version 1.248, 2015/04/29 18:35:00 |
version 1.285, 2022/01/10 18:01:35 |
|
|
/* $Id$ */ |
/* $Id$ */ |
/* |
/* |
|
* Copyright (c) 2010-2022 Ingo Schwarze <schwarze@openbsd.org> |
* Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv> |
* Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv> |
* Copyright (c) 2010-2015 Ingo Schwarze <schwarze@openbsd.org> |
|
* |
* |
* Permission to use, copy, modify, and distribute this software for any |
* Permission to use, copy, modify, and distribute this software for any |
* purpose with or without fee is hereby granted, provided that the above |
* purpose with or without fee is hereby granted, provided that the above |
|
|
|
|
#include <assert.h> |
#include <assert.h> |
#include <ctype.h> |
#include <ctype.h> |
|
#include <stdint.h> |
#include <stdio.h> |
#include <stdio.h> |
#include <stdlib.h> |
#include <stdlib.h> |
#include <string.h> |
#include <string.h> |
|
|
#include "main.h" |
#include "main.h" |
|
|
static size_t cond_width(const struct termp *, int, int *); |
static size_t cond_width(const struct termp *, int, int *); |
static void adjbuf(struct termp *p, size_t); |
static void adjbuf(struct termp_col *, size_t); |
static void bufferc(struct termp *, char); |
static void bufferc(struct termp *, char); |
static void encode(struct termp *, const char *, size_t); |
static void encode(struct termp *, const char *, size_t); |
static void encode1(struct termp *, int); |
static void encode1(struct termp *, int); |
|
static void endline(struct termp *); |
|
static void term_field(struct termp *, size_t, size_t); |
|
static void term_fill(struct termp *, size_t *, size_t *, |
|
size_t); |
|
|
|
|
void |
void |
term_free(struct termp *p) |
term_setcol(struct termp *p, size_t maxtcol) |
{ |
{ |
|
if (maxtcol > p->maxtcol) { |
|
p->tcols = mandoc_recallocarray(p->tcols, |
|
p->maxtcol, maxtcol, sizeof(*p->tcols)); |
|
p->maxtcol = maxtcol; |
|
} |
|
p->lasttcol = maxtcol - 1; |
|
p->tcol = p->tcols; |
|
} |
|
|
free(p->buf); |
void |
|
term_free(struct termp *p) |
|
{ |
|
term_tab_free(); |
|
for (p->tcol = p->tcols; p->tcol < p->tcols + p->maxtcol; p->tcol++) |
|
free(p->tcol->buf); |
|
free(p->tcols); |
free(p->fontq); |
free(p->fontq); |
free(p); |
free(p); |
} |
} |
Line 69 term_end(struct termp *p) |
|
Line 88 term_end(struct termp *p) |
|
* Flush a chunk of text. By default, break the output line each time |
* Flush a chunk of text. By default, break the output line each time |
* the right margin is reached, and continue output on the next line |
* the right margin is reached, and continue output on the next line |
* at the same offset as the chunk itself. By default, also break the |
* at the same offset as the chunk itself. By default, also break the |
* output line at the end of the chunk. |
* output line at the end of the chunk. There are many flags modifying |
* The following flags may be specified: |
* this behaviour, see the comments in the body of the function. |
* |
|
* - TERMP_NOBREAK: Do not break the output line at the right margin, |
|
* but only at the max right margin. Also, do not break the output |
|
* line at the end of the chunk, such that the next call can pad to |
|
* the next column. However, if less than p->trailspace blanks, |
|
* which can be 0, 1, or 2, remain to the right margin, the line |
|
* will be broken. |
|
* - TERMP_BRIND: If the chunk does not fit and the output line has |
|
