279 The \lpr {Ustack} command will scan the next brace and start a new math group
280 with the correct (numerator) math style.
281
282 \subsection{Cramped math styles}
283
284 \topicindex {math+styles}
285 \topicindex {math+spacing}
286 \topicindex {math+cramped}
287
288 \LUATEX\ has four new primitives to set the cramped math styles directly:
289
290 \starttyping
291 \crampeddisplaystyle
292 \crampedtextstyle
293 \crampedscriptstyle
294 \crampedscriptscriptstyle
295 \stoptyping
296
297 These additional commands are not all that valuable on their own, but they come
298 in handy as arguments to the math parameter settings that will be added shortly.
299
300 In Eijkhouts \quotation {\TEX\ by Topic} the rules for handling styles in scripts
301 are described as follows:
302
303 \startitemize
304 \startitem
305 In any style superscripts and subscripts are taken from the next smaller style.
306 Exception: in display style they are in script style.
307 \stopitem
308 \startitem
309 Subscripts are always in the cramped variant of the style; superscripts are only
310 cramped if the original style was cramped.
311 \stopitem
312 \startitem
313 In an \type {..\over..} formula in any style the numerator and denominator are
314 taken from the next smaller style.
315 \stopitem
316 \startitem
317 The denominator is always in cramped style; the numerator is only in cramped
318 style if the original style was cramped.
319 \stopitem
320 \startitem
321 Formulas under a \type {\sqrt} or \prm {overline} are in cramped style.
322 \stopitem
323 \stopitemize
324
325 In \LUATEX\ one can set the styles in more detail which means that you sometimes
326 have to set both normal and cramped styles to get the effect you want. (Even) if
327 we force styles in the script using \prm {scriptstyle} and \lpr
328 {crampedscriptstyle} we get this:
329
330 \startbuffer[demo]
331 \starttabulate
332 \DB style \BC example \NC \NR
333 \TB
334 \NC default \NC $b_{x=xx}^{x=xx}$ \NC \NR
335 \NC script \NC $b_{\scriptstyle x=xx}^{\scriptstyle x=xx}$ \NC \NR
336 \NC crampedscript \NC $b_{\crampedscriptstyle x=xx}^{\crampedscriptstyle x=xx}$ \NC \NR
337 \LL
338 \stoptabulate
339 \stopbuffer
340
341 \getbuffer[demo]
342
343 Now we set the following parameters
344
345 \startbuffer[setup]
346 \Umathordrelspacing\scriptstyle=30mu
347 \Umathordordspacing\scriptstyle=30mu
348 \stopbuffer
349
350 \typebuffer[setup]
351
352 This gives a different result:
353
354 \start\getbuffer[setup,demo]\stop
355
356 But, as this is not what is expected (visually) we should say:
357
358 \startbuffer[setup]
359 \Umathordrelspacing\scriptstyle=30mu
360 \Umathordordspacing\scriptstyle=30mu
361 \Umathordrelspacing\crampedscriptstyle=30mu
362 \Umathordordspacing\crampedscriptstyle=30mu
363 \stopbuffer
364
365 \typebuffer[setup]
366
367 Now we get:
368
369 \start\getbuffer[setup,demo]\stop
370
371 \stopsection
372
373 \startsection[title={Math parameter settings}]
374
375 \subsection {Many new \lpr {Umath*} primitives}
376
377 \topicindex {math+parameters}
378
379 In \LUATEX, the font dimension parameters that \TEX\ used in math typesetting are
380 now accessible via primitive commands. In fact, refactoring of the math engine
381 has resulted in many more parameters than were not accessible before.
382
383 \starttabulate
384 \DB primitive name \BC description \NC \NR
385 \TB
386 \NC \lpr {Umathquad} \NC the width of 18 mu's \NC \NR
387 \NC \lpr {Umathaxis} \NC height of the vertical center axis of
388 the math formula above the baseline \NC \NR
389 \NC \lpr {Umathoperatorsize} \NC minimum size of large operators in display mode \NC \NR
390 \NC \lpr {Umathoverbarkern} \NC vertical clearance above the rule \NC \NR
391 \NC \lpr {Umathoverbarrule} \NC the width of the rule \NC \NR
392 \NC \lpr {Umathoverbarvgap} \NC vertical clearance below the rule \NC \NR
393 \NC \lpr {Umathunderbarkern} \NC vertical clearance below the rule \NC \NR
394 \NC \lpr {Umathunderbarrule} \NC the width of the rule \NC \NR
395 \NC \lpr {Umathunderbarvgap} \NC vertical clearance above the rule \NC \NR
396 \NC \lpr {Umathradicalkern} \NC vertical clearance above the rule \NC \NR
397 \NC \lpr {Umathradicalrule} \NC the width of the rule \NC \NR
398 \NC \lpr {Umathradicalvgap} \NC vertical clearance below the rule \NC \NR
399 \NC \lpr {Umathradicaldegreebefore}\NC the forward kern that takes place before placement of
400 the radical degree \NC \NR
401 \NC \lpr {Umathradicaldegreeafter} \NC the backward kern that takes place after placement of
402 the radical degree \NC \NR
403 \NC \lpr {Umathradicaldegreeraise} \NC this is the percentage of the total height and depth of
404 the radical sign that the degree is raised by; it is
405 expressed in \type {percents}, so 60\% is expressed as the
406 integer $60$ \NC \NR
407 \NC \lpr {Umathstackvgap} \NC vertical clearance between the two
408 elements in a \prm {atop} stack \NC \NR
409 \NC \lpr {Umathstacknumup} \NC numerator shift upward in \prm {atop} stack \NC \NR
410 \NC \lpr {Umathstackdenomdown} \NC denominator shift downward in \prm {atop} stack \NC \NR
411 \NC \lpr {Umathfractionrule} \NC the width of the rule in a \prm {over} \NC \NR
412 \NC \lpr {Umathfractionnumvgap} \NC vertical clearance between the numerator and the rule \NC \NR
413 \NC \lpr {Umathfractionnumup} \NC numerator shift upward in \prm {over} \NC \NR
414 \NC \lpr {Umathfractiondenomvgap} \NC vertical clearance between the denominator and the rule \NC \NR
415 \NC \lpr {Umathfractiondenomdown} \NC denominator shift downward in \prm {over} \NC \NR
416 \NC \lpr {Umathfractiondelsize} \NC minimum delimiter size for \type {\...withdelims} \NC \NR
417 \NC \lpr {Umathlimitabovevgap} \NC vertical clearance for limits above operators \NC \NR
418 \NC \lpr {Umathlimitabovebgap} \NC vertical baseline clearance for limits above operators \NC \NR
419 \NC \lpr {Umathlimitabovekern} \NC space reserved at the top of the limit \NC \NR
420 \NC \lpr {Umathlimitbelowvgap} \NC vertical clearance for limits below operators \NC \NR
421 \NC \lpr {Umathlimitbelowbgap} \NC vertical baseline clearance for limits below operators \NC \NR
422 \NC \lpr {Umathlimitbelowkern} \NC space reserved at the bottom of the limit \NC \NR
423 \NC \lpr {Umathoverdelimitervgap} \NC vertical clearance for limits above delimiters \NC \NR
424 \NC \lpr {Umathoverdelimiterbgap} \NC vertical baseline clearance for limits above delimiters \NC \NR
425 \NC \lpr {Umathunderdelimitervgap} \NC vertical clearance for limits below delimiters \NC \NR
426 \NC \lpr {Umathunderdelimiterbgap} \NC vertical baseline clearance for limits below delimiters \NC \NR
427 \NC \lpr {Umathsubshiftdrop} \NC subscript drop for boxes and subformulas \NC \NR
428 \NC \lpr {Umathsubshiftdown} \NC subscript drop for characters \NC \NR
429 \NC \lpr {Umathsupshiftdrop} \NC superscript drop (raise, actually) for boxes and subformulas \NC \NR
430 \NC \lpr {Umathsupshiftup} \NC superscript raise for characters \NC \NR
431 \NC \lpr {Umathsubsupshiftdown} \NC subscript drop in the presence of a superscript \NC \NR
432 \NC \lpr {Umathsubtopmax} \NC the top of standalone subscripts cannot be higher than this
433 above the baseline \NC \NR
434 \NC \lpr {Umathsupbottommin} \NC the bottom of standalone superscripts cannot be less than
435 this above the baseline \NC \NR
436 \NC \lpr {Umathsupsubbottommax} \NC the bottom of the superscript of a combined super- and subscript
437 be at least as high as this above the baseline \NC \NR
438 \NC \lpr {Umathsubsupvgap} \NC vertical clearance between super- and subscript \NC \NR
439 \NC \lpr {Umathspaceafterscript} \NC additional space added after a super- or subscript \NC \NR
440 \NC \lpr {Umathconnectoroverlapmin}\NC minimum overlap between parts in an extensible recipe \NC \NR
441 \LL
442 \stoptabulate
443
444 Each of the parameters in this section can be set by a command like this:
445
446 \starttyping
447 \Umathquad\displaystyle=1em
448 \stoptyping
449
450 they obey grouping, and you can use \type {\the\Umathquad\displaystyle} if
451 needed.
