Debian Codegraph @

texmf-dist/doc/latex/base/ltnews33.tex

raw
1 % \iffalse meta-comment
2 %
3 % Copyright (C) 2019-2025
4 % The LaTeX Project and any individual authors listed elsewhere
5 % in this file.
6 %
7 % This file is part of the LaTeX base system.
8 % -------------------------------------------
9 %
10 % It may be distributed and/or modified under the
11 % conditions of the LaTeX Project Public License, either version 1.3c
12 % of this license or (at your option) any later version.
13 % The latest version of this license is in
14 % https://www.latex-project.org/lppl.txt
15 % and version 1.3c or later is part of all distributions of LaTeX
16 % version 2008 or later.
17 %
18 % This file has the LPPL maintenance status "maintained".
19 %
20 % The list of all files belonging to the LaTeX base distribution is
21 % given in the file `manifest.txt'. See also `legal.txt' for additional
22 % information.
23 %
24 % The list of derived (unpacked) files belonging to the distribution
25 % and covered by LPPL is defined by the unpacking scripts (with
26 % extension .ins) which are part of the distribution.
27 %
28 % \fi
29 % Filename: ltnews33.tex
30 %
31 % This is issue 33 of LaTeX News.
32
33 \NeedsTeXFormat{LaTeX2e}[2020-02-02]
34
35 \documentclass{ltnews}
36
37 %%CCC Temporary definitions:
38 \providecommand\Dash {\unskip ---}
39
40
41
42 %% NOTE: Chris' preferred hyphens!
43 %%\showhyphens{parameters}
44 \hyphenation{because para-me-ters}
45
46 \usepackage[T1]{fontenc}
47
48 \usepackage{lmodern,url,hologo}
49
50 \usepackage{csquotes}
51 \usepackage{multicol}
52
53 \providecommand\hook[1]{\texttt{#1}}
54
55 \providecommand\meta[1]{$\langle$\textrm{\itshape#1}$\rangle$}
56 \providecommand\option[1]{\texttt{#1}}
57 \providecommand\env[1]{\texttt{#1}}
58 \providecommand\Arg[1]{\texttt\{\meta{#1}\texttt\}}
59
60
61 \providecommand\eTeX{\hologo{eTeX}}
62 \providecommand\XeTeX{\hologo{XeTeX}}
63 \providecommand\LuaTeX{\hologo{LuaTeX}}
64 \providecommand\pdfTeX{\hologo{pdfTeX}}
65 \providecommand\MiKTeX{\hologo{MiKTeX}}
66 \providecommand\CTAN{\textsc{ctan}}
67 \providecommand\TL{\TeX\,Live}
68 \providecommand\githubissue[2][]{\ifhmode\unskip\fi
69 \quad\penalty500\strut\nobreak\hfill
70 \mbox{\small\slshape(%
71 \href{https://github.com/latex3/latex2e/issues/\getfirstgithubissue#2 \relax}%
72 {github issue#1 #2}%
73 )}%
74 \par\smallskip}
75
76 % simple solution right now (just link to the first issue if there are more)
77 \def\getfirstgithubissue#1 #2\relax{#1}
78
79 \providecommand\sxissue[1]{\ifhmode\unskip\fi
80 \quad\penalty500\strut\nobreak\hfill
81 \mbox{\small\slshape(\url{https://tex.stackexchange.com/#1})}\par}
82
83 \providecommand\gnatsissue[2]{\ifhmode\unskip\fi
84 \quad\penalty500\strut\nobreak\hfill
85 \mbox{\small\slshape(%
86 \href{https://www.latex-project.org/cgi-bin/ltxbugs2html?pr=#1\%2F\getfirstgithubissue#2 \relax}%
87 {gnats issue #1/#2}%
88 )}%
89 \par}
90
91 \let\cls\pkg
92 \providecommand\env[1]{\texttt{#1}}
93 \providecommand\acro[1]{\textsc{#1}}
94
95 \vbadness=1400 % accept slightly empty columns
96
97
98 \makeatletter
99 % maybe not the greatest design but normally we wouldn't have subsubsections
100 \renewcommand{\subsubsection}{%
101 \@startsection {subsubsection}{2}{0pt}{1.5ex \@plus 1ex \@minus .2ex}%
102 {-1em}{\@subheadingfont\colonize}%
103 }
104 \providecommand\colonize[1]{#1:}
105 \makeatother
106
107 \let\finalvspace\vspace % for document layout fixes
108
109
110
111 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
112 \providecommand\tubcommand[1]{}
113 \tubcommand{\input{tubltmac}}
114
115 \publicationmonth{June}
116 \publicationyear{2021} %% --- Draft Version 3p
117
118 \publicationissue{33}
119
120 \begin{document}
121
122 \tubcommand{\addtolength\textheight{4.2pc}} % only for TUB
123
124 \maketitle
125 {\hyphenpenalty=10000 \spaceskip=3.33pt \hbadness=10000 \tableofcontents}
126
127 \setlength\rightskip{0pt plus 3em}
128
129
130 \medskip
131
132
133 \section{Introduction}
134
135 The focus of the June 2021 release is to provide further important
136 building blocks for the future production of
137 reliable tagged PDF output
138 (see \cite{33:blueprint}); these enhancements
139 are discussed in the next two
140 sections.