* to be broken, start the next line at the right margin instead |
|
* of at the offset. Used together with TERMP_NOBREAK for the tags |
|
* in various kinds of tagged lists. |
|
* - TERMP_DANGLE: Do not break the output line at the right margin, |
|
* append the next chunk after it even if this one is too long. |
|
* To be used together with TERMP_NOBREAK. |
|
* - TERMP_HANG: Like TERMP_DANGLE, and also suppress padding before |
|
* the next chunk if this column is not full. |
|
*/ |
*/ |
void |
void |
term_flushln(struct termp *p) |
term_flushln(struct termp *p) |
{ |
{ |
size_t i; /* current input position in p->buf */ |
size_t vbl; /* Number of blanks to prepend to the output. */ |
int ntab; /* number of tabs to prepend */ |
size_t vbr; /* Actual visual position of the end of field. */ |
size_t vis; /* current visual position on output */ |
size_t vfield; /* Desired visual field width. */ |
size_t vbl; /* number of blanks to prepend to output */ |
size_t vtarget; /* Desired visual position of the right margin. */ |
size_t vend; /* end of word visual position on output */ |
size_t ic; /* Character position in the input buffer. */ |
size_t bp; /* visual right border position */ |
size_t nbr; /* Number of characters to print in this field. */ |
size_t dv; /* temporary for visual pos calculations */ |
|
size_t j; /* temporary loop index for p->buf */ |
|
size_t jhy; /* last hyph before overflow w/r/t j */ |
|
size_t maxvis; /* output position of visible boundary */ |
|
|
|
/* |
/* |
* First, establish the maximum columns of "visible" content. |
* Normally, start writing at the left margin, but with the |
* This is usually the difference between the right-margin and |
* NOPAD flag, start writing at the current position instead. |
* an indentation, but can be, for tagged lists or columns, a |
|
* small set of values. |
|
* |
|
* The following unsigned-signed subtractions look strange, |
|
* but they are actually correct. If the int p->overstep |
|
* is negative, it gets sign extended. Subtracting that |
|
* very large size_t effectively adds a small number to dv. |
|
*/ |
*/ |
dv = p->rmargin > p->offset ? p->rmargin - p->offset : 0; |
|
maxvis = (int)dv > p->overstep ? dv - (size_t)p->overstep : 0; |
|
|
|
if (p->flags & TERMP_NOBREAK) { |
vbl = (p->flags & TERMP_NOPAD) || p->tcol->offset < p->viscol ? |
dv = p->maxrmargin > p->offset ? |
0 : p->tcol->offset - p->viscol; |
p->maxrmargin - p->offset : 0; |
if (p->minbl && vbl < p->minbl) |
bp = (int)dv > p->overstep ? |
vbl = p->minbl; |
dv - (size_t)p->overstep : 0; |
|
} else |
|
bp = maxvis; |
|
|
|
/* |
if ((p->flags & TERMP_MULTICOL) == 0) |
* Calculate the required amount of padding. |
p->tcol->col = 0; |
*/ |
|
vbl = p->offset + p->overstep > p->viscol ? |
|
p->offset + p->overstep - p->viscol : 0; |
|
|
|
vis = vend = 0; |
/* Loop over output lines. */ |
i = 0; |
|
|
|
while (i < p->col) { |
for (;;) { |
|
vfield = p->tcol->rmargin > p->viscol + vbl ? |
|
p->tcol->rmargin - p->viscol - vbl : 0; |
|
|
/* |
/* |
* Handle literal tab characters: collapse all |
* Normally, break the line at the the right margin |
* subsequent tabs into a single huge set of spaces. |
* of the field, but with the NOBREAK flag, only |
|
* break it at the max right margin of the screen, |
|
* and with the BRNEVER flag, never break it at all. |
*/ |
*/ |
ntab = 0; |
|
while (i < p->col && '\t' == p->buf[i]) { |
|
vend = (vis / p->tabwidth + 1) * p->tabwidth; |
|
