]>
git.cameronkatri.com Git - mandoc.git/blob - term.c
1 /* $Id: term.c,v 1.163 2010/07/21 20:35:03 kristaps Exp $ */
3 * Copyright (c) 2008, 2009, 2010 Kristaps Dzonsons <kristaps@bsd.lv>
4 * Copyright (c) 2010 Ingo Schwarze <schwarze@openbsd.org>
6 * Permission to use, copy, modify, and distribute this software for any
7 * purpose with or without fee is hereby granted, provided that the above
8 * copyright notice and this permission notice appear in all copies.
10 * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
11 * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
12 * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
13 * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
14 * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
15 * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
16 * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
22 #include <sys/types.h>
37 static void spec(struct termp
*, enum roffdeco
,
38 const char *, size_t);
39 static void res(struct termp
*, const char *, size_t);
40 static void buffera(struct termp
*, const char *, size_t);
41 static void bufferc(struct termp
*, char);
42 static void adjbuf(struct termp
*p
, size_t);
43 static void encode(struct termp
*, const char *, size_t);
47 term_free(struct termp
*p
)
53 chars_free(p
->symtab
);
60 term_begin(struct termp
*p
, term_margin head
,
61 term_margin foot
, const void *arg
)
72 term_end(struct termp
*p
)
80 term_alloc(enum termenc enc
)
84 p
= calloc(1, sizeof(struct termp
));
96 * Flush a line of text. A "line" is loosely defined as being something
97 * that should be followed by a newline, regardless of whether it's
98 * broken apart by newlines getting there. A line can also be a
99 * fragment of a columnar list (`Bl -tag' or `Bl -column'), which does
100 * not have a trailing newline.
102 * The following flags may be specified:
104 * - TERMP_NOLPAD: when beginning to write the line, don't left-pad the
105 * offset value. This is useful when doing columnar lists where the
106 * prior column has right-padded.
108 * - TERMP_NOBREAK: this is the most important and is used when making
109 * columns. In short: don't print a newline and instead pad to the
110 * right margin. Used in conjunction with TERMP_NOLPAD.
112 * - TERMP_TWOSPACE: when padding, make sure there are at least two
113 * space characters of padding. Otherwise, rather break the line.
115 * - TERMP_DANGLE: don't newline when TERMP_NOBREAK is specified and
116 * the line is overrun, and don't pad-right if it's underrun.
118 * - TERMP_HANG: like TERMP_DANGLE, but doesn't newline when
119 * overruning, instead save the position and continue at that point
120 * when the next invocation.
122 * In-line line breaking:
124 * If TERMP_NOBREAK is specified and the line overruns the right
125 * margin, it will break and pad-right to the right margin after
126 * writing. If maxrmargin is violated, it will break and continue
127 * writing from the right-margin, which will lead to the above scenario
128 * upon exit. Otherwise, the line will break at the right margin.
131 term_flushln(struct termp
*p
)
133 int i
; /* current input position in p->buf */
134 size_t vis
; /* current visual position on output */
135 size_t vbl
; /* number of blanks to prepend to output */
136 size_t vend
; /* end of word visual position on output */
137 size_t bp
; /* visual right border position */
138 int j
; /* temporary loop index for p->buf */
139 int jhy
; /* last hyph before overflow w/r/t j */
140 size_t maxvis
; /* output position of visible boundary */
141 size_t mmax
; /* used in calculating bp */
144 * First, establish the maximum columns of "visible" content.
145 * This is usually the difference between the right-margin and
146 * an indentation, but can be, for tagged lists or columns, a
147 * small set of values.
150 assert(p
->offset
< p
->rmargin
);
152 maxvis
= (int)(p
->rmargin
- p
->offset
) - p
->overstep
< 0 ?
154 0 : p
->rmargin
- p
->offset
- p
->overstep
;
155 mmax
= (int)(p
->maxrmargin
- p
->offset
) - p
->overstep
< 0 ?
157 0 : p
->maxrmargin
- p
->offset
- p
->overstep
;
159 bp
= TERMP_NOBREAK
& p
->flags
? mmax
: maxvis
;
162 * Indent the first line of a paragraph.
164 vbl
= p
->flags
& TERMP_NOLPAD
? 0 : p
->offset
;
168 while (i
< (int)p
->col
) {
170 * Handle literal tab characters: collapse all
171 * subsequent tabs into a single huge set of spaces.
173 for (j
= i
; j
< (int)p
->col
; j
++) {
174 if ('\t' != p
->buf
[j
])
176 vend
= (vis
/ p
->tabwidth
+ 1) * p
->tabwidth
;
182 * Count up visible word characters. Control sequences
183 * (starting with the CSI) aren't counted. A space
184 * generates a non-printing word, which is valid (the
185 * space is printed according to regular spacing rules).