141
142 Subsequent sections describe quite a number of recent smaller
143 enhancements and fixes. As usual, more detail on individual changes
144 can be found in the \file{changes.txt} files in the distribution
145 and, of course, in the documented sources~\cite{33:source2e}.
146
147
148
149 \section{Extending the hook concept to paragraphs}
150
151 Largely triggered by the need for better control of paragraph text
152 processing, in particular when producing tagged PDF output, we have
153 changed \LaTeX{} so that the kernel gains control both at the start
154 and at the end of each paragraph. This is done in a manner that is (or
155 should be) transparent to both packages and documents.
156
157 Besides the addition of internal control points for the exclusive use
158 of the \LaTeX{} kernel, we also implemented four public hooks that can
159 be used in packages or documents (via the normal hook management
160 declarations) to achieve special effects, etc. Until now, such
161 enhancements required redefinitions of \cs{everypar} or \cs{par},
162 which led to the usual issues since such changes can easily conflict
163 with changes made by other packages.
164
165 The documentation of these new \enquote{paragraph hooks}, together
166 with a few examples, is in \file{ltpara-doc.pdf} and, for those who
167 want to study it, the (quite interesting) code can be found in
168 \file{ltpara-code.pdf}. Additionally, both of these files are included
169 as part of the full kernel documentation in \file{source2e.pdf}.
170
171
172
173 \section{Extending the hook concept to commands}
174
175 Up to now, hook management covered hooks for only a few core areas,
176 such as the hooks for the \cs{shipout} process or those in the
177 \env{document} environment, as well as some \enquote {generic} hooks,
178 both for file loading (helpful for patching such files) and for
179 arbitrary environments (the hooks executed within \cs{begin} and
180 \cs{end}). This concept of \enquote{generic hooks} has now been
181 extended to provide \hook{/before} and \hook{/after} hooks for any
182 (document-level) command\Dash in theory at least.
183
184
185 In practice, these new generic \hook{cmd} hooks, especially the
186 \hook{cmd/.../after}, hooks may fail with commands that are too
187 complex to be automatically patched, breaking if the hook contains any
188 code. These restrictions are documented in
189 \file{ltcmdhooks-doc.pdf}.
190 %
191 However, given that these hooks are mainly meant for developers who
192 wish to provide better interoperability between different packages,
193 and between packages and the \LaTeX{} kernel, these restrictions are,
194 we hope, of minor importance. Indeed, for commands where this
195 mechanism can't be applied, one is in the same situation as before;
196 and for all others there will be a noticeable improvement.
197
198 These hooks will be especially important for our current project to
199 provide accessible and tagged PDF output~\cite{33:blueprint} because
200 we will eventually have to patch many third-party packages, and this
201 must be done in controlled and standardized ways.
202
203
204
205
206 \section{Other hook business}
207
208
209 \subsection{Shipping out a page while bypassing hooks}
210
211 In the 2020 October release, several hooks were added to control the
212 process of constructing and shipping out a page box: these support,
213 for example, the addition of background or foreground material
214 to some or all pages.
215
216
217 We have now added a command, called \cs{RawShipout}, which does not do
218 any rebuilding of the page box and so does not run most of these
219 hooks. When using this new command, essential internal book-keeping
220 is still carried out, such as updating the \texttt{totalpages} counter
221 and adding \texttt{shipout/firstpage} or \texttt{shipout/lastpage}
222 material when appropriate.
223
224
225 \subsection{A new Lua callback in \pkg{ltshipout},
226 for custom attributes}
227
228 For use just before shipping out a page, there is now a \LuaTeX{}
229 callback \texttt{pre\_shipout\_filter} to contain final adjustments to
230 the box being shipped out. This is particularly useful for
231 Lua\TeX\ packages which flag (using, for example, attributes or
232 properties) elements on a page in order to apply effects (such as the
233 insertion of \enquote{color commands}) to these elements at shipout.