vbl += vend - vis; |
|
vis = vend; |
|
ntab++; |
|
i++; |
|
} |
|
|
|
|
vtarget = (p->flags & TERMP_NOBREAK) == 0 ? vfield : |
|
p->maxrmargin > p->viscol + vbl ? |
|
p->maxrmargin - p->viscol - vbl : 0; |
|
|
/* |
/* |
* Count up visible word characters. Control sequences |
* Figure out how much text will fit in the field. |
* (starting with the CSI) aren't counted. A space |
* If there is whitespace only, print nothing. |
* generates a non-printing word, which is valid (the |
|
* space is printed according to regular spacing rules). |
|
*/ |
*/ |
|
|
for (j = i, jhy = 0; j < p->col; j++) { |
term_fill(p, &nbr, &vbr, |
if (' ' == p->buf[j] || '\t' == p->buf[j]) |
p->flags & TERMP_BRNEVER ? SIZE_MAX : vtarget); |
break; |
if (nbr == 0) |
|
break; |
|
|
/* Back over the the last printed character. */ |
/* |
if (8 == p->buf[j]) { |
* With the CENTER or RIGHT flag, increase the indentation |
assert(j); |
* to center the text between the left and right margins |
vend -= (*p->width)(p, p->buf[j - 1]); |
* or to adjust it to the right margin, respectively. |
|
*/ |
|
|
|
if (vbr < vtarget) { |
|
if (p->flags & TERMP_CENTER) |
|
vbl += (vtarget - vbr) / 2; |
|
else if (p->flags & TERMP_RIGHT) |
|
vbl += vtarget - vbr; |
|
} |
|
|
|
/* Finally, print the field content. */ |
|
|
|
term_field(p, vbl, nbr); |
|
|
|
/* |
|
* If there is no text left in the field, exit the loop. |
|
* If the BRTRSP flag is set, consider trailing |
|
* whitespace significant when deciding whether |
|
* the field fits or not. |
|
*/ |
|
|
|
for (ic = p->tcol->col; ic < p->tcol->lastcol; ic++) { |
|
switch (p->tcol->buf[ic]) { |
|
case '\t': |
|
if (p->flags & TERMP_BRTRSP) |
|
vbr = term_tab_next(vbr); |
continue; |
continue; |
|
case ' ': |
|
if (p->flags & TERMP_BRTRSP) |
|
vbr += (*p->width)(p, ' '); |
|
continue; |
|
case '\n': |
|
case ASCII_BREAK: |
|
continue; |
|
default: |
|
break; |
} |
} |
|
break; |
|
} |
|
if (ic == p->tcol->lastcol) |
|
break; |
|
|
/* Regular word. */ |
/* |
/* Break at the hyphen point if we overrun. */ |
* At the location of an automtic line break, input |
if (vend > vis && vend < bp && |
* space characters are consumed by the line break. |
(ASCII_HYPH == p->buf[j] || |
*/ |
ASCII_BREAK == p->buf[j])) |
|
jhy = j; |
|
|
|
/* |
while (p->tcol->col < p->tcol->lastcol && |
* Hyphenation now decided, put back a real |
p->tcol->buf[p->tcol->col] == ' ') |
* hyphen such that we get the correct width. |
p->tcol->col++; |
*/ |
|
if (ASCII_HYPH == p->buf[j]) |
|
p->buf[j] = '-'; |
|
|
|
vend += (*p->width)(p, p->buf[j]); |
/* |
} |
* In multi-column mode, leave the rest of the text |
|
* in the buffer to be handled by a subsequent |
|
* invocation, such that the other columns of the |
|
* table can be handled first. |
|
* In single-column mode, simply break the line. |
|
*/ |
|
|
|
if (p->flags & TERMP_MULTICOL) |
|
return; |
|
|
|
endline(p); |
|
p->viscol = 0; |
|
|
/* |
/* |
* Find out whether we would exceed the right margin. |
* Normally, start the next line at the same indentation |
* If so, break to the next line. |
* as this one, but with the BRIND flag, start it at the |
|
* right margin instead. This is used together with |
|
* NOBREAK for the tags in various kinds of tagged lists. |
*/ |
*/ |
if (vend > bp && 0 == jhy && vis > 0) { |
|
vend -= vis; |
|
(*p->endline)(p); |
|
p->viscol = 0; |
|
if (TERMP_BRIND & p->flags) { |
|
vbl = p->rmargin; |
|
vend += p->rmargin; |
|
vend -= p->offset; |