452
453 \subsection{Font|-|based math parameters}
454
455 \topicindex {math+parameters}
456
457 While it is nice to have these math parameters available for tweaking, it would
458 be tedious to have to set each of them by hand. For this reason, \LUATEX\
459 initializes a bunch of these parameters whenever you assign a font identifier to
460 a math family based on either the traditional math font dimensions in the font
461 (for assignments to math family~2 and~3 using \TFM|-|based fonts like \type
462 {cmsy} and \type {cmex}), or based on the named values in a potential \type
463 {MathConstants} table when the font is loaded via Lua. If there is a \type
464 {MathConstants} table, this takes precedence over font dimensions, and in that
465 case no attention is paid to which family is being assigned to: the \type
466 {MathConstants} tables in the last assigned family sets all parameters.
467
468 In the table below, the one|-|letter style abbreviations and symbolic tfm font
469 dimension names match those used in the \TeX book. Assignments to \prm
470 {textfont} set the values for the cramped and uncramped display and text styles,
471 \prm {scriptfont} sets the script styles, and \prm {scriptscriptfont} sets the
472 scriptscript styles, so we have eight parameters for three font sizes. In the
473 \TFM\ case, assignments only happen in family~2 and family~3 (and of course only
474 for the parameters for which there are font dimensions).
475
476 Besides the parameters below, \LUATEX\ also looks at the \quote {space} font
477 dimension parameter. For math fonts, this should be set to zero.
478
479 \def\MathLine#1#2#3#4#5%
480 {\TB
481 \NC \llap{\high{\tx #2\enspace}}\ttbf \string #1 \NC \tt #5 \NC \NR
482 \NC \tx #3 \NC \tt #4 \NC \NR}
483
484 \starttabulate[|l|l|]
485 \DB variable / style \BC tfm / opentype \NC \NR
486 \MathLine{\Umathaxis} {} {} {AxisHeight} {axis_height}
487 \MathLine{\Umathoperatorsize} {6} {D, D'} {DisplayOperatorMinHeight} {\emdash}
488 \MathLine{\Umathfractiondelsize} {9} {D, D'} {FractionDelimiterDisplayStyleSize} {delim1}
489 \MathLine{\Umathfractiondelsize} {9} {T, T', S, S', SS, SS'}{FractionDelimiterSize} {delim2}
490 \MathLine{\Umathfractiondenomdown} {} {D, D'} {FractionDenominatorDisplayStyleShiftDown}{denom1}
491 \MathLine{\Umathfractiondenomdown} {} {T, T', S, S', SS, SS'}{FractionDenominatorShiftDown} {denom2}
492 \MathLine{\Umathfractiondenomvgap} {} {D, D'} {FractionDenominatorDisplayStyleGapMin} {3*default_rule_thickness}
493 \MathLine{\Umathfractiondenomvgap} {} {T, T', S, S', SS, SS'}{FractionDenominatorGapMin} {default_rule_thickness}
494 \MathLine{\Umathfractionnumup} {} {D, D'} {FractionNumeratorDisplayStyleShiftUp} {num1}
495 \MathLine{\Umathfractionnumup} {} {T, T', S, S', SS, SS'}{FractionNumeratorShiftUp} {num2}
496 \MathLine{\Umathfractionnumvgap} {} {D, D'} {FractionNumeratorDisplayStyleGapMin} {3*default_rule_thickness}
497 \MathLine{\Umathfractionnumvgap} {} {T, T', S, S', SS, SS'}{FractionNumeratorGapMin} {default_rule_thickness}
498 \MathLine{\Umathfractionrule} {} {} {FractionRuleThickness} {default_rule_thickness}
499 \MathLine{\Umathskewedfractionhgap} {} {} {SkewedFractionHorizontalGap} {math_quad/2}
500 \MathLine{\Umathskewedfractionvgap} {} {} {SkewedFractionVerticalGap} {math_x_height}
501 \MathLine{\Umathlimitabovebgap} {} {} {UpperLimitBaselineRiseMin} {big_op_spacing3}
502 \MathLine{\Umathlimitabovekern} {1} {} {0} {big_op_spacing5}
503 \MathLine{\Umathlimitabovevgap} {} {} {UpperLimitGapMin} {big_op_spacing1}
504 \MathLine{\Umathlimitbelowbgap} {} {} {LowerLimitBaselineDropMin} {big_op_spacing4}
505 \MathLine{\Umathlimitbelowkern} {1} {} {0} {big_op_spacing5}
506 \MathLine{\Umathlimitbelowvgap} {} {} {LowerLimitGapMin} {big_op_spacing2}
507 \MathLine{\Umathoverdelimitervgap} {} {} {StretchStackGapBelowMin} {big_op_spacing1}
508 \MathLine{\Umathoverdelimiterbgap} {} {} {StretchStackTopShiftUp} {big_op_spacing3}
509 \MathLine{\Umathunderdelimitervgap} {} {} {StretchStackGapAboveMin} {big_op_spacing2}
510 \MathLine{\Umathunderdelimiterbgap} {} {} {StretchStackBottomShiftDown} {big_op_spacing4}
511 \MathLine{\Umathoverbarkern} {} {} {OverbarExtraAscender} {default_rule_thickness}
512 \MathLine{\Umathoverbarrule} {} {} {OverbarRuleThickness} {default_rule_thickness}
513 \MathLine{\Umathoverbarvgap} {} {} {OverbarVerticalGap} {3*default_rule_thickness}
514 \MathLine{\Umathquad} {1} {} {<font_size(f)>} {math_quad}
515 \MathLine{\Umathradicalkern} {} {} {RadicalExtraAscender} {default_rule_thickness}
516 \MathLine{\Umathradicalrule} {2} {} {RadicalRuleThickness} {<not set>}
517 \MathLine{\Umathradicalvgap} {3} {D, D'} {RadicalDisplayStyleVerticalGap} {default_rule_thickness+abs(math_x_height)/4}
518 \MathLine{\Umathradicalvgap} {3} {T, T', S, S', SS, SS'}{RadicalVerticalGap} {default_rule_thickness+abs(default_rule_thickness)/4}
519 \MathLine{\Umathradicaldegreebefore}{2} {} {RadicalKernBeforeDegree} {<not set>}
520 \MathLine{\Umathradicaldegreeafter} {2} {} {RadicalKernAfterDegree} {<not set>}
521 \MathLine{\Umathradicaldegreeraise} {2,7}{} {RadicalDegreeBottomRaisePercent} {<not set>}
522 \MathLine{\Umathspaceafterscript} {4} {} {SpaceAfterScript} {script_space}
523 \MathLine{\Umathstackdenomdown} {} {D, D'} {StackBottomDisplayStyleShiftDown} {denom1}
524 \MathLine{\Umathstackdenomdown} {} {T, T', S, S', SS, SS'}{StackBottomShiftDown} {denom2}
525 \MathLine{\Umathstacknumup} {} {D, D'} {StackTopDisplayStyleShiftUp} {num1}
526 \MathLine{\Umathstacknumup} {} {T, T', S, S', SS, SS'}{StackTopShiftUp} {num3}
527 \MathLine{\Umathstackvgap} {} {D, D'} {StackDisplayStyleGapMin} {7*default_rule_thickness}
528 \MathLine{\Umathstackvgap} {} {T, T', S, S', SS, SS'}{StackGapMin} {3*default_rule_thickness}
529 \MathLine{\Umathsubshiftdown} {} {} {SubscriptShiftDown} {sub1}
530 \MathLine{\Umathsubshiftdrop} {} {} {SubscriptBaselineDropMin} {sub_drop}
531 \MathLine{\Umathsubsupshiftdown} {8} {} {SubscriptShiftDownWithSuperscript} {\emdash}
532 \MathLine{\Umathsubtopmax} {} {} {SubscriptTopMax} {abs(math_x_height*4)/5}
533 \MathLine{\Umathsubsupvgap} {} {} {SubSuperscriptGapMin} {4*default_rule_thickness}
534 \MathLine{\Umathsupbottommin} {} {} {SuperscriptBottomMin} {abs(math_x_height/4)}
535 \MathLine{\Umathsupshiftdrop} {} {} {SuperscriptBaselineDropMax} {sup_drop}
536 \MathLine{\Umathsupshiftup} {} {D} {SuperscriptShiftUp} {sup1}
537 \MathLine{\Umathsupshiftup} {} {T, S, SS,} {SuperscriptShiftUp} {sup2}
538 \MathLine{\Umathsupshiftup} {} {D', T', S', SS'} {SuperscriptShiftUpCramped} {sup3}
539 \MathLine{\Umathsupsubbottommax} {} {} {SuperscriptBottomMaxWithSubscript} {abs(math_x_height*4)/5}
540 \MathLine{\Umathunderbarkern} {} {} {UnderbarExtraDescender} {default_rule_thickness}
541 \MathLine{\Umathunderbarrule} {} {} {UnderbarRuleThickness} {default_rule_thickness}
542 \MathLine{\Umathunderbarvgap} {} {} {UnderbarVerticalGap} {3*default_rule_thickness}
543 \MathLine{\Umathconnectoroverlapmin}{5} {} {MinConnectorOverlap} {0}
544 \LL
545 \stoptabulate
546
547 Note 1: \OPENTYPE\ fonts set \lpr {Umathlimitabovekern} and \lpr
548 {Umathlimitbelowkern} to zero and set \lpr {Umathquad} to the font size of the
549 used font, because these are not supported in the \type {MATH} table,
550
551 Note 2: Traditional \TFM\ fonts do not set \lpr {Umathradicalrule} because
552 \TEX82\ uses the height of the radical instead. When this parameter is indeed not
553 set when \LUATEX\ has to typeset a radical, a backward compatibility mode will
554 kick in that assumes that an oldstyle \TEX\ font is used. Also, they do not set
555 \lpr {Umathradicaldegreebefore}, \lpr {Umathradicaldegreeafter}, and \lpr
556 {Umathradicaldegreeraise}. These are then automatically initialized to
557 $5/18$quad, $-10/18$quad, and 60.