234
235
236
237 \section{Improved handling of file names}
238
239 \subsection[File names with spaces, multiple dots or\\
240 \acro{utf-8} characters]
241 {File names with spaces, multiple dots or \acro{utf-8} characters}
242
243 In one of the recent \LaTeX{} releases we improved the interface
244 for specifying file names so that they can now safely contain spaces
245 (as is common these days),
246 more than one dot character, and also \acro{utf-8} characters
247 outside the \acro{ascii} range.
248 In the past this was only possible by applying a special syntax
249 in the case of spaces,
250 while file names with several dots often failed,
251 as did most \acro{utf-8} characters.
252
253
254 \subsubsection{Consequences for file names in \cs{include}}
255
256 \TeX{} has a built-in rule saying that you can normally leave out the
257 extension if it is \texttt{.tex}. Thus \verb=\input{file}= and
258 \verb=\input{file.tex}= both load \file{file.tex} (if it exists).
259 While this is convenient most of the time, it is a little awkward in
260 some scenarios (for example, when both \file{file} and \file{file.tex}
261 exist) and also when you manually try to implement the rule.
262
263 \LaTeX{} therefore had one special syntax for \cs{include} and
264 \cs{includeonly}: they always expected that
265 their arguments contain a
266 file name\footnote{In the case of \cs{includeonly}, a comma-separated list of such names.}
267 with no extension given,
268 so that it had to be \texttt{.tex}. Thus,
269 when you mistakenly wrote
270 \verb=\include{mychap.tex}= (for example,
271 because you changed from \cs{input}
272 to \cs{include}),
273 \LaTeX{} went ahead and looked for the
274 file \file{mychap.tex.tex} for inclusion and tried to
275 use the file \file{mychap.tex.aux} for internal (auxiliary) information. The reason was that
276 \cs{include} had to construct both
277 of these file names from the given
278 argument and it didn't bother to do
279 anything special
280 with the supplied
281 extension \texttt{.tex}.
282
283 With the new implementation this has
284 changed:
285 the extension \texttt{.tex}
286 now gets removed/ignored if it was
287 supplied.
288 Thus \verb=\include{mychap.tex}= now
289 no longer looks for \file{mychap.tex.tex}
290 but loads
291 \file{mychap.tex}
292 and uses \file{mychap.aux}.
293 %
294 \githubissue{486}
295
296
297
298 \subsection{Normalization of robust commands in file names}
299
300 The handling of file names has been modified so that \verb|\string| is
301 applied to normalize robust commands within the file name.
302 Previously, for example, \verb|\input{\sqrt{2}}| would cause
303 \LaTeX\ to loop indefinitely whereas with
304 the new normalization
305 it looks for the file named \verb|sqrt {2}.tex|
306 (and therefore very likely reports ``file not found'').
307 %
308 \githubissue{481}
309
310
311
312
313 \subsection{Fix for \env{filecontents} with \acro{utf-8}
314 chars in the file name}
315
316 Since a few releases back, the \env{filecontents} environment has
317 allowed \acro{utf-8} characters in the file name. There was, however,
318 a bug that would not allow \emph{over}writing a file with \acro{utf-8}
319 characters in its name. This has been fixed and now
320 \env{filecontents} allows any characters in the file name.
321 %
322 \githubissue{415}
323
324
325
326 \section{Updates to the font selection scheme}
327
328
329 \subsection{A new hook in \cs{selectfont}}
330
331 After \cs{selectfont} has changed the font, we now run a hook (\hook{selectfont})
332 so that packages can make final \mbox{adjustments}. This functionality was
333 originally provided by the \pkg{everysel} package but our
334 implementation is slightly different and uses the standard hook
335 management.
336 %
337 \githubissue{444}
338
339
340 \subsection{Change of font series/shape delayed until \cs{selectfont}}
341
342 With the NFSS extensions introduced in 2020, the font series and shape
343 settings can be influenced by changes to the font family. The
344 settings of these two are now therefore delayed until \cs{selectfont}
345 is executed; this avoids unnecessary or incorrect substitutions that
346 may otherwise happen due to the order of declarations.
347 %
348 \githubissue{444}
349
350
351
352 \section{Glyphs, characters \& encodings}
353
354 \subsection{Improved copy\,\&\,paste for \pdfTeX{} documents}
355
356 When compiling with \pdfTeX{}, additional information
357 (from the file \file{glyphtounicode.tex}) is now added automatically
358 to the PDF file in order to improve copying from, and searching in,
359 text.
360
361 In particular, this allows the most common ligatures to be copied as
362 intended from all generated PDF files without the need to explicitly
363 load the package \pkg{cmap}.