|
} else |
|
vbl = p->offset; |
|
|
|
/* use pending tabs on the new line */ |
vbl = p->flags & TERMP_BRIND ? |
|
p->tcol->rmargin : p->tcol->offset; |
|
} |
|
|
if (0 < ntab) |
/* Reset output state in preparation for the next field. */ |
vbl += ntab * p->tabwidth; |
|
|
|
/* |
p->col = p->tcol->col = p->tcol->lastcol = 0; |
* Remove the p->overstep width. |
p->minbl = p->trailspace; |
* Again, if p->overstep is negative, |
p->flags &= ~(TERMP_BACKAFTER | TERMP_BACKBEFORE | TERMP_NOPAD); |
* sign extension does the right thing. |
|
*/ |
|
|
|
bp += (size_t)p->overstep; |
if (p->flags & TERMP_MULTICOL) |
p->overstep = 0; |
return; |
} |
|
|
|
/* Write out the [remaining] word. */ |
/* |
for ( ; i < p->col; i++) { |
* The HANG flag means that the next field |
if (vend > bp && jhy > 0 && i > jhy) |
* always follows on the same line. |
|
* The NOBREAK flag means that the next field |
|
* follows on the same line unless the field was overrun. |
|
* Normally, break the line at the end of each field. |
|
*/ |
|
|
|
if ((p->flags & TERMP_HANG) == 0 && |
|
((p->flags & TERMP_NOBREAK) == 0 || |
|
vbr + term_len(p, p->trailspace) > vfield)) |
|
endline(p); |
|
} |
|
|
|
/* |
|
* Store the number of input characters to print in this field in *nbr |
|
* and their total visual width to print in *vbr. |
|
* If there is only whitespace in the field, both remain zero. |
|
* The desired visual width of the field is provided by vtarget. |
|
* If the first word is longer, the field will be overrun. |
|
*/ |
|
static void |
|
term_fill(struct termp *p, size_t *nbr, size_t *vbr, size_t vtarget) |
|
{ |
|
size_t ic; /* Character position in the input buffer. */ |
|
size_t vis; /* Visual position of the current character. */ |
|
size_t vn; /* Visual position of the next character. */ |
|
int breakline; /* Break at the end of this word. */ |
|
int graph; /* Last character was non-blank. */ |
|
|
|
*nbr = *vbr = vis = 0; |
|
breakline = graph = 0; |
|
for (ic = p->tcol->col; ic < p->tcol->lastcol; ic++) { |
|
switch (p->tcol->buf[ic]) { |
|
case '\b': /* Escape \o (overstrike) or backspace markup. */ |
|
assert(ic > 0); |
|
vis -= (*p->width)(p, p->tcol->buf[ic - 1]); |
|
continue; |
|
|
|
case '\t': /* Normal ASCII whitespace. */ |
|
case ' ': |
|
case ASCII_BREAK: /* Escape \: (breakpoint). */ |
|
switch (p->tcol->buf[ic]) { |
|
case '\t': |
|
vn = term_tab_next(vis); |
break; |
break; |
if ('\t' == p->buf[i]) |
case ' ': |
|
vn = vis + (*p->width)(p, ' '); |
break; |
break; |
if (' ' == p->buf[i]) { |
case ASCII_BREAK: |
j = i; |
vn = vis; |
while (i < p->col && ' ' == p->buf[i]) |
|
i++; |
|
dv = (i - j) * (*p->width)(p, ' '); |
|
vbl += dv; |
|
vend += dv; |
|
break; |
break; |
|
default: |
|
abort(); |
} |
} |
if (ASCII_NBRSP == p->buf[i]) { |
/* Can break at the end of a word. */ |
vbl += (*p->width)(p, ' '); |
if (breakline || vn > vtarget) |
continue; |
break; |
|
if (graph) { |
|
*nbr = ic; |
|
*vbr = vis; |
|
graph = 0; |
} |
} |
if (ASCII_BREAK == p->buf[i]) |
vis = vn; |
continue; |
continue; |
|
|
|
case '\n': /* Escape \p (break at the end of the word). */ |
|
breakline = 1; |
|
continue; |
|
|
|
case ASCII_HYPH: /* Breakable hyphen. */ |
|
graph = 1; |
/* |
/* |
* Now we definitely know there will be |
* We are about to decide whether to break the |
* printable characters to output, |
* line or not, so we no longer need this hyphen |
* so write preceding white space now. |