189 for (jhy
= 0; j
< (int)p
->col
; j
++) {
190 if ((j
&& ' ' == p
->buf
[j
]) || '\t' == p
->buf
[j
])
193 /* Back over the the last printed character. */
194 if (8 == p
->buf
[j
]) {
196 vend
-= (*p
->width
)(p
, p
->buf
[j
- 1]);
201 /* Break at the hyphen point if we overrun. */
202 if (vend
> vis
&& vend
< bp
&&
203 ASCII_HYPH
== p
->buf
[j
])
206 vend
+= (*p
->width
)(p
, p
->buf
[j
]);
210 * Find out whether we would exceed the right margin.
211 * If so, break to the next line.
213 if (vend
> bp
&& 0 == jhy
&& vis
> 0) {
216 if (TERMP_NOBREAK
& p
->flags
) {
217 p
->viscol
= p
->rmargin
;
218 (*p
->advance
)(p
, p
->rmargin
);
219 vend
+= p
->rmargin
- p
->offset
;
225 /* Remove the p->overstep width. */
227 bp
+= (int)/* LINTED */
233 * Skip leading tabs, they were handled above.
235 while (i
< (int)p
->col
&& '\t' == p
->buf
[i
])
238 /* Write out the [remaining] word. */
239 for ( ; i
< (int)p
->col
; i
++) {
240 if (vend
> bp
&& jhy
> 0 && i
> jhy
)
242 if ('\t' == p
->buf
[i
])
244 if (' ' == p
->buf
[i
]) {
245 while (' ' == p
->buf
[i
]) {
246 vbl
+= (*p
->width
)(p
, p
->buf
[i
]);
251 if (ASCII_NBRSP
== p
->buf
[i
]) {
252 vbl
+= (*p
->width
)(p
, ' ');
257 * Now we definitely know there will be
258 * printable characters to output,
259 * so write preceding white space now.
262 (*p
->advance
)(p
, vbl
);
267 if (ASCII_HYPH
== p
->buf
[i
]) {
268 (*p
->letter
)(p
, '-');
269 p
->viscol
+= (*p
->width
)(p
, '-');
271 (*p
->letter
)(p
, p
->buf
[i
]);
272 p
->viscol
+= (*p
->width
)(p
, p
->buf
[i
]);
282 if ( ! (TERMP_NOBREAK
& p
->flags
)) {
288 if (TERMP_HANG
& p
->flags
) {
289 /* We need one blank after the tag. */
290 p
->overstep
= /* LINTED */
291 vis
- maxvis
+ (*p
->width
)(p
, ' ');
294 * Behave exactly the same way as groff:
295 * If we have overstepped the margin, temporarily move
296 * it to the right and flag the rest of the line to be
298 * If we landed right at the margin, be happy.
299 * If we are one step before the margin, temporarily
300 * move it one step LEFT and flag the rest of the line
303 if (p
->overstep
>= -1) {
304 assert((int)maxvis
+ p
->overstep
>= 0);
306 maxvis
+= p
->overstep
;
310 } else if (TERMP_DANGLE
& p
->flags
)
314 if (maxvis
> vis
+ /* LINTED */
315 ((TERMP_TWOSPACE
& p
->flags
) ?
316 (*p
->width
)(p
, ' ') : 0)) {
317 p
->viscol
+= maxvis
- vis
;
318 (*p
->advance
)(p
, maxvis
- vis
);
319 vis
+= (maxvis
- vis
);
320 } else { /* ...or newline break. */
322 p
->viscol
= p
->rmargin
;
323 (*p
->advance
)(p
, p
->rmargin
);
329 * A newline only breaks an existing line; it won't assert vertical
330 * space. All data in the output buffer is flushed prior to the newline
334 term_newln(struct termp
*p
)
337 p
->flags
|= TERMP_NOSPACE
;
338 if (0 == p
->col
&& 0 == p
->viscol
) {
339 p
->flags
&= ~TERMP_NOLPAD
;
343 p
->flags
&= ~TERMP_NOLPAD
;
348 * Asserts a vertical space (a full, empty line-break between lines).
349 * Note that if used twice, this will cause two blank spaces and so on.