364 %
365 \githubissue{465}
366
367
368
369
370
371 \subsection{Support for more Unicode characters}
372
373
374 \LaTeX\ is quite capable of typesetting characters such as
375 \enquote{\d{m}}, but until now it could not access some Unicode
376 characters from the Latin Extended Additional block. This meant that,
377 for example, there were no Unicode mappings for some characters that
378 are used to write Sanskrit words in Latin transliteration (as seen in
379 books about yoga, Buddhist philosophy, etc.).
380 %
381 These characters have now been added so that they can be entered
382 directly instead of using \verb=\d{m}=, etc.
383 %
384 \githubissue{484}
385
386
387
388
389
390 \subsection{More ``dashes'' in encodings \texttt{OT1},
391 \texttt{T1} and \texttt{TU}}
392
393 When pasting in text from external sources, one can encounter these
394 three Unicode characters
395 %
396 \texttt{"2011} (non-breaking hyphen),
397 \texttt{"2012} (figure dash) and
398 \texttt{"2015} (horizontal bar),
399 %
400 in addition to the more common
401 %
402 \texttt{"2013} (en-dash) and \texttt{"2014} (em-dash).
403 %
404 In the past, these first three produced an error message when used
405 with \pdfTeX{} (since they are not available in \texttt{OT1} or
406 \texttt{T1} encoded fonts). They now typeset an approximation to the
407 glyph: e.g., the ``figure dash'' is approximated by an en-dash.
408
409 With Unicode engines they either work (when the glyph is contained in
410 the selected Unicode font) or they typeset nothing, producing a
411 ``Missing character'' warning in the log file.
412
413 With all engines these characters can also now be accessed using the
414 command names \cs{textnonbreakinghyphen}, \cs{textfiguredash} and
415 \cs{texthorizontalbar}, respectively.
416 %
417 \githubissue{404}
418
419
420 \subsection{Poor man's \cs{textasteriskcentered}}
421
422 The \cs{textasteriskcentered} symbol, used as part of the set of
423 footnote symbols in \LaTeX{}, is assumed to be implemented by every
424 font with the \texttt{TS1} encoding (when \pdfTeX{} is used) or with
425 the \texttt{TU} encoding for the Unicode engines. That assumption is
426 unfortunately not correct for all fonts since, for example, the
427 \texttt{stix2} fonts don't provide this glyph. A result is that one
428 gets missing glyph messages when using \cs{thanks}, etc.
429
430 Therefore \cs{textasteriskcentered} now checks whether there is such a
431 glyph and, if not, uses a normal \enquote{*}, but slightly enlarged
432 and lowered. This may not be perfect in all cases, but it is
433 certainly better than no glyph showing up.
434 %
435 \githubissue{502}
436
437 \subsection {The characters from \pkg{textcomp} are in the kernel}
438
439 A couple of releases back, the functionality of the \pkg{textcomp}
440 package was integrated into the \LaTeX{} kernel. Thus it is no longer
441 necessary to load this package in order to access glyphs such as
442 \cs{textcopyright}, \cs{texteuro} or \cs{textyen}.
443
444
445 At this time the opportunity was also taken to bring some order to the
446 chaos surrounding the question: \enquote{which glyphs from the
447 \texttt{TS1} encoding are available in a given font?}. This was
448 done using an approach based on font families and collections, with
449 the differing glyph coverage of the \enquote*{text symbols} being
450 indicated by assigning to a font family or collection a ``sub-encoding
451 number'' that indicates which glyphs from the \texttt{TS1} encoding are
452 guaranteed to be available when using a font from that family or
453 collection. This assignment ensures that \LaTeX{} always errs on the
454 side of caution, possibly claiming that a glyph is not available even
455 when it in fact is.
456
457 \iffalse %%FMi but drop that
458
459 The documented code for this can be found now in the file
460 \file{lttextcomp.dtx} but we hope to publish a full explanation of
461 the approach very soon now.
462
463 \fi %% FMi potential drop
464
465
466
467 \subsubsection
468 [A note on the history of ``text symbols'']
469 {A note on the history of ``text symbols'' and
470 the \texttt{TS1} encoding}
471
472
473 The \enquote{text symbol encoding} (\texttt{TS1}) was originally
474 designed at the Cork Conference as a companion to the \texttt{T1}
475 encoding. In it various symbols that are not subject to hyphenation
476 got assembled and the \pkg{textcomp} package was developed to make
477 them accessible. Unfortunately the \TeX{} community was a bit too
478 enthusiastic and included several symbols only available in a few
479 \TeX{} fonts and some, such as the capital accents, not available at
480 all but developed as part of the reference font implementation.