* to be marked as breakable. Put back a real |
|
* hyphen such that we get the correct width. |
*/ |
*/ |
if (vbl) { |
p->tcol->buf[ic] = '-'; |
(*p->advance)(p, vbl); |
vis += (*p->width)(p, '-'); |
p->viscol += vbl; |
if (vis > vtarget) { |
vbl = 0; |
ic++; |
|
break; |
} |
} |
|
*nbr = ic + 1; |
|
*vbr = vis; |
|
continue; |
|
|
(*p->letter)(p, p->buf[i]); |
case ASCII_NBRSP: /* Non-breakable space. */ |
if (8 == p->buf[i]) |
p->tcol->buf[ic] = ' '; |
p->viscol -= (*p->width)(p, p->buf[i-1]); |
/* FALLTHROUGH */ |
else |
default: /* Printable character. */ |
p->viscol += (*p->width)(p, p->buf[i]); |
graph = 1; |
|
vis += (*p->width)(p, p->tcol->buf[ic]); |
|
if (vis > vtarget && *nbr > 0) |
|
return; |
|
continue; |
} |
} |
vis = vend; |
break; |
} |
} |
|
|
/* |
/* |
* If there was trailing white space, it was not printed; |
* If the last word extends to the end of the field without any |
* so reset the cursor position accordingly. |
* trailing whitespace, the loop could not check yet whether it |
|
* can remain on this line. So do the check now. |
*/ |
*/ |
if (vis > vbl) |
|
vis -= vbl; |
|
else |
|
vis = 0; |
|
|
|
p->col = 0; |
if (graph && (vis <= vtarget || *nbr == 0)) { |
p->overstep = 0; |
*nbr = ic; |
p->flags &= ~(TERMP_BACKAFTER | TERMP_BACKBEFORE); |
*vbr = vis; |
|
|
if ( ! (TERMP_NOBREAK & p->flags)) { |
|
p->viscol = 0; |
|
(*p->endline)(p); |
|
return; |
|
} |
} |
|
} |
|
|
if (TERMP_HANG & p->flags) { |
/* |
p->overstep += (int)(p->offset + vis - p->rmargin + |
* Print the contents of one field |
p->trailspace * (*p->width)(p, ' ')); |
* with an indentation of vbl visual columns, |
|
* and an input string length of nbr characters. |
|
*/ |
|
static void |
|
term_field(struct termp *p, size_t vbl, size_t nbr) |
|
{ |
|
size_t ic; /* Character position in the input buffer. */ |
|
size_t vis; /* Visual position of the current character. */ |
|
size_t dv; /* Visual width of the current character. */ |
|
size_t vn; /* Visual position of the next character. */ |
|
|
|
vis = 0; |
|
for (ic = p->tcol->col; ic < nbr; ic++) { |
|
|
/* |
/* |
* If we have overstepped the margin, temporarily move |
* To avoid the printing of trailing whitespace, |
* it to the right and flag the rest of the line to be |
* do not print whitespace right away, only count it. |
* shorter. |
|
* If there is a request to keep the columns together, |
|
* allow negative overstep when the column is not full. |
|
*/ |
*/ |
if (p->trailspace && p->overstep < 0) |
|
p->overstep = 0; |
|
return; |
|
|
|
} else if (TERMP_DANGLE & p->flags) |
switch (p->tcol->buf[ic]) { |
return; |
case '\n': |
|
case ASCII_BREAK: |
|
continue; |
|
case '\t': |
|
vn = term_tab_next(vis); |
|
vbl += vn - vis; |
|
vis = vn; |
|
continue; |
|
case ' ': |
|
case ASCII_NBRSP: |
|
dv = (*p->width)(p, ' '); |
|
vbl += dv; |
|
vis += dv; |
|
continue; |
|
default: |
|
break; |
|
} |
|
|
/* If the column was overrun, break the line. */ |
/* |
if (maxvis < vis + p->trailspace * (*p->width)(p, ' ')) { |
* We found a non-blank character to print, |
(*p->endline)(p); |
* so write preceding white space now. |
p->viscol = 0; |
*/ |
|
|
|
if (vbl > 0) { |
|
(*p->advance)(p, vbl); |
|
p->viscol += vbl; |
|
vbl = 0; |
|
} |
|
|
|
/* Print the character and adjust the visual position. */ |
|
|
|
(*p->letter)(p, p->tcol->buf[ic]); |
|
if (p->tcol->buf[ic] == '\b') { |
|
dv = (*p->width)(p, p->tcol->buf[ic - 1]); |
|
p->viscol -= dv; |
|
vis -= dv; |
|