350 * All data in the output buffer is flushed prior to the newline
354 term_vspace(struct termp
*p
)
364 spec(struct termp
*p
, enum roffdeco d
, const char *word
, size_t len
)
369 rhs
= chars_spec2str(p
->symtab
, word
, len
, &sz
);
372 else if (DECO_SSPECIAL
== d
)
373 encode(p
, word
, len
);
378 res(struct termp
*p
, const char *word
, size_t len
)
383 rhs
= chars_res2str(p
->symtab
, word
, len
, &sz
);
390 term_fontlast(struct termp
*p
)
395 p
->fontl
= p
->fontq
[p
->fonti
];
396 p
->fontq
[p
->fonti
] = f
;
401 term_fontrepl(struct termp
*p
, enum termfont f
)
404 p
->fontl
= p
->fontq
[p
->fonti
];
405 p
->fontq
[p
->fonti
] = f
;
410 term_fontpush(struct termp
*p
, enum termfont f
)
413 assert(p
->fonti
+ 1 < 10);
414 p
->fontl
= p
->fontq
[p
->fonti
];
415 p
->fontq
[++p
->fonti
] = f
;
420 term_fontq(struct termp
*p
)
423 return(&p
->fontq
[p
->fonti
]);
428 term_fonttop(struct termp
*p
)
431 return(p
->fontq
[p
->fonti
]);
436 term_fontpopq(struct termp
*p
, const void *key
)
439 while (p
->fonti
>= 0 && key
!= &p
->fontq
[p
->fonti
])
441 assert(p
->fonti
>= 0);
446 term_fontpop(struct termp
*p
)
455 * Handle pwords, partial words, which may be either a single word or a
456 * phrase that cannot be broken down (such as a literal string). This
457 * handles word styling.
460 term_word(struct termp
*p
, const char *word
)
462 const char *sv
, *seq
;
469 if (word
[0] && '\0' == word
[1])
486 if ( ! (TERMP_IGNDELIM
& p
->flags
))
487 p
->flags
|= TERMP_NOSPACE
;
493 if ( ! (TERMP_NOSPACE
& p
->flags
)) {
494 if ( ! (TERMP_KEEP
& p
->flags
)) {
495 if (TERMP_PREKEEP
& p
->flags
)
496 p
->flags
|= TERMP_KEEP
;
498 if (TERMP_SENTENCE
& p
->flags
)
501 bufferc(p
, ASCII_NBRSP
);
504 if ( ! (p
->flags
& TERMP_NONOSPACE
))
505 p
->flags
&= ~TERMP_NOSPACE
;
507 p
->flags
&= ~TERMP_SENTENCE
;
510 if ((ssz
= strcspn(word
, "\\")) > 0)
511 encode(p
, word
, ssz
);
518 sz
= a2roffdeco(&deco
, &seq
, &ssz
);
521 case (DECO_RESERVED
):
526 case (DECO_SSPECIAL
):
527 spec(p
, deco
, seq
, ssz
);
530 term_fontrepl(p
, TERMFONT_BOLD
);
533 term_fontrepl(p
, TERMFONT_UNDER
);
536 term_fontrepl(p
, TERMFONT_NONE
);
538 case (DECO_PREVIOUS
):
546 if (DECO_NOSPACE
== deco
&& '\0' == *word
)
547 p
->flags
|= TERMP_NOSPACE
;
551 * Note that we don't process the pipe: the parser sees it as
552 * punctuation, but we don't in terms of typography.
554 if (sv
[0] && '\0' == sv
[1])
559 p
->flags
|= TERMP_NOSPACE
;
568 adjbuf(struct termp
*p
, size_t sz
)
573 while (sz
>= p
->maxcols
)
576 p
->buf
= realloc(p
->buf
, p
->maxcols
);
577 if (NULL
== p
->buf
) {
585 buffera(struct termp
*p
, const char *word
, size_t sz
)
588 if (p
->col
+ sz
>= p
->maxcols
)
589 adjbuf(p
, p
->col
+ sz
);
591 memcpy(&p
->buf
[(int)p
->col
], word
, sz
);
597 bufferc(struct termp
*p
, char c
)
600 if (p
->col
+ 1 >= p
->maxcols
)
601 adjbuf(p
, p
->col
+ 1);
603 p
->buf
[(int)p
->col
++] = c
;
608 encode(struct termp
*p
, const char *word
, size_t sz
)
614 * Encode and buffer a string of characters. If the current
615 * font mode is unset, buffer directly, else encode then buffer
616 * character by character.
619 if (TERMFONT_NONE
== (f
= term_fonttop(p
))) {
620 buffera(p
, word
, sz
);
624 for (i
= 0; i
< (int)sz
; i
++) {
625 if ( ! isgraph((u_char
)word
[i
])) {
630 if (TERMFONT_UNDER
== f
)
642 term_len(const struct termp
*p
, size_t sz
)
645 return((*p
->width
)(p
, ' ') * sz
);
650 term_strlen(const struct termp
*p
, const char *cp
)
654 for (sz
= 0; *cp
; cp
++)
655 sz
+= (*p
->width
)(p
, *cp
);
663 term_vspan(const struct termp
*p
, const struct roffsu
*su
)
681 r
= su
->scale
/ 1000;
693 return(/* LINTED */(size_t)
699 term_hspan(const struct termp
*p
, const struct roffsu
*su
)
703 v
= ((*p
->hspan
)(p
, su
));
706 return((size_t) /* LINTED */