481
482 In hindsight that was a very bad idea because it meant that other
483 existing fonts (at the time) and later new fonts that got developed
484 were unable to provide the full set of glyphs that made up the
485 \texttt{TS1} encoding. For existing free PostScript fonts people
486 took the extra effort and produced virtual fonts that faked (some) of
487 the missing glyphs. But this was and is a time-consuming effort so it
488 was done for only a few basic fonts. But even then, only some fonts
489 included all glyphs from \texttt{TS1} so the \pkg{textcomp} already
490 back then contained a long list, dividing fonts into 5 categories
491 according to which glyphs were implemented and which were missing.
492
493
494 When we recently integrated the functionality of the \pkg{textcomp}
495 into the \LaTeX{} kernel
496 many new free fonts had appeared and
497 unfortunately the chaos around the question \enquote{which glyphs of
498 the \texttt{TS1} encoding are implemented by which font} had
499 increased with it. Not only did one find many new holes, it was next to
500 impossible to order the set of fonts into a reasonable set of
501 sub-encodings that are contained in each other in a single sequence.
502
503 In the end we decided on nine or ten sub-encodings with a reasonable
504 number of fonts in each so that all fonts implemented all glyphs of the
505 sub-encoding they got mapped to. Thus when typesetting with a font one
506 could be sure that a command like \cs{textcopyleft} would either
507 typeset the requested character (if the glyph was part of the
508 sub-encoding the font belonged to) or it would raise an error, saying
509 that the glyph is unavailable in that font. The mapping would ensure
510 that \LaTeX{} always errs on the side of caution, because it might
511 claim a glyph is unavailable even though in fact it is.\looseness-1
512
513 For example, the old \texttt{pcr} (PostScript Courier) font (as well
514 as most other older PS fonts) is mapped to sub-encoding 5 and
515 therefore claims that \cs{textasciigrave} is unavailable even though
516 in fact for Courier this is not true. If one uses such a font and this
517 becomes an issue then there are a couple (suboptimal) possibilities.
518 For one, one can alter the mapping of Courier and pretend that belongs
519 to a fuller sub-encoding, e.g.
520 \begin{verbatim}
521 \DeclareEncodingSubset{TS1}{pcr}{2}
522 \end{verbatim}
523 The downside is, that \LaTeX{} then believes other glyphs that are in fact
524 unavailable are also there, so that it is important to check that the
525 final document doesn't have some missing glyphs.
526
527 An alternative is to pretend that \cs{textasciigrave} can always be
528 taken from the \texttt{TS1} encoding (no questions asked):
529 \begin{verbatim}
530 \DeclareTextSymbolDefault{\textasciigrave}{TS1}
531 \end{verbatim}
532 Again there is a danger that this is not true when it is used with a
533 different font and would then generate a missing glyph.
534
535 Finally, and possibly the best solution, if not impossible for other
536 reasons, is to simply use a different font, for example, to use the
537 \TeX{} Gyre Cursor font (a reimplementation of Courier with a
538 much more complete glyph set).
539
540
541
542
543 \section{New or improved commands}
544
545 \subsection{Adjusting \env{itemize} labels with \cs{labelitemfont}}
546
547 The command \cs{labelitemfont} was introduced already with the
548 \LaTeX\ release 2020-02-02, but back then we forgot to describe it, so
549 we do this now. Its purpose is to resolve some bad formatting issues
550 with the \env{itemize} environment and also to make it easier to
551 adjust the layout when necessary. What could happen in the past was
552 that the \env{itemize} labels (e.g., the \textbullet{}) would
553 sometimes react to surrounding font changes and could then suddenly
554 change shape, for example to \textit{\textbullet}.
555
556 This new command \cs{labelitemfont}, which defaults to \cs{normalfont},
557 can be used to provide additional control in the typesetting of
558 each label. Thus by choosing
559 different settings other effects can be achieved. Here are two
560 examples:
561 \begin{verbatim}
562 \renewcommand\labelitemfont
563 {\normalfont\fontfamily{lmss}\selectfont}
564 \renewcommand\labelitemfont
565 {\rmfamily\normalshape}
566 \end{verbatim}
567 The first definition will take the symbols from the font Latin Modern
568 Sans, so that you get
569 %
570 \def\myfont#1{{\let\labelitemfont\empty\fontfamily{lmss}\selectfont#1}}%
571 %
572 \myfont\labelitemi, \myfont\labelitemii, \myfont\labelitemiii\ and
573 \!\!\myfont\labelitemiv\,; while the second variant freezes the font family
574 and shape, but leaves the series as a variable quantity, so that an
575 \env{itemize} in a bold context would show bolder symbols. Making
576 \cs{labelitemfont} empty would give you back the buggy old behavior.