558
559 Note 3: If \TFM\ fonts are used, then the \lpr {Umathradicalvgap} is not set
560 until the first time \LUATEX\ has to typeset a formula because this needs
561 parameters from both family~2 and family~3. This provides a partial backward
562 compatibility with \TEX82, but that compatibility is only partial: once the \lpr
563 {Umathradicalvgap} is set, it will not be recalculated any more.
564
565 Note 4: When \TFM\ fonts are used a similar situation arises with respect to \lpr
566 {Umathspaceafterscript}: it is not set until the first time \LUATEX\ has to
567 typeset a formula. This provides some backward compatibility with \TEX82. But
568 once the \lpr {Umathspaceafterscript} is set, \prm {scriptspace} will never be
569 looked at again.
570
571 Note 5: Traditional \TFM\ fonts set \lpr {Umathconnectoroverlapmin} to zero
572 because \TEX82\ always stacks extensibles without any overlap.
573
574 Note 6: The \lpr {Umathoperatorsize} is only used in \prm {displaystyle}, and is
575 only set in \OPENTYPE\ fonts. In \TFM\ font mode, it is artificially set to one
576 scaled point more than the initial attempt's size, so that always the \quote
577 {first next} will be tried, just like in \TEX82.
578
579 Note 7: The \lpr {Umathradicaldegreeraise} is a special case because it is the
580 only parameter that is expressed in a percentage instead of a number of scaled
581 points.
582
583 Note 8: \type {SubscriptShiftDownWithSuperscript} does not actually exist in the
584 \quote {standard} \OPENTYPE\ math font Cambria, but it is useful enough to be
585 added.
586
587 Note 9: \type {FractionDelimiterDisplayStyleSize} and \type
588 {FractionDelimiterSize} do not actually exist in the \quote {standard} \OPENTYPE\
589 math font Cambria, but were useful enough to be added.
590
591 \stopsection
592
593 \startsection[title={Math spacing}]
594
595 \subsection{Inline surrounding space}
596
597 \topicindex {math+spacing}
598
599 Inline math is surrounded by (optional) \prm {mathsurround} spacing but that is a fixed
600 dimension. There is now an additional parameter \lpr {mathsurroundskip}. When set to a
601 non|-|zero value (or zero with some stretch or shrink) this parameter will replace
602 \prm {mathsurround}. By using an additional parameter instead of changing the nature
603 of \prm {mathsurround}, we can remain compatible. In the meantime a bit more
604 control has been added via \lpr {mathsurroundmode}. This directive can take 6 values
605 with zero being the default behaviour.
606
607 \start
608
609 \def\OneLiner#1#2%
610 {\NC \type{#1}
611 \NC \dontleavehmode\inframed[align=normal,offset=0pt,frame=off]{\mathsurroundmode#1\relax\hsize 100pt x$x$x}
612 \NC \dontleavehmode\inframed[align=normal,offset=0pt,frame=off]{\mathsurroundmode#1\relax\hsize 100pt x $x$ x}
613 \NC #2
614 \NC \NR}
615
616 \startbuffer
617 \mathsurround 10pt
618 \mathsurroundskip20pt
619 \stopbuffer
620
621 \typebuffer \getbuffer
622
623 \starttabulate[|c|c|c|pl|]
624 \DB mode \BC x\$x\$x \BC x \$x\$ x \BC effect \NC \NR
625 \TB
626 \OneLiner{0}{obey \prm {mathsurround} when \lpr {mathsurroundskip} is 0pt}
627 \OneLiner{1}{only add skip to the left}
628 \OneLiner{2}{only add skip to the right}
629 \OneLiner{3}{add skip to the left and right}
630 \OneLiner{4}{ignore the skip setting, obey \prm {mathsurround}}
631 \OneLiner{5}{disable all spacing around math}
632 \OneLiner{6}{only apply \lpr {mathsurroundskip} when also spacing}
633 \OneLiner{7}{only apply \lpr {mathsurroundskip} when no spacing}
634 \LL
635 \stoptabulate
636
637 \stop
638
639 Method six omits the surround glue when there is (x)spacing glue present while
640 method seven does the opposite, the glue is only applied when there is (x)space
641 glue present too. Anything more fancy, like checking the begining or end of a
642 paragraph (or edges of a box) would not be robust anyway. If you want that you
643 can write a callback that runs over a list and analyzes a paragraph. Actually, in
644 that case you could also inject glue (or set the properties of a math node)
645 explicitly. So, these modes are in practice mostly useful for special purposes
646 and experiments (they originate in a tracker item). Keep in mind that this glue
647 is part of the math node and not always treated as normal glue: it travels with
648 the begin and end math nodes. Also, method 6 and 7 will zero the skip related
649 fields in a node when applicable in the first occasion that checks them
650 (linebreaking or packaging).
651
652 \subsection{Pairwise spacing}
653
654 \topicindex {math+spacing}
655
656 Besides the parameters mentioned in the previous sections, there are also 64 new
657 primitives to control the math spacing table (as explained in Chapter~18 of the
658 \TEX book). The primitive names are a simple matter of combining two math atom
659 types, but for completeness' sake, here is the whole list:
660
661 \starttwocolumns
662 \startlines
663 \lpr {Umathordordspacing}
664 \lpr {Umathordopspacing}
665 \lpr {Umathordbinspacing}
666 \lpr {Umathordrelspacing}
667 \lpr {Umathordopenspacing}
668 \lpr {Umathordclosespacing}
669 \lpr {Umathordpunctspacing}
670 \lpr {Umathordinnerspacing}
671 \lpr {Umathopordspacing}
672 \lpr {Umathopopspacing}
673 \lpr {Umathopbinspacing}
674 \lpr {Umathoprelspacing}
675 \lpr {Umathopopenspacing}
676 \lpr {Umathopclosespacing}
677 \lpr {Umathoppunctspacing}
678 \lpr {Umathopinnerspacing}
679 \lpr {Umathbinordspacing}
680 \lpr {Umathbinopspacing}
681 \lpr {Umathbinbinspacing}
682 \lpr {Umathbinrelspacing}
683 \lpr {Umathbinopenspacing}
684 \lpr {Umathbinclosespacing}
685 \lpr {Umathbinpunctspacing}
686 \lpr {Umathbininnerspacing}
687 \lpr {Umathrelordspacing}
688 \lpr {Umathrelopspacing}
689 \lpr {Umathrelbinspacing}
690 \lpr {Umathrelrelspacing}
691 \lpr {Umathrelopenspacing}
692 \lpr {Umathrelclosespacing}
693 \lpr {Umathrelpunctspacing}
694 \lpr {Umathrelinnerspacing}
695 \lpr {Umathopenordspacing}
696 \lpr {Umathopenopspacing}
697 \lpr {Umathopenbinspacing}
698 \lpr {Umathopenrelspacing}
699 \lpr {Umathopenopenspacing}
700 \lpr {Umathopenclosespacing}
701 \lpr {Umathopenpunctspacing}
702 \lpr {Umathopeninnerspacing}
703 \lpr {Umathcloseordspacing}
704 \lpr {Umathcloseopspacing}
705 \lpr {Umathclosebinspacing}
706 \lpr {Umathcloserelspacing}
707 \lpr {Umathcloseopenspacing}
708 \lpr {Umathcloseclosespacing}
709 \lpr {Umathclosepunctspacing}
710 \lpr {Umathcloseinnerspacing}
711 \lpr {Umathpunctordspacing}
712 \lpr {Umathpunctopspacing}
713 \lpr {Umathpunctbinspacing}
714 \lpr {Umathpunctrelspacing}
715 \lpr {Umathpunctopenspacing}
716 \lpr {Umathpunctclosespacing}
717 \lpr {Umathpunctpunctspacing}
718 \lpr {Umathpunctinnerspacing}
719 \lpr {Umathinnerordspacing}
720 \lpr {Umathinneropspacing}
721 \lpr {Umathinnerbinspacing}
722 \lpr {Umathinnerrelspacing}
723 \lpr {Umathinneropenspacing}
724 \lpr {Umathinnerclosespacing}
725 \lpr {Umathinnerpunctspacing}
726 \lpr {Umathinnerinnerspacing}
727 \stoplines
728 \stoptwocolumns
729
730 These parameters are of type \prm {muskip}, so setting a parameter can be done
731 like this:
732
733 \starttyping
734 \Umathopordspacing\displaystyle=4mu plus 2mu
735 \stoptyping
736
737 They are all initialized by \type {initex} to the values mentioned in the table
738 in Chapter~18 of the \TEX book.
739
740 Note 1: for ease of use as well as for backward compatibility, \prm {thinmuskip},
741 \prm {medmuskip} and \prm {thickmuskip} are treated specially. In their case a
742 pointer to the corresponding internal parameter is saved, not the actual \prm
743 {muskip} value. This means that any later changes to one of these three
744 parameters will be taken into account.