} else { |
|
dv = (*p->width)(p, p->tcol->buf[ic]); |
|
p->viscol += dv; |
|
vis += dv; |
|
} |
} |
} |
|
p->tcol->col = nbr; |
} |
} |
|
|
|
static void |
|
endline(struct termp *p) |
|
{ |
|
if ((p->flags & (TERMP_NEWMC | TERMP_ENDMC)) == TERMP_ENDMC) { |
|
p->mc = NULL; |
|
p->flags &= ~TERMP_ENDMC; |
|
} |
|
if (p->mc != NULL) { |
|
if (p->viscol && p->maxrmargin >= p->viscol) |
|
(*p->advance)(p, p->maxrmargin - p->viscol + 1); |
|
p->flags |= TERMP_NOBUF | TERMP_NOSPACE; |
|
term_word(p, p->mc); |
|
p->flags &= ~(TERMP_NOBUF | TERMP_NEWMC); |
|
} |
|
p->viscol = 0; |
|
p->minbl = 0; |
|
(*p->endline)(p); |
|
} |
|
|
/* |
/* |
* A newline only breaks an existing line; it won't assert vertical |
* A newline only breaks an existing line; it won't assert vertical |
* space. All data in the output buffer is flushed prior to the newline |
* space. All data in the output buffer is flushed prior to the newline |
Line 308 term_newln(struct termp *p) |
|
Line 438 term_newln(struct termp *p) |
|
{ |
{ |
|
|
p->flags |= TERMP_NOSPACE; |
p->flags |= TERMP_NOSPACE; |
if (p->col || p->viscol) |
if (p->tcol->lastcol || p->viscol) |
term_flushln(p); |
term_flushln(p); |
} |
} |
|
|
Line 324 term_vspace(struct termp *p) |
|
Line 454 term_vspace(struct termp *p) |
|
|
|
term_newln(p); |
term_newln(p); |
p->viscol = 0; |
p->viscol = 0; |
|
p->minbl = 0; |
if (0 < p->skipvsp) |
if (0 < p->skipvsp) |
p->skipvsp--; |
p->skipvsp--; |
else |
else |
Line 359 term_fontpush(struct termp *p, enum termfont f) |
|
Line 490 term_fontpush(struct termp *p, enum termfont f) |
|
if (++p->fonti == p->fontsz) { |
if (++p->fonti == p->fontsz) { |
p->fontsz += 8; |
p->fontsz += 8; |
p->fontq = mandoc_reallocarray(p->fontq, |
p->fontq = mandoc_reallocarray(p->fontq, |
p->fontsz, sizeof(enum termfont *)); |
p->fontsz, sizeof(*p->fontq)); |
} |
} |
p->fontq[p->fonti] = f; |
p->fontq[p->fonti] = f; |
} |
} |
Line 391 term_fontpop(struct termp *p) |
|
Line 522 term_fontpop(struct termp *p) |
|
void |
void |
term_word(struct termp *p, const char *word) |
term_word(struct termp *p, const char *word) |
{ |
{ |
|
struct roffsu su; |
const char nbrsp[2] = { ASCII_NBRSP, 0 }; |
const char nbrsp[2] = { ASCII_NBRSP, 0 }; |
const char *seq, *cp; |
const char *seq, *cp; |
int sz, uc; |
int sz, uc; |
size_t ssz; |
size_t csz, lsz, ssz; |
enum mandoc_esc esc; |
enum mandoc_esc esc; |
|
|
if ( ! (TERMP_NOSPACE & p->flags)) { |
if ((p->flags & TERMP_NOBUF) == 0) { |
if ( ! (TERMP_KEEP & p->flags)) { |
if ((p->flags & TERMP_NOSPACE) == 0) { |
bufferc(p, ' '); |
if ((p->flags & TERMP_KEEP) == 0) { |
if (TERMP_SENTENCE & p->flags) |
|
bufferc(p, ' '); |
bufferc(p, ' '); |
} else |
if (p->flags & TERMP_SENTENCE) |
bufferc(p, ASCII_NBRSP); |
bufferc(p, ' '); |
|
} else |
|
bufferc(p, ASCII_NBRSP); |
|
} |
|
if (p->flags & TERMP_PREKEEP) |
|
p->flags |= TERMP_KEEP; |
|
if (p->flags & TERMP_NONOSPACE) |
|
p->flags |= TERMP_NOSPACE; |
|
else |
|
p->flags &= ~TERMP_NOSPACE; |
|
p->flags &= ~(TERMP_SENTENCE | TERMP_NONEWLINE); |
|
p->skipvsp = 0; |
} |
} |
if (TERMP_PREKEEP & p->flags) |
|
p->flags |= TERMP_KEEP; |
|
|
|
if ( ! (p->flags & TERMP_NONOSPACE)) |
|
p->flags &= ~TERMP_NOSPACE; |
|
else |
|
p->flags |= TERMP_NOSPACE; |
|
|
|
p->flags &= ~(TERMP_SENTENCE | TERMP_NONEWLINE); |
|
p->skipvsp = 0; |
|
|
|
while ('\0' != *word) { |
while ('\0' != *word) { |
if ('\\' != *word) { |
if ('\\' != *word) { |
if (TERMP_NBRWORD & p->flags) { |