577 %
578 \githubissue{497}
579
580
581 \subsection{Producing several marks for one footnote}
582
583 It is sometimes necessary to reference the same footnote several
584 times: i.e., to produce several footnote marks using the same number
585 or symbol. This is now easily possible by placing a \cs{label} within
586 the referenced \cs{footnote} and referencing this label by using the
587 new command \cs{footref}. This means that footnote marks can be
588 generated to refer to arbitrary footnotes (including those in
589 \env{minipage}s).
590
591 This \cs{footref} command has previously been available, but only when
592 using certain classes or the \pkg{footmisc} package.
593 %
594 \githubissue{482}
595
596
597 \subsection{Allow \cs{nocite} in the preamble}
598
599 A natural place for \verb=\nocite{*}= would be the preamble of the
600 document, but for historical reasons \LaTeX{} issued an error message
601 if it was placed there. This command is now allowed in the preamble.
602 %
603 \githubissue{424}
604
605
606 \subsection{Made \cs{\textbackslash} generally robust}
607
608 In 2018 most \LaTeX{} user-level commands were made robust, including
609 the \cs{\textbackslash} command. However, \cs{\textbackslash} gets
610 redefined in various environments and not all these cases were caught:
611 such as, in particular, its use as the row delimiter in \env{tabular}
612 structures. This has been corrected so that \cs{\textbackslash}
613 should now be robust in all circumstances.
614
615
616 This change also fixed one anomaly present in the past:
617 in a tabular preamble of the form
618 \finalvspace*{-.3\baselineskip}
619 \begin{quote}
620 \hspace*{-.15em}\verb={l=\texttt{\string>}\verb={\raggedright}p{10cm}r}= % stupid class
621 \end{quote}
622 \finalvspace*{-.3\baselineskip}
623 a \cs{\textbackslash} in the second column would have the definition
624 used within \cs{raggedright} and so it would not indicate the
625 (premature) end of the \env{tabular}. Thus, for example,%
626 \finalvspace*{-.3\baselineskip}
627 \begin{quote}
628 \verb=a & b1 \\ b2 & c \\=
629 \end{quote}
630 \finalvspace*{-.3\baselineskip}
631 was interpreted as a single row of the \env{tabular} (as intended),
632 whereas
633 \finalvspace*{-.3\baselineskip}
634 \begin{quote}
635 \verb=a & \\ b2 & c \\=
636 \end{quote}
637 \finalvspace*{-.3\baselineskip}
638 resulted in two rows! This happened because the \cs{\textbackslash}
639 directly following the \verb=&= got interpreted while it still had the
640 \enquote{end the row} meaning and not yet the \enquote{start a new
641 line within the second column} meaning.
642
643 With \cs{\textbackslash} now being robust, the special scanning mode
644 initiated by the \verb=&= ends immediately when this command is seen:
645 the second column is therefore then started, which results in the
646 \cs{\textbackslash} being interpreted as being within that column and
647 hence as having its expected, within-column, meaning.
648
649 We have restored consistency here: now both of the above lines
650 produce a single \env{tabular} row.
651 %
652 As before, you can
653 put \cs{raggedright}\cs{arraybackslash} in the \env{tabular}'s
654 preamble for a column to ensure that \cs{\textbackslash} is always
655 interpreted as a tabular row separator when used in that column. And
656 you can use \cs{tabularnewline} to explicitly ask for a new table row,
657 even when \cs{\textbackslash} has a different meaning within the
658 current column.
659 %
660 \githubissue{548}
661
662
663
664
665 \subsection{Allow extra space between name and address in \pkg{letter} class}
666
667 The \cs{opening} command in the \pkg{letter} class expects the name
668 and address to be separated by \verb=\\=, but it didn't allow the use
669 of an optional argument to add some extra space after the name. The
670 code has now been slightly altered to allow this.
671 %
672 \githubissue{427}
673
674
675
676 \subsection{Additions to \cs{tracingall}}
677
678 In July 2020 David Jones suggested an extension to \TeX{} engines,
679 that added the possibility to set \cs{tracinglostchars}\texttt{=3} in
680 order to generate an error message in case some character is missing
681 from a font. In previous years, a warning about a missing character
682 was silently printed to the \texttt{.log} file\linebreak
683 (if $\cs{tracinglostchars}>0$) and to the terminal\linebreak
684 (if ${}>1$). This extension was added for \TL{} and \MiKTeX{}
685 (except in Knuth's \TeX, of course),
686 so that with $\cs{tracinglostchars}>2$ you now also get an
687 error message for each missing glyph.