745
746 Note 2: Careful readers will realise that there are also primitives for the items
747 marked \type {*} in the \TEX book. These will not actually be used as those
748 combinations of atoms cannot actually happen, but it seemed better not to break
749 orthogonality. They are initialized to zero.
750
751 \subsection{Skips around display math}
752
753 \topicindex {math+spacing}
754
755 The injection of \prm {abovedisplayskip} and \prm {belowdisplayskip} is not
756 symmetrical. An above one is always inserted, also when zero, but the below is
757 only inserted when larger than zero. Especially the latter makes it sometimes hard
758 to fully control spacing. Therefore \LUATEX\ comes with a new directive: \lpr
759 {mathdisplayskipmode}. The following values apply:
760
761 \starttabulate[|c|l|]
762 \DB value \BC meaning \NC \NR
763 \TB
764 \NC 0 \NC normal \TEX\ behaviour \NC \NR
765 \NC 1 \NC always (same as 0) \NC \NR
766 \NC 2 \NC only when not zero \NC \NR
767 \NC 3 \NC never, not even when not zero \NC \NR
768 \LL
769 \stoptabulate
770
771 By default the short skip detection is not adapted to r2l typesetting and that
772 hasn't been the case since the start of the project. Changing it could break
773 hacks that users came up with but when you set \lpr {matheqdirmode} to a positive
774 value direction will be taken into account.
775
776 % \subsection {Nolimit correction}
777 %
778 % \topicindex {math+limits}
779 %
780 % There are two extra math parameters \lpr {Umathnolimitsupfactor} and \lpr
781 % {Umathnolimitsubfactor} that were added to provide some control over how limits
782 % are spaced (for example the position of super and subscripts after integral
783 % operators). They relate to an extra parameter \lpr {mathnolimitsmode}. The half
784 % corrections are what happens when scripts are placed above and below. The
785 % problem with italic corrections is that officially that correction italic is used
786 % for above|/|below placement while advanced kerns are used for placement at the
787 % right end. The question is: how often is this implemented, and if so, do the
788 % kerns assume correction too. Anyway, with this parameter one can control it.
789
790 % \starttabulate[|l|ck1|ck1|ck1|ck1|ck1|ck1|]
791 % \NC % probably not ok, we need a raw int here
792 % \NC \mathnolimitsmode0 $\displaystyle\mathop{\normalint}\nolimits^0_1$
793 % \NC \mathnolimitsmode1 $\displaystyle\mathop{\normalint}\nolimits^0_1$
794 % \NC \mathnolimitsmode2 $\displaystyle\mathop{\normalint}\nolimits^0_1$
795 % \NC \mathnolimitsmode3 $\displaystyle\mathop{\normalint}\nolimits^0_1$
796 % \NC \mathnolimitsmode4 $\displaystyle\mathop{\normalint}\nolimits^0_1$
797 % \NC \mathnolimitsmode8000 $\displaystyle\mathop{\normalint}\nolimits^0_1$
798 % \NC \NR
799 % \TB
800 % \BC mode
801 % \NC \tttf 0
802 % \NC \tttf 1
803 % \NC \tttf 2
804 % \NC \tttf 3
805 % \NC \tttf 4
806 % \NC \tttf 8000
807 % \NC \NR
808 % \BC superscript
809 % \NC 0
810 % \NC font
811 % \NC 0
812 % \NC 0
813 % \NC +ic/2
814 % \NC 0
815 % \NC \NR
816 % \BC subscript
817 % \NC -ic
818 % \NC font
819 % \NC 0
820 % \NC -ic/2
821 % \NC -ic/2
822 % \NC 8000ic/1000
823 % \NC \NR
824 % \stoptabulate
825 %
826 % When the mode is set to one, the math parameters are used. This way a macro
827 % package writer can decide what looks best. Given the current state of fonts in
828 % \CONTEXT\ we currently use mode 1 with factor 0 for the superscript and 750 for
829 % the subscripts. Positive values are used for both parameters but the subscript
830 % shifts to the left. A \lpr {mathnolimitsmode} larger that 15 is considered to
831 % be a factor for the subscript correction. This feature can be handy when
832 % experimenting.
833
834 % \subsection {Math italic mess}
835 %
836 % \topicindex {math+italics}
837 %
838 % The \lpr {mathitalicsmode} parameter was introduced to deal with the difference
839 % in applying italic correction in traditional and \OPENTYPE\ math fonts. There are
840 % \OPENTYPE\ fonts out there that have italic correction and assume them to be
841 % applied like traditional \TEX\ fonts. This parameter takes several values:
842 %
843 % When set to zero, you get what was decided when the two code paths (traditional
844 % and \OPENTYPE) were introduced.
845 %
846 % Values larger than zero will add the italic correction between simple noads (it
847 % checks some classes so you might pay attention to for instance punctuation
848 % classes assigned).
849 %
850 % When set to zero or one, italics are independent, so we separate width from
851 % italic, while values larger than one combine both in the width but later
852 % selectively has to get rid of it (depending on code path).
853 %
854 % A value larger than two will backtrack italics for large operators, because there
855 % italic correction is used for anchoring scripts (limits and no limits). In fact,
856 % \OPENTYPE\ uses italics either between characters or for this purpose but as
857 % mentioned fonts are sort of messy here.
858 %
859 % We tested our version of plain \TEX\ and recommend to use the value of three to
860 % get the best average results. More about this italic correction dilemma in
861 % rendering math can be found in articles (in for instance \TUGBOAT) and various
862 % documents in the \CONTEXT\ distribution, especially those that discuss the
863 % upgraded math engine in \LUAMETATEX.
864 %
865 % % The \lpr {mathitalicsmode} parameter can be set to~1 to force italic correction
866 % % before noads that represent some more complex structure (read: everything that is
867 % % not an ord, bin, rel, open, close, punct or inner). A value of~2 will enforce the
868 % % old school font code path for all italics. We show a Cambria example.
869 % %
870 % % \starttexdefinition Whatever #1
871 % % \NC \type{\mathitalicsmode = #1}
872 % % \NC \mathitalicsmode#1\ruledhbox{$\left|T^1\right|$}
873 % % \NC \mathitalicsmode#1\ruledhbox{$\left|T\right|$}
874 % % \NC \mathitalicsmode#1\ruledhbox{$T+1$}
875 % % \NC \mathitalicsmode#1\ruledhbox{$T{1\over2}$}
876 % % \NC \mathitalicsmode#1\ruledhbox{$T\sqrt{1}$}
877 % % \NC \NR
878 % % \stoptexdefinition
879 % %
880 % % \start
881 % % \switchtobodyfont[cambria]
882 % % \starttabulate[|c|c|c|c|c|c|]
883 % % \Whatever{0}%
884 % % \Whatever{1}%
885 % % \stoptabulate
886 % % \stop
887 % %
888 % % This kind of parameters relate to the fact that italic correction in \OPENTYPE\
889 % % math is bound to fuzzy rules. So, control is the solution.
890
891 % \subsection {Script and kerning}
892 %
893 % \topicindex {math+kerning}
894 % \topicindex {math+scripts}
895 %
896 % If you want to typeset text in math macro packages often provide something \type
897 % {\text} which obeys the script sizes. As the definition can be anything there is
898 % a good chance that the kerning doesn't come out well when used in a script. Given
899 % that the first glyph ends up in a \prm {hbox} we have some control over this.
900 % And, as a bonus we also added control over the normal sublist kerning. The \lpr
901 % {mathscriptboxmode} parameter defaults to~1.
902 %
903 % \starttabulate[|c|l|]
904 % \DB value \BC meaning \NC \NR
905 % \TB
906 % \NC \type {0} \NC forget about kerning \NC \NR
907 % \NC \type {1} \NC kern math sub lists with a valid glyph \NC \NR
908 % \NC \type {2} \NC also kern math sub boxes that have a valid glyph \NC \NR
909 % \NC \type {2} \NC only kern math sub boxes with a boundary node present\NC \NR
910 % \LL
911 % \stoptabulate
912 %
913 % Here we show some examples. Of course this doesn't solve all our problems, if
914 % only because some fonts have characters with bounding boxes that compensate for
915 % italics, while other fonts can lack kerns.