if (TERMP_NBRWORD & p->flags) { |
Line 434 term_word(struct termp *p, const char *word) |
|
Line 566 term_word(struct termp *p, const char *word) |
|
|
|
word++; |
word++; |
esc = mandoc_escape(&word, &seq, &sz); |
esc = mandoc_escape(&word, &seq, &sz); |
if (ESCAPE_ERROR == esc) |
|
continue; |
|
|
|
switch (esc) { |
switch (esc) { |
case ESCAPE_UNICODE: |
case ESCAPE_UNICODE: |
uc = mchars_num2uc(seq + 1, sz - 1); |
uc = mchars_num2uc(seq + 1, sz - 1); |
Line 448 term_word(struct termp *p, const char *word) |
|
Line 577 term_word(struct termp *p, const char *word) |
|
break; |
break; |
case ESCAPE_SPECIAL: |
case ESCAPE_SPECIAL: |
if (p->enc == TERMENC_ASCII) { |
if (p->enc == TERMENC_ASCII) { |
cp = mchars_spec2str(p->symtab, |
cp = mchars_spec2str(seq, sz, &ssz); |
seq, sz, &ssz); |
|
if (cp != NULL) |
if (cp != NULL) |
encode(p, cp, ssz); |
encode(p, cp, ssz); |
} else { |
} else { |
uc = mchars_spec2cp(p->symtab, seq, sz); |
uc = mchars_spec2cp(seq, sz); |
if (uc > 0) |
if (uc > 0) |
encode1(p, uc); |
encode1(p, uc); |
} |
} |
continue; |
continue; |
|
case ESCAPE_UNDEF: |
|
uc = *seq; |
|
break; |
case ESCAPE_FONTBOLD: |
case ESCAPE_FONTBOLD: |
|
case ESCAPE_FONTCB: |
term_fontrepl(p, TERMFONT_BOLD); |
term_fontrepl(p, TERMFONT_BOLD); |
continue; |
continue; |
case ESCAPE_FONTITALIC: |
case ESCAPE_FONTITALIC: |
|
case ESCAPE_FONTCI: |
term_fontrepl(p, TERMFONT_UNDER); |
term_fontrepl(p, TERMFONT_UNDER); |
continue; |
continue; |
case ESCAPE_FONTBI: |
case ESCAPE_FONTBI: |
term_fontrepl(p, TERMFONT_BI); |
term_fontrepl(p, TERMFONT_BI); |
continue; |
continue; |
case ESCAPE_FONT: |
case ESCAPE_FONT: |
/* FALLTHROUGH */ |
case ESCAPE_FONTCR: |
case ESCAPE_FONTROMAN: |
case ESCAPE_FONTROMAN: |
term_fontrepl(p, TERMFONT_NONE); |
term_fontrepl(p, TERMFONT_NONE); |
continue; |
continue; |
case ESCAPE_FONTPREV: |
case ESCAPE_FONTPREV: |
term_fontlast(p); |
term_fontlast(p); |
continue; |
continue; |
|
case ESCAPE_BREAK: |
|
bufferc(p, '\n'); |
|
continue; |
case ESCAPE_NOSPACE: |
case ESCAPE_NOSPACE: |
if (p->flags & TERMP_BACKAFTER) |
if (p->flags & TERMP_BACKAFTER) |
p->flags &= ~TERMP_BACKAFTER; |
p->flags &= ~TERMP_BACKAFTER; |
else if (*word == '\0') |
else if (*word == '\0') |
p->flags |= (TERMP_NOSPACE | TERMP_NONEWLINE); |
p->flags |= (TERMP_NOSPACE | TERMP_NONEWLINE); |
continue; |
continue; |
|
case ESCAPE_DEVICE: |
|
if (p->type == TERMTYPE_PDF) |
|
encode(p, "pdf", 3); |
|
else if (p->type == TERMTYPE_PS) |
|
encode(p, "ps", 2); |
|
else if (p->enc == TERMENC_ASCII) |
|
encode(p, "ascii", 5); |
|
else |
|
encode(p, "utf8", 4); |
|
continue; |
|
case ESCAPE_HORIZ: |
|
if (*seq == '|') { |
|
seq++; |
|
uc = -p->col; |
|
} else |
|
uc = 0; |
|
if (a2roffsu(seq, &su, SCALE_EM) == NULL) |
|
continue; |
|
uc += term_hen(p, &su); |
|
if (uc > 0) { |
|
while (uc > 0) { |
|
bufferc(p, ASCII_NBRSP); |
|
uc -= term_len(p, 1); |
|
} |
|
} else if (p->col > (size_t)(-uc)) { |
|
p->col += uc; |
|
} else { |
|
uc += p->col; |
|
p->col = 0; |
|
if (p->tcol->offset > (size_t)(-uc)) { |
|
p->ti += uc; |
|
p->tcol->offset += uc; |
|
} else { |
|
p->ti -= p->tcol->offset; |
|
p->tcol->offset = 0; |
|
} |
|
} |
|
continue; |
|
case ESCAPE_HLINE: |
|
if ((cp = a2roffsu(seq, &su, SCALE_EM)) == NULL) |
|
continue; |
|
uc = term_hen(p, &su); |
|
if (uc <= 0) { |
|
if (p->tcol->rmargin <= p->tcol->offset) |
|
continue; |
|
lsz = p->tcol->rmargin - p->tcol->offset; |
|
} else |
|
lsz = uc; |
|