688
689 Later, in January 2021, Petr Olšák suggested yet another extension:
690 a new primitive
691 \mbox{parameter}
692 %
693 \cs{tracingstacklevels} that, when both it and \cs{tracingmacros} are
694 positive, will add to the \mbox{tracing} information for each
695 macro a visual indication (using dots) of
696 its nesting level in the macro expansion stack.
697
698 These changes have both now been added to \LaTeX's debugging macros
699 \cs{tracingall} and \cs{tracingnone}, so that these two new extensions
700 are activated/deactivated as appropriate, so long as the \TeX{} engine
701 supports them. An example document demonstrating these parameters is
702 in the linked GitHub issue.
703 %
704 \githubissue{524}
705
706
707
708
709 \section{Code improvements}
710
711 \subsection{Execute \cs{par} at the end of \cs{marginpar}}
712
713 Previously, \LaTeX{} ended a \cs{marginpar} without ever explicitly
714 calling \cs{par}. This command is now explicitly added because it is
715 essential to the correct working of the paragraph hooks.
716
717 Another case where this issue caused problems was the \pkg{lineno}
718 package, where the last line was not numbered if the \cs{marginpar}
719 ended without an explicit \cs{par}.
720 %
721 \githubissue{489}
722
723
724
725 \subsection{Execute \cs{AtEndDocument} hook in vertical mode}
726
727 Until now \verb=\end{document}= executed the code from the
728 \cs{AtEndDocument} hook as its first action. This meant that this
729 hook was executed in horizontal mode if the user left no empty line
730 after the last paragraph. As a result, one could get a spurious space
731 added when, for example, that code contained a \cs{write}
732 statement. This was fixed and now \cs{enddocument} first issues a
733 \cs{par} to ensure that it always goes into vertical mode.
734 %
735 \githubissue{385}
736
737
738 \subsection{Color groups made permanent}
739
740 The use of color in certain \LaTeX{} constructs, especially boxes,
741 needs an extra layer of grouping to ensure that the color setting does
742 not \emph{escape} and continue outside the box when it shouldn't.
743 %
744 To support this, the \LaTeX{} kernel defines a number of commands,
745 e.g., \cs{color@begingroup} to be used in such places.
746
747
748 Until now, these commands were initially set as no-ops and only the
749 color packages redefined them to become real groups; this methodology
750 complicates the coding as one has to account for a group being present
751 or not (\mbox{depending} on what is loaded in the document).
752 %
753 The kernel therefore now permanently adds these \enquote{color groups}.
754 %
755 \githubissue{488}
756
757
758 \subsection{Provide the raw option list to key/value option handlers}
759
760 Before any further processing of the option list, the original
761 (un-normalized, \enquote{raw} and unchanged) list of package or class
762 options is now saved, as \cs{@raw@opt@...}; this list is not used by
763 the standard option processing code but it is now available for use by
764 extended class/package processing systems. Note that, for
765 compatibility reasons, the standard option processing code has not
766 been changed.
767
768
769
770 One
771 aspect of this
772 change does affect the standard \mbox{processing}: any tokens to the
773 right of an \texttt{=}
774 sign are \mbox{removed}
775 from consideration
776 when constructing
777 the \enquote{\mbox{unused} option list}.
778 For example, in
779 this release \texttt{clip=true} and
780 \texttt{clip=false} both contribute \texttt{clip} to the list of
781 options that have been used.
782 %
783 \githubissue{85}
784
785
786 \subsection{New for \pkg{latexrelease}\,: \cs{NewModuleRelease}}
787
788 To explain the need for this new feature, we shall consider the
789 following example: in the 2020-10-01 release, \LaTeX's new hook
790 management system was added to the kernel (see \cite{33:ltnews32})
791 and, as with all changes to the kernel, it was added to
792 \pkg{latexrelease}; this made it possible to roll back to a date where
793 this module didn't yet exist, or to roll forward from an older
794 \LaTeX{} release to get the hook management system (by loading the
795 \pkg{latexrelease} package).
796 %
797 However, this method of rolling back from a later release to the
798 2020-10-01 release didn't quite work because it would try to define
799 all the commands from \pkg{lthooks} again; and this would of course
800 result in the expected errors from commands defined with
801 \cs{newcommand} or (as in \pkg{lthooks}) \cs{cs\_new:Npn}.
802
803 To solve such issues, we now provide \cs{NewModuleRelease} so that
804 completely new modules can be defined using the facilities of
805 \pkg{latexrelease} in such a way that, when rolling back or forward,
806 the system will know whether the code of the new module has to be read
807 or completely ignored. More details on this can be found in the
808 \pkg{latexrelease} documentation (get this with
809 \verb|texdoc latexrelease|).