916 %
917 % \startbuffer[1]
918 % $T_{\tf fluff}$
919 % \stopbuffer
920 %
921 % \startbuffer[2]
922 % $T_{\text{fluff}}$
923 % \stopbuffer
924 %
925 % \startbuffer[3]
926 % $T_{\text{\boundary1 fluff}}$
927 % \stopbuffer
928 %
929 % \unexpanded\def\Show#1#2#3%
930 % {\doifelsenothing{#3}
931 % {\small\tx\typeinlinebuffer[#1]}
932 % {\doifelse{#3}{-}
933 % {\small\bf\tt mode #2}
934 % {\switchtobodyfont[#3]\showfontkerns\showglyphs\mathscriptboxmode#2\relax\inlinebuffer[#1]}}}
935 %
936 % \starttabulate[|lBT|c|c|c|c|c|]
937 % \NC \NC \Show{1}{0}{} \NC\Show{1}{1}{} \NC \Show{2}{1}{} \NC \Show{2}{2}{} \NC \Show{3}{3}{} \NC \NR
938 % \NC \NC \Show{1}{0}{-} \NC\Show{1}{1}{-} \NC \Show{2}{1}{-} \NC \Show{2}{2}{-} \NC \Show{3}{3}{-} \NC \NR
939 % \NC modern \NC \Show{1}{0}{modern} \NC\Show{1}{1}{modern} \NC \Show{2}{1}{modern} \NC \Show{2}{2}{modern} \NC \Show{3}{3}{modern} \NC \NR
940 % \NC lucidaot \NC \Show{1}{0}{lucidaot} \NC\Show{1}{1}{lucidaot} \NC \Show{2}{1}{lucidaot} \NC \Show{2}{2}{lucidaot} \NC \Show{3}{3}{lucidaot} \NC \NR
941 % \NC pagella \NC \Show{1}{0}{pagella} \NC\Show{1}{1}{pagella} \NC \Show{2}{1}{pagella} \NC \Show{2}{2}{pagella} \NC \Show{3}{3}{pagella} \NC \NR
942 % \NC cambria \NC \Show{1}{0}{cambria} \NC\Show{1}{1}{cambria} \NC \Show{2}{1}{cambria} \NC \Show{2}{2}{cambria} \NC \Show{3}{3}{cambria} \NC \NR
943 % \NC dejavu \NC \Show{1}{0}{dejavu} \NC\Show{1}{1}{dejavu} \NC \Show{2}{1}{dejavu} \NC \Show{2}{2}{dejavu} \NC \Show{3}{3}{dejavu} \NC \NR
944 % \stoptabulate
945 %
946 % % Kerning between a character subscript is controlled by \lpr {mathscriptcharmode}
947 % % which also defaults to~1.
948 %
949 % Here is another example. Internally we tag kerns as italic kerns or font kerns
950 % where font kerns result from the staircase kern tables. In 2018 fonts like Latin
951 % Modern and Pagella rely on cheats with the boundingbox, Cambria uses staircase
952 % kerns and Lucida a mixture. Depending on how fonts evolve we might add some more
953 % control over what one can turn on and off.
954 %
955 % \def\MathSample#1#2#3%
956 % {\NC
957 % #1 \NC
958 % #2 \NC
959 % \showglyphdata \switchtobodyfont[#2,17.3pt]$#3T_{f}$ \NC
960 % \showglyphdata \switchtobodyfont[#2,17.3pt]$#3\gamma_{e}$ \NC
961 % \showglyphdata \switchtobodyfont[#2,17.3pt]$#3\gamma_{ee}$ \NC
962 % \showglyphdata \switchtobodyfont[#2,17.3pt]$#3T_{\tf fluff}$ \NC
963 % \NR}
964 %
965 % \starttabulate[|Tl|Tl|l|l|l|l|]
966 % \FL
967 % \MathSample{normal}{modern} {\mr}
968 % \MathSample{} {pagella} {\mr}
969 % \MathSample{} {cambria} {\mr}
970 % \MathSample{} {lucidaot}{\mr}
971 % \ML
972 % \MathSample{bold} {modern} {\mb}
973 % \MathSample{} {pagella} {\mb}
974 % \MathSample{} {cambria} {\mb}
975 % \MathSample{} {lucidaot}{\mb}
976 % \LL
977 % \stoptabulate
978
979 \subsection{Fixed scripts}
980
981 We have three parameters that are used for this fixed anchoring:
982
983 \starttabulate[|c|l|]
984 \DB parameter \BC register \NC \NR
985 \NC $d$ \NC \lpr {Umathsubshiftdown} \NC \NR
986 \NC $u$ \NC \lpr {Umathsupshiftup} \NC \NR
987 \NC $s$ \NC \lpr {Umathsubsupshiftdown} \NC \NR
988 \LL
989 \stoptabulate
990
991 When we set \lpr {mathscriptsmode} to a value other than zero these are used
992 for calculating fixed positions. This is something that is needed for instance
993 for chemistry. You can manipulate the mentioned variables to achieve different
994 effects.
995
996 \def\SampleMath#1%
997 {$\mathscriptsmode#1\mathupright CH_2 + CH^+_2 + CH^2_2$}
998
999 \starttabulate[|c|c|c|p|]
1000 \DB mode \BC down \BC up \BC example \NC \NR
1001 \TB
1002 \NC 0 \NC dynamic \NC dynamic \NC \SampleMath{0} \NC \NR
1003 \NC 1 \NC $d$ \NC $u$ \NC \SampleMath{1} \NC \NR
1004 \NC 2 \NC $s$ \NC $u$ \NC \SampleMath{2} \NC \NR
1005 \NC 3 \NC $s$ \NC $u + s - d$ \NC \SampleMath{3} \NC \NR
1006 \NC 4 \NC $d + (s-d)/2$ \NC $u + (s-d)/2$ \NC \SampleMath{4} \NC \NR
1007 \NC 5 \NC $d$ \NC $u + s - d$ \NC \SampleMath{5} \NC \NR
1008 \LL
1009 \stoptabulate
1010
1011 The value of this parameter obeys grouping but applies to the whole current
1012 formula.
1013
1014 % if needed we can put the value in stylenodes but maybe more should go there
1015
1016 \subsection{Penalties: \lpr {mathpenaltiesmode}}
1017
1018 \topicindex {math+penalties}
1019
1020 Only in inline math penalties will be added in a math list. You can force
1021 penalties (also in display math) by setting:
1022
1023 \starttyping
1024 \mathpenaltiesmode = 1
1025 \stoptyping
1026
1027 This primnitive is not really needed in \LUATEX\ because you can use the callback
1028 \cbk {mlist_to_hlist} to force penalties by just calling the regular routine
1029 with forced penalties. However, as part of opening up and control this primitive
1030 makes sense. As a bonus we also provide two extra penalties:
1031
1032 \starttyping
1033 \prebinoppenalty = -100 % example value
1034 \prerelpenalty = 900 % example value
1035 \stoptyping
1036
1037 They default to inifinite which signals that they don't need to be inserted. When
1038 set they are injected before a binop or rel noad. This is an experimental feature.
1039
1040 \subsection{Equation spacing: \lpr {matheqnogapstep}}
1041
1042 By default \TEX\ will add one quad between the equation and the number. This is
1043 hard coded. A new primitive can control this:
1044
1045 \startsyntax
1046 \matheqnogapstep = 1000
1047 \stopsyntax
1048
1049 Because a math quad from the math text font is used instead of a dimension, we
1050 use a step to control the size. A value of zero will suppress the gap. The step
1051 is divided by 1000 which is the usual way to mimmick floating point factors in
1052 \TEX.
1053
1054 \stopsection
1055
1056 \startsection[title={Math constructs}]
1057
1058 % \subsection {Unscaled fences}
1059 %
1060 % \topicindex {math+fences}
1061 %
1062 % The \lpr {mathdelimitersmode} primitive is experimental and deals with the
1063 % following (potential) problems. Three bits can be set. The first bit prevents an
1064 % unwanted shift when the fence symbol is not scaled (a cambria side effect). The
1065 % second bit forces italic correction between a preceding character ordinal and the
1066 % fenced subformula, while the third bit turns that subformula into an ordinary so
1067 % that the same spacing applies as with unfenced variants. Here we show Cambria
1068 % (with \lpr {mathitalicsmode} enabled).
1069 %
1070 % \starttexdefinition Whatever #1
1071 % \NC \type{\mathdelimitersmode = #1}
1072 % \NC \mathitalicsmode1\mathdelimitersmode#1\ruledhbox{\showglyphs\showfontkerns\showfontitalics$f(x)$}
1073 % \NC \mathitalicsmode1\mathdelimitersmode#1\ruledhbox{\showglyphs\showfontkerns\showfontitalics$f\left(x\right)$}
1074 % \NC \NR
1075 % \stoptexdefinition
1076 %
1077 % \start
1078 % \switchtobodyfont[cambria]
1079 % \starttabulate[|l|l|l|]
1080 % \Whatever{0}\Whatever{1}\Whatever{2}\Whatever{3}%
1081 % \Whatever{4}\Whatever{5}\Whatever{6}\Whatever{7}%
1082 % \stoptabulate
1083 % \stop
1084 %
1085 % So, when set to 7 fenced subformulas with unscaled delimiters come out the same
1086 % as unfenced ones. This can be handy for cases where one is forced to use \prm
1087 % {left} and \prm {right} always because of unpredictable content. As said, it's an
1088 % experimental feature (which somehow fits in the exceptional way fences are dealt
1089 % with in the engine). The full list of flags is given in the next table:
1090 %
1091 % \starttabulate[|c|l|]
1092 % \DB value \BC meaning \NC \NR
1093 % \TB
1094 % \NC \type{"01} \NC don't apply the usual shift \NC \NR
1095 % \NC \type{"02} \NC apply italic correction when possible \NC \NR
1096 % \NC \type{"04} \NC force an ordinary subformula \NC \NR
1097 % \NC \type{"08} \NC no shift when a base character \NC \NR
1098 % \NC \type{"10} \NC only shift when an extensible \NC \NR
1099 % \LL
1100 % \stoptabulate
1101 %
1102 % The effect can depend on the font (and for Cambria one can use for instance \type {"16}).
1103
1104 \subsection[mathacc]{Accent handling}
1105
1106 \topicindex {math+accents}
1107
1108 \LUATEX\ supports both top accents and bottom accents in math mode, and math
1109 accents stretch automatically (if this is supported by the font the accent comes
1110 from, of course). Bottom and combined accents as well as fixed-width math accents
1111 are controlled by optional keywords following \lpr {Umathaccent}.
1112
1113 The keyword \type {bottom} after \lpr {Umathaccent} signals that a bottom accent
1114 is needed, and the keyword \type {both} signals that both a top and a bottom
1115 accent are needed (in this case two accents need to be specified, of course).
1116
1117 Then the set of three integers defining the accent is read. This set of integers
1118 can be prefixed by the \type {fixed} keyword to indicate that a non-stretching
1119 variant is requested (in case of both accents, this step is repeated).
1120
1121 A simple example:
1122
1123 \starttyping
1124 \Umathaccent both fixed 0 0 "20D7 fixed 0 0 "20D7 {example}
1125 \stoptyping
1126
1127 If a math top accent has to be placed and the accentee is a character and has a
1128 non-zero \type {top_accent} value, then this value will be used to place the
1129 accent instead of the \prm {skewchar} kern used by \TEX82.
1130
1131 The \type {top_accent} value represents a vertical line somewhere in the
1132 accentee. The accent will be shifted horizontally such that its own \type
1133 {top_accent} line coincides with the one from the accentee. If the \type
1134 {top_accent} value of the accent is zero, then half the width of the accent
1135 followed by its italic correction is used instead.
1136
1137 The vertical placement of a top accent depends on the \type {x_height} of the
1138 font of the accentee (as explained in the \TEX book), but if a value turns out
1139 to be zero and the font had a \type {MathConstants} table, then \type
1140 {AccentBaseHeight} is used instead.
1141
1142 The vertical placement of a bottom accent is straight below the accentee, no
1143 correction takes place.
1144
1145 Possible locations are \type {top}, \type {bottom}, \type {both} and \type
1146 {center}. When no location is given \type {top} is assumed. An additional
1147 parameter \nod {fraction} can be specified followed by a number; a value of for
1148 instance 1200 means that the criterium is 1.2 times the width of the nucleus. The
1149 fraction only applies to the stepwise selected shapes and is mostly meant for the
1150 \type {overlay} location. It also works for the other locations but then it
1151 concerns the width.
1152
1153 \subsection{Radical extensions}
1154
1155 \topicindex {math+radicals}
1156
1157 The new primitive \lpr {Uroot} allows the construction of a radical noad
1158 including a degree field. Its syntax is an extension of \lpr {Uradical}:
1159
1160 \starttyping
1161 \Uradical <fam integer> <char integer> <radicand>
1162 \Uroot <fam integer> <char integer> <degree> <radicand>
1163 \stoptyping
1164
1165 The placement of the degree is controlled by the math parameters \lpr
1166 {Umathradicaldegreebefore}, \lpr {Umathradicaldegreeafter}, and \lpr
1167 {Umathradicaldegreeraise}. The degree will be typeset in \prm
1168 {scriptscriptstyle}.
1169
1170 \subsection{Super- and subscripts}
1171
1172 The character fields in a \LUA|-|loaded \OPENTYPE\ math font can have a \quote
1173 {mathkern} table. The format of this table is the same as the \quote {mathkern}
1174 table that is returned by the \type {fontloader} library, except that all height
1175 and kern values have to be specified in actual scaled points.
1176
1177 When a super- or subscript has to be placed next to a math item, \LUATEX\ checks
1178 whether the super- or subscript and the nucleus are both simple character items.
1179 If they are, and if the fonts of both character items are \OPENTYPE\ fonts (as
1180 opposed to legacy \TEX\ fonts), then \LUATEX\ will use the \OPENTYPE\ math
1181 algorithm for deciding on the horizontal placement of the super- or subscript.
1182
1183 This works as follows:
1184
1185 \startitemize
1186 \startitem
1187 The vertical position of the script is calculated.
1188 \stopitem
1189 \startitem
1190 The default horizontal position is flat next to the base character.
1191 \stopitem
1192 \startitem
1193 For superscripts, the italic correction of the base character is added.
1194 \stopitem
1195 \startitem
1196 For a superscript, two vertical values are calculated: the bottom of the
1197 script (after shifting up), and the top of the base. For a subscript, the two
1198 values are the top of the (shifted down) script, and the bottom of the base.
1199 \stopitem
1200 \startitem
1201 For each of these two locations:
1202 \startitemize
1203 \startitem
1204 find the math kern value at this height for the base (for a subscript
1205 placement, this is the bottom_right corner, for a superscript
1206 placement the top_right corner)
1207 \stopitem
1208 \startitem
1209 find the math kern value at this height for the script (for a
1210 subscript placement, this is the top_left corner, for a superscript
1211 placement the bottom_left corner)
1212 \stopitem
1213 \startitem
1214 add the found values together to get a preliminary result.
1215 \stopitem
1216 \stopitemize
1217 \stopitem
1218 \startitem
1219 The horizontal kern to be applied is the smallest of the two results from
1220 previous step.
1221 \stopitem
1222 \stopitemize
1223
1224 The math kern value at a specific height is the kern value that is specified by the
1225 next higher height and kern pair, or the highest one in the character (if there is no
1226 value high enough in the character), or simply zero (if the character has no math kern
1227 pairs at all).
1228
1229 \subsection{Scripts on extensibles}
1230
1231 \topicindex {math+scripts}
1232 \topicindex {math+delimiters}
1233 \topicindex {math+extensibles}
1234
1235 The primitives \lpr {Uunderdelimiter} and \lpr {Uoverdelimiter} allow the
1236 placement of a subscript or superscript on an automatically extensible item and
1237 \lpr {Udelimiterunder} and \lpr {Udelimiterover} allow the placement of an
1238 automatically extensible item as a subscript or superscript on a nucleus. The
1239 input:
1240
1241 % these produce radical noads .. in fact the code base has the numbers wrong for
1242 % quite a while, so no one seems to use this
1243
1244 \startbuffer
1245 $\Uoverdelimiter 0 "2194 {\hbox{\strut overdelimiter}}$
1246 $\Uunderdelimiter 0 "2194 {\hbox{\strut underdelimiter}}$
1247 $\Udelimiterover 0 "2194 {\hbox{\strut delimiterover}}$
1248 $\Udelimiterunder 0 "2194 {\hbox{\strut delimiterunder}}$
1249 \stopbuffer
1250
1251 \typebuffer will render this:
1252
1253 \blank \startnarrower \getbuffer \stopnarrower \blank
1254
1255 The vertical placements are controlled by \lpr {Umathunderdelimiterbgap}, \lpr
1256 {Umathunderdelimitervgap}, \lpr {Umathoverdelimiterbgap}, and \lpr
1257 {Umathoverdelimitervgap} in a similar way as limit placements on large operators.
1258 The superscript in \lpr {Uoverdelimiter} is typeset in a suitable scripted style,
1259 the subscript in \lpr {Uunderdelimiter} is cramped as well.
1260
1261 These primitives accepts an option \type {width} specification. When used the
1262 also optional keywords \type {left}, \type {middle} and \type {right} will
1263 determine what happens when a requested size can't be met (which can happen when
1264 we step to successive larger variants).
1265
1266 An extra primitive \lpr {Uhextensible} is available that can be used like this:
1267
1268 \startbuffer
1269 $\Uhextensible width 10cm 0 "2194$
1270 \stopbuffer
1271
1272 \typebuffer This will render this:
1273
1274 \blank \startnarrower \getbuffer \stopnarrower \blank
1275
1276 Here you can also pass options, like:
1277
1278 \startbuffer
1279 $\Uhextensible width 1pt middle 0 "2194$
1280 \stopbuffer
1281
1282 \typebuffer This gives:
1283
1284 \blank \startnarrower \getbuffer \stopnarrower \blank
1285
1286 \LUATEX\ internally uses a structure that supports \OPENTYPE\ \quote
1287 {MathVariants} as well as \TFM\ \quote {extensible recipes}. In most cases where
1288 font metrics are involved we have a different code path for traditional fonts end
1289 \OPENTYPE\ fonts.
1290
1291 Sometimes you might want to act upon the size of a delimiter, something that is
1292 not really possible because of the fact that they are calculated {\em after} most
1293 has been typeset already. In the following example the all|-|zero specification
1294 is the trigger to make a fake box with the last delimiter dimensions and shift.
1295 It's an ugly hack but its relative simple and not intrusive implementation has no
1296 side effects. Any other heuristic solution would not satisfy possible demands
1297 anyway. Here is a rather low level example:
1298
1299 \startbuffer
1300 \startformula
1301 \Uleft \Udelimiter 5 0 "222B
1302 \frac{\frac{a}{b}}{\frac{c}{d}}
1303 \Uright \Udelimiter 5 0 "222B
1304 \kern-2\fontcharwd\textfont0 "222B
1305 \mathlimop{\Uvextensible \Udelimiter 0 0 0}_1^2 x
1306 \stopformula
1307 \stopbuffer
1308
1309 \typebuffer
1310
1311 The last line, by passing zero values, results in a fake operator that has the
1312 dimensions of the previous delimiter. We can then backtrack over the (presumed)
1313 width and the two numbers become limit operators. As said, it's not pretty but it
1314 works.
1315
1316 \subsection{Fractions}
1317
1318 \topicindex {math+fractions}
1319
1320 The \prm {abovewithdelims} command accepts keywords \type {exact} and \type
1321 {norule}. When issued the extra space relative to the rule thickness is not
1322 added. One can of course use the \type {\Umathfraction..gap} commands to
1323 influence the spacing. Also the rule is still positioned around the math axis.
1324
1325 \starttyping
1326 $$ { {a} \abovewithdelims() exact 4pt {b} }$$
1327 \stoptyping
1328
1329 The math parameter table contains some parameters that specify a horizontal and
1330 vertical gap for skewed fractions. Of course some guessing is needed in order to
1331 implement something that uses them. And so we now provide a primitive similar to the
1332 other fraction related ones but with a few options so that one can influence the
1333 rendering. Of course a user can also mess around a bit with the parameters
1334 \lpr {Umathskewedfractionhgap} and \lpr {Umathskewedfractionvgap}.
1335
1336 The syntax used here is:
1337
1338 \starttyping
1339 { {1} \Uskewed / <options> {2} }
1340 { {1} \Uskewedwithdelims / () <options> {2} }
1341 \stoptyping
1342
1343 where the options can be \type {noaxis} and \type {exact}. By default we add half
1344 the axis to the shifts and by default we zero the width of the middle character.
1345 For Latin Modern the result looks as follows:
1346
1347 \def\ShowA#1#2#3{$x + { {#1} \Uskewed / #3 {#2} } + x$}
1348 \def\ShowB#1#2#3{$x + { {#1} \Uskewedwithdelims / () #3 {#2} } + x$}
1349
1350 \start
1351 \switchtobodyfont[modern]
1352 \starttabulate[||||||]
1353 \NC \NC
1354 \ShowA{a}{b}{} \NC
1355 \ShowA{1}{2}{} \NC
1356 \ShowB{a}{b}{} \NC
1357 \ShowB{1}{2}{} \NC
1358 \NR
1359 \NC \type{exact} \NC
1360 \ShowA{a}{b}{exact} \NC
1361 \ShowA{1}{2}{exact} \NC
1362 \ShowB{a}{b}{exact} \NC
1363 \ShowB{1}{2}{exact} \NC
1364 \NR
1365 \NC \type{noaxis} \NC
1366 \ShowA{a}{b}{noaxis} \NC
1367 \ShowA{1}{2}{noaxis} \NC
1368 \ShowB{a}{b}{noaxis} \NC
1369 \ShowB{1}{2}{noaxis} \NC
1370 \NR
1371 \NC \type{exact noaxis} \NC
1372 \ShowA{a}{b}{exact noaxis} \NC
1373 \ShowA{1}{2}{exact noaxis} \NC
1374 \ShowB{a}{b}{exact noaxis} \NC
1375 \ShowB{1}{2}{exact noaxis} \NC
1376 \NR
1377 \stoptabulate
1378 \stop
1379
1380 \subsection {Delimiters: \type{\Uleft}, \prm {Umiddle} and \prm {Uright}}
1381
1382 \topicindex {math+delimiters}
1383
1384 Normally you will force delimiters to certain sizes by putting an empty box or
1385 rule next to it. The resulting delimiter will either be a character from the
1386 stepwise size range or an extensible. The latter can be quite differently
1387 positioned than the characters as it depends on the fit as well as the fact if
1388 the used characters in the font have depth or height. Commands like (plain \TEX
1389 s) \type {\big} need use this feature. In \LUATEX\ we provide a bit more control
1390 by three variants that support optional parameters \type {height}, \type {depth}
1391 and \type {axis}. The following example uses this:
1392
1393 \startbuffer
1394 \Uleft height 30pt depth 10pt \Udelimiter "0 "0 "000028
1395 \quad x\quad
1396 \Umiddle height 40pt depth 15pt \Udelimiter "0 "0 "002016
1397 \quad x\quad
1398 \Uright height 30pt depth 10pt \Udelimiter "0 "0 "000029
1399 \quad \quad \quad
1400 \Uleft height 30pt depth 10pt axis \Udelimiter "0 "0 "000028
1401 \quad x\quad
1402 \Umiddle height 40pt depth 15pt axis \Udelimiter "0 "0 "002016
1403 \quad x\quad
1404 \Uright height 30pt depth 10pt axis \Udelimiter "0 "0 "000029
1405 \stopbuffer
1406
1407 \typebuffer
1408
1409 \startlinecorrection
1410 \ruledhbox{\mathematics{\getbuffer}}
1411 \stoplinecorrection
1412
1413 The keyword \type {exact} can be used as directive that the real dimensions
1414 should be applied when the criteria can't be met which can happen when we're
1415 still stepping through the successively larger variants. When no dimensions are
1416 given the \type {noaxis} command can be used to prevent shifting over the axis.
1417
1418 You can influence the final class with the keyword \type {class} which will
1419 influence the spacing. The numbers are the same as for character classes.
1420
1421 \stopsection
1422
1423 \startsection[title={Extracting values}]
1424
1425 \subsection{Codes}
1426
1427 \topicindex {math+codes}
1428
1429 You can extract the components of a math character. Say that we have defined:
1430
1431 \starttyping
1432 \Umathcode 1 2 3 4
1433 \stoptyping
1434
1435 then
1436
1437 \starttyping
1438 [\Umathcharclass1] [\Umathcharfam1] [\Umathcharslot1]
1439 \stoptyping
1440
1441 will return:
1442
1443 \starttyping
1444 [2] [3] [4]
1445 \stoptyping
1446
1447 These commands are provides as convenience. Before they come available you could
1448 do the following:
1449
1450 \starttyping
1451 \def\Umathcharclass{\directlua{tex.print(tex.getmathcode(token.scan_int())[1])}}
1452 \def\Umathcharfam {\directlua{tex.print(tex.getmathcode(token.scan_int())[2])}}
1453 \def\Umathcharslot {\directlua{tex.print(tex.getmathcode(token.scan_int())[3])}}
1454 \stoptyping
1455
1456 \subsection {Last lines}
1457
1458 \topicindex {math+last line}
1459
1460 There is a new primitive to control the overshoot in the calculation of the
1461 previous line in mid|-|paragraph display math. The default value is 2 times
1462 the em width of the current font:
1463
1464 \starttyping
1465 \predisplaygapfactor=2000
1466 \stoptyping
1467
1468 If you want to have the length of the last line independent of math i.e.\ you don't
1469 want to revert to a hack where you insert a fake display math formula in order to
1470 get the length of the last line, the following will often work too:
1471
1472 \starttyping
1473 \def\lastlinelength{\dimexpr
1474 \directlua {tex.sprint (
1475 (nodes.dimensions(node.tail(tex.lists.page_head).list))
1476 )}sp
1477 \relax}
1478 \stoptyping
1479
1480 \stopsection
1481
1482 \startsection[title={Math mode}]
1483
1484 \subsection {Verbose versions of single|-|character math commands}
1485
1486 \topicindex {math+styles}
1487
1488 \LUATEX\ defines six new primitives that have the same function as
1489 \type {^}, \type {_}, \type {$}, and \type {$$}:
1490
1491 \starttabulate[|l|l|]
1492 \DB primitive \BC explanation \NC \NR
1493 \TB
1494 \NC \lpr {Usuperscript} \NC duplicates the functionality of \type {^} \NC \NR
1495 \NC \lpr {Usubscript} \NC duplicates the functionality of \type {_} \NC \NR
1496 \NC \lpr {Ustartmath} \NC duplicates the functionality of \type {$}, % $
1497 when used in non-math mode. \NC \NR
1498 \NC \lpr {Ustopmath} \NC duplicates the functionality of \type {$}, % $
1499 when used in inline math mode. \NC \NR
1500 \NC \lpr {Ustartdisplaymath} \NC duplicates the functionality of \type {$$}, % $$
1501 when used in non-math mode. \NC \NR
1502 \NC \lpr {Ustopdisplaymath} \NC duplicates the functionality of \type {$$}, % $$
1503 when used in display math mode. \NC \NR
1504 \LL
1505 \stoptabulate
1506
1507 The \lpr {Ustopmath} and \lpr {Ustopdisplaymath} primitives check if the current
1508 math mode is the correct one (inline vs.\ displayed), but you can freely intermix
1509 the four mathon|/|mathoff commands with explicit dollar sign(s).
1510
1511 \subsection{Script commands \lpr {Unosuperscript} and \lpr {Unosubscript}}
1512
1513 \topicindex {math+styles}
1514 \topicindex {math+scripts}
1515
1516 These two commands result in super- and subscripts but with the current style (at the
1517 time of rendering). So,
1518
1519 \startbuffer[script]
1520 $
1521 x\Usuperscript {1}\Usubscript {2} =
1522 x\Unosuperscript{1}\Unosubscript{2} =
1523 x\Usuperscript {1}\Unosubscript{2} =
1524 x\Unosuperscript{1}\Usubscript {2}
1525 $
1526 \stopbuffer
1527
1528 \typebuffer[script]
1529
1530 results in \inlinebuffer[script].
1531
1532 \subsection{Allowed math commands in non|-|math modes}
1533
1534 \topicindex {math+text}
1535 \topicindex {text+math}
1536
1537 The commands \prm {mathchar}, and \lpr {Umathchar} and control sequences that are
1538 the result of \prm {mathchardef} or \lpr {Umathchardef} are also acceptable in
1539 the horizontal and vertical modes. In those cases, the \prm {textfont} from the
1540 requested math family is used.
1541
1542 % \startsection[title={Math todo}]
1543 %
1544 % The following items are still todo.
1545 %
1546 % \startitemize
1547 % \startitem
1548 % Pre-scripts.
1549 % \stopitem
1550 % \startitem
1551 % Multi-story stacks.
1552 % \stopitem
1553 % \startitem
1554 % Flattened accents for high characters (maybe).
1555 % \stopitem
1556 % \startitem
1557 % Better control over the spacing around displays and handling of equation numbers.
1558 % \stopitem
1559 % \startitem
1560 % Support for multi|-|line displays using \MATHML\ style alignment points.
1561 % \stopitem
1562 % \stopitemize
1563 %
1564 % \stopsection
1565
1566 \stopsection
1567
1568 \startsection[title={Goodies}]
1569
1570 \subsection {Obsolete}
1571
1572 Per TL 2026 the following primitives are \quote {ignored}. They were not used in
1573 two decades unless by \CONTEXT\ for exploring variants but we can do that
1574 differently now. Obsolete commands that trigger a warning: \typ {\mathoption},
1575 \typ {\mathitalicssmode}, \typ {\mathnolimitsmode}, \typ {\mathscriptboxmode},
1576 \typ {\mathscriptcharmode}, %\typ {\mathflattenmode},
1577 \typ {\mathdefaultsmode},
1578 \typ {\mathrulethicknessmode}, \typ {\mathdelimitersmode}. Also no longer used
1579 are the parameters \typ {\Umathnolimitsupfactor} and \typ
1580 {\Umathnolimitsubfactor}.
1581
1582 \subsection {Flattening: \lpr {mathflattenmode}}
1583
1584 \topicindex {math+flattening}
1585
1586 The \TEX\ math engine collapses \type {ord} noads without sub- and superscripts
1587 and a character as nucleus. and which has the side effect that in \OPENTYPE\ mode
1588 italic corrections are applied (given that they are enabled).
1589
1590 \startbuffer[sample]
1591 \switchtobodyfont[modern]
1592 $V \mathbin{\mathbin{v}} V$\par
1593 $V \mathord{\mathord{v}} V$\par
1594 \stopbuffer
1595
1596 \typebuffer[sample]
1597
1598 This renders as:
1599
1600 \blank \start \mathflattenmode\plusone \getbuffer[sample] \stop \blank
1601
1602 When we set \lpr {mathflattenmode} to 31 we get:
1603
1604 \blank \start \mathflattenmode\numexpr1+2+4+8+16\relax \getbuffer[sample] \stop \blank
1605
1606 When you see no difference, then the font probably has the proper character
1607 dimensions and no italic correction is needed. For Latin Modern (at least till
1608 2018) there was a visual difference. In that respect this parameter is not always
1609 needed unless of course you want efficient math lists anyway.
1610
1611 You can influence flattening by adding the appropriate number to the value of the
1612 mode parameter. The default value is~1.
1613
1614 \starttabulate[|Tc|c|]
1615 \DB mode \BC class \NC \NR
1616 \TB
1617 \NC 1 \NC ord \NC \NR
1618 \NC 2 \NC bin \NC \NR
1619 \NC 4 \NC rel \NC \NR
1620 \NC 8 \NC punct \NC \NR
1621 \NC 16 \NC inner \NC \NR
1622 \LL
1623 \stoptabulate
1624
1625 \subsection {Less Tracing}
1626
1627 \topicindex {math+tracing}
1628
1629 Because there are quite some math related parameters and values, it is possible
1630 to limit tracing. Only when \type {tracingassigns} and|/|or \type
1631 {tracingrestores} are set to~2 or more they will be traced.
1632
1633 % \subsection {Math options with \lpr {mathdefaultsmode}}
1634 %
1635 % This option has been introduced because \LATEX\ developers wanted some of the
1636 % defaults to be different from the ones that were set in stone when we froze
1637 % \LUATEX. The default values are:
1638 %
1639 % \starttabulate[|l|c|c|]
1640 % \DB \BC scanning \BC rendering \NC \NR
1641 % \TB
1642 % \NC radical/root \NC cramped \NC cramped \NC \NR
1643 % \NC under delimiter \NC cramped \NC supstyle \NC \NR
1644 % \NC over delimiter \NC cramped \NC substyle \NC \NR
1645 % \NC delimiter under \NC cramped \NC current \NC \NR
1646 % \NC delimiter over \NC cramped \NC current \NC \NR
1647 % \LL
1648 % \stoptabulate
1649 %
1650 % When \type {\mathdefaultsmode} is larger than zero, we have:
1651 %
1652 % \starttabulate[|l|c|c|]
1653 % \DB \BC scanning \BC rendering \NC \NR
1654 % \TB
1655 % \NC radical/root \NC cramped \NC cramped \NC \NR
1656 % \NC under delimiter \NC substyle \NC substyle \NC \NR
1657 % \NC over delimiter \NC supstyle \NC supstyle \NC \NR
1658 % \NC delimiter under \NC current \NC current \NC \NR
1659 % \NC delimiter over \NC cramped \NC cramped \NC \NR
1660 % \LL
1661 % \stoptabulate
1662 %
1663 % It is outside the scope of this manual to discuss the rationale behind these
1664 % defaults. The zero values date back from the early times. If needed you can
1665 % explicitly set the style in the content argument.
1666
1667 % \subsection {Math options with \lpr {mathoption}}
1668 %
1669 % This command is now obsolete and triggers an error message. It was only meant
1670 % for experiments.
1671
1672 % % even more obsolete:
1673
1674 % The logic in the math engine is rather complex and there are often no universal
1675 % solutions (read: what works out well for one font, fails for another). Therefore
1676 % some variations in the implementation are driven by parameters (modes). In
1677 % addition there is a new primitive \lpr {mathoption} which will be used for
1678 % testing. Don't rely on any option to be there in a production version as they are
1679 % meant for development.
1680 %
1681 % This option was introduced for testing purposes when the math engine got split
1682 % code paths and it forces the engine to treat new fonts as old ones with respect
1683 % to italic correction etc. There are no guarantees given with respect to the final
1684 % result and unexpected side effects are not seen as bugs as they relate to font
1685 % properties. There is currently only one option:
1686 %
1687 % \startbuffer
1688 % \mathoption old 1
1689 % \stopbuffer
1690 %
1691 % The \type {oldmath} boolean flag in the \LUA\ font table is the official way to
1692 % force old treatment as it's bound to fonts. Like with all options we may
1693 % temporarily introduce with this command this feature is not meant for production.
1694
1695 % % obsolete:
1696 %
1697 % \subsubsection {\type {\mathoption noitaliccompensation}}
1698 %
1699 % This option compensates placement for characters with a built|-|in italic
1700 % correction.
1701 %
1702 % \startbuffer
1703 % {\showboxes\int}\quad
1704 % {\showboxes\int_{|}^{|}}\quad
1705 % {\showboxes\int\limits_{|}^{|}}
1706 % \stopbuffer
1707 %
1708 % \typebuffer
1709 %
1710 % Gives (with computer modern that has such italics):
1711 %
1712 % \startlinecorrection[blank]
1713 % \switchtobodyfont[modern]
1714 % \startcombination[nx=2,ny=2,distance=5em]
1715 % {\mathoption noitaliccompensation 0\relax \mathematics{\getbuffer}}
1716 % {\nohyphens\type{0:inline}}
1717 % {\mathoption noitaliccompensation 0\relax \mathematics{\displaymath\getbuffer}}
1718 % {\nohyphens\type{0:display}}
1719 % {\mathoption noitaliccompensation 1\relax \mathematics{\getbuffer}}
1720 % {\nohyphens\type{1:inline}}
1721 % {\mathoption noitaliccompensation 1\relax \mathematics{\displaymath\getbuffer}}
1722 % {\nohyphens\type{1:display}}
1723 % \stopcombination
1724 % \stoplinecorrection
1725
1726 % % obsolete:
1727 %
1728 % \subsubsection {\type {\mathoption nocharitalic}}
1729 %
1730 % When two characters follow each other italic correction can interfere. The
1731 % following example shows what this option does:
1732 %
1733 % \startbuffer
1734 % \catcode"1D443=11
1735 % \catcode"1D444=11
1736 % \catcode"1D445=11
1737 % P( PP PQR
1738 % \stopbuffer
1739 %
1740 % \typebuffer
1741 %
1742 % Gives (with computer modern that has such italics):
1743 %
1744 % \startlinecorrection[blank]
1745 % \switchtobodyfont[modern]
1746 % \startcombination[nx=2,ny=2,distance=5em]
1747 % {\mathoption nocharitalic 0\relax \mathematics{\getbuffer}}
1748 % {\nohyphens\type{0:inline}}
1749 % {\mathoption nocharitalic 0\relax \mathematics{\displaymath\getbuffer}}
1750 % {\nohyphens\type{0:display}}
1751 % {\mathoption nocharitalic 1\relax \mathematics{\getbuffer}}
1752 % {\nohyphens\type{1:inline}}
1753 % {\mathoption nocharitalic 1\relax \mathematics{\displaymath\getbuffer}}
1754 % {\nohyphens\type{1:display}}
1755 % \stopcombination
1756 % \stoplinecorrection
1757
1758 % % obsolete:
1759 %
1760 % \subsubsection {\type {\mathoption useoldfractionscaling}}
1761 %
1762 % This option has been introduced as solution for tracker item 604 for fuzzy cases
1763 % around either or not present fraction related settings for new fonts.
1764
1765 \stopsection
1766
1767 \stopchapter
1768
1769 \stopcomponent