if (*cp == seq[-1]) |
|
uc = -1; |
|
else if (*cp == '\\') { |
|
seq = cp + 1; |
|
esc = mandoc_escape(&seq, &cp, &sz); |
|
switch (esc) { |
|
case ESCAPE_UNICODE: |
|
uc = mchars_num2uc(cp + 1, sz - 1); |
|
break; |
|
case ESCAPE_NUMBERED: |
|
uc = mchars_num2char(cp, sz); |
|
break; |
|
case ESCAPE_SPECIAL: |
|
uc = mchars_spec2cp(cp, sz); |
|
break; |
|
case ESCAPE_UNDEF: |
|
uc = *seq; |
|
break; |
|
default: |
|
uc = -1; |
|
break; |
|
} |
|
} else |
|
uc = *cp; |
|
if (uc < 0x20 || (uc > 0x7E && uc < 0xA0)) |
|
uc = '_'; |
|
if (p->enc == TERMENC_ASCII) { |
|
cp = ascii_uc2str(uc); |
|
csz = term_strlen(p, cp); |
|
ssz = strlen(cp); |
|
} else |
|
csz = (*p->width)(p, uc); |
|
while (lsz >= csz) { |
|
if (p->enc == TERMENC_ASCII) |
|
encode(p, cp, ssz); |
|
else |
|
encode1(p, uc); |
|
lsz -= csz; |
|
} |
|
continue; |
case ESCAPE_SKIPCHAR: |
case ESCAPE_SKIPCHAR: |
p->flags |= TERMP_BACKAFTER; |
p->flags |= TERMP_BACKAFTER; |
continue; |
continue; |
Line 499 term_word(struct termp *p, const char *word) |
|
Line 723 term_word(struct termp *p, const char *word) |
|
p->flags |= TERMP_BACKBEFORE; |
p->flags |= TERMP_BACKBEFORE; |
} |
} |
} |
} |
|
/* Trim trailing backspace/blank pair. */ |
|
if (p->tcol->lastcol > 2 && |
|
(p->tcol->buf[p->tcol->lastcol - 1] == ' ' || |
|
p->tcol->buf[p->tcol->lastcol - 1] == '\t')) |
|
p->tcol->lastcol -= 2; |
|
if (p->col > p->tcol->lastcol) |
|
p->col = p->tcol->lastcol; |
continue; |
continue; |
default: |
default: |
continue; |
continue; |
Line 523 term_word(struct termp *p, const char *word) |
|
Line 754 term_word(struct termp *p, const char *word) |
|
} |
} |
|
|
static void |
static void |
adjbuf(struct termp *p, size_t sz) |
adjbuf(struct termp_col *c, size_t sz) |
{ |
{ |
|
if (c->maxcols == 0) |
if (0 == p->maxcols) |
c->maxcols = 1024; |
p->maxcols = 1024; |
while (c->maxcols <= sz) |
while (sz >= p->maxcols) |
c->maxcols <<= 2; |
p->maxcols <<= 2; |
c->buf = mandoc_reallocarray(c->buf, c->maxcols, sizeof(*c->buf)); |
|
|
p->buf = mandoc_reallocarray(p->buf, p->maxcols, sizeof(int)); |
|
} |
} |
|
|
static void |
static void |
bufferc(struct termp *p, char c) |
bufferc(struct termp *p, char c) |
{ |
{ |
|
if (p->flags & TERMP_NOBUF) { |
if (p->col + 1 >= p->maxcols) |
(*p->letter)(p, c); |
adjbuf(p, p->col + 1); |
return; |
|
} |
p->buf[p->col++] = c; |
if (p->col + 1 >= p->tcol->maxcols) |
|
adjbuf(p->tcol, p->col + 1); |
|
if (p->tcol->lastcol <= p->col || (c != ' ' && c != ASCII_NBRSP)) |
|
p->tcol->buf[p->col] = c; |
|
if (p->tcol->lastcol < ++p->col) |
|
p->tcol->lastcol = p->col; |
} |
} |
|
|
/* |
/* |
Line 554 encode1(struct termp *p, int c) |
|
Line 788 encode1(struct termp *p, int c) |
|
{ |
{ |
enum termfont f; |
enum termfont f; |
|
|
if (p->col + 7 >= p->maxcols) |
if (p->flags & TERMP_NOBUF) { |
adjbuf(p, p->col + 7); |
(*p->letter)(p, c); |
|
return; |
|
} |
|
|
f = (c == ASCII_HYPH || isgraph(c)) ? |
if (p->col + 7 >= p->tcol->maxcols) |
|
adjbuf(p->tcol, p->col + 7); |
|
|
|
f = (c == ASCII_HYPH || c > 127 || isgraph(c)) ? |
p->fontq[p->fonti] : TERMFONT_NONE; |
p->fontq[p->fonti] : TERMFONT_NONE; |
|
|
if (p->flags & TERMP_BACKBEFORE) { |
if (p->flags & TERMP_BACKBEFORE) { |
p->buf[p->col++] = 8; |
if (p->tcol->buf[p->col - 1] == ' ' || |
|
p->tcol->buf[p->col - 1] == '\t') |
|
p->col--; |
|
else |
|
p->tcol->buf[p->col++] = '\b'; |
p->flags &= ~TERMP_BACKBEFORE; |
p->flags &= ~TERMP_BACKBEFORE; |
} |
} |
if (TERMFONT_