810 %
811 \githubissue{479}
812
813
814 \subsection{Small fix for rolling back prior to 2020-02-02}
815
816 Whereas the \pkg{latexrelease} package can usually emulate an older
817 \LaTeX{} kernel without much problem, rolling back to before the
818 2020-02-02 release didn't work properly: this is because the
819 management of the \cs{ExplSyntaxOn/Off} status for packages (after an
820 \pkg{expl3}-based package is loaded) cannot be removed by the rollback
821 without messing up the catcodes. This has been fixed so that rollback
822 is now more careful not to leave \cs{ExplSyntaxOn} after a package
823 ends.
824 %
825 \githubissue{504}
826
827
828
829 \section{Changes to packages in the \pkg{graphics} category}
830
831
832 \subsection{Removed warning when loading graphics files}
833
834 A previous release sometimes mistakenly caused a (false) warning
835 message to appear when \mbox{using} a generic graphics rule to find
836 and load a graphics file with an unknown \mbox{extension}.
837 %%CCC removing hyphenation here makes this one line longer.
838 This warning would incorrectly say that the file was not found,
839 whereas the file would in fact be correctly loaded. The warning now
840 doesn't show up in that case.
841 %
842 \githubissue{516}
843
844
845 \subsection{Fixed loading of \texttt{gzip}ped
846 PostScript files}
847
848 A previous release mistakenly changed the file searching mechanism so
849 that compressed PostScript graphics files would raise an error when
850 being loaded with \cs{includegraphics}. This has been fixed so that
851 \texttt{gzip}ped graphics files now load correctly.
852 %
853 \githubissue{519}
854
855
856
857
858 \section{Changes to packages in the \pkg{tools} category}
859
860 \subsection{\pkg{layout}: Added language options}
861
862 This package now recognizes \option{japanese} and \option{romanian} as
863 language options.
864 %
865 \githubissue[s]{353 and 529}
866
867
868 \subsection{\pkg{array} and \pkg{longtable}: Make \cs{\textbackslash} generally robust}
869
870 The fix for this issue was also applied to these packages; see above.
871 %
872 \githubissue{548}
873
874
875 \subsection{\pkg{longtable}: General bug fix update}
876
877 This is a minor update to the \pkg{longtable} package that fixes several
878 reported bugs: notably the possibility of incorrect page breaks when
879 floats appear on the page where a \env{longtable} starts. As this may
880 affect page breaking in existing documents, a rollback to
881 \pkg{longtable 4.13} (\file{longtable-2020-01-07.sty}) is supported.
882 %
883 \gnatsissue{tools}{2914 3396 3512}
884 \githubissue{133 137 183 464 561}
885
886
887 \subsection{\pkg{trace}: Additions to \cs{traceon}}
888
889 The \cs{tracingstacklevels} and \cs{tracinglostchars} extensions to
890 \cs{tracingall} (see above) were also added to \cs{traceon} in the
891 \pkg{trace} package, so its users can also benefit from these new
892 debugging possibilities.
893 %
894 \githubissue{524}
895
896
897 \subsection{\pkg{bm}: Better support for commands with optional arguments}
898
899 Some uses of optional arguments in \cs{bm} stopped being supported (in
900 2004) when \cs{kernel@ifnextchar} was used internally by the format
901 instead of \cs{@ifnextchar}. This update handles both versions of this
902 command and restores the original \mbox{behavior}.
903
904 In addition, package options for guiding the use of \enquote{poor
905 man's bold} in fallback situations were added.
906 %
907 \githubissue{554}
908
909
910
911 \section{Changes to packages in the \pkg{amsmath} category}
912
913 The fix for issue 548 was also applied in \pkg{amsmath}; see above.
914 %
915 \githubissue{548}
916
917
918
919
920 \medskip
921
922 \begin{thebibliography}{9}
923
924 \fontsize{9.3}{11.3}\selectfont
925
926 \bibitem{33:blueprint} Frank Mittelbach and Chris Rowley:
927 \emph{\LaTeX{} Tagged PDF \Dash A blueprint for a large project}.\\
928 \url{https://latex-project.org/publications/indexbyyear/2020/}
929
930 \bibitem{33:source2e}
931 \emph{\LaTeX{} documentation on the \LaTeX{} Project Website}.\\
932 \url{https://latex-project.org/help/documentation/}
933
934 \bibitem{33:ltnews32} \LaTeX{} Project Team:
935 \emph{\LaTeXe{} news 32}.\\
936 \url{https://latex-project.org/news/latex2e-news/ltnews32.pdf}
937
938 \end{thebibliography}
939
940
941
942 \end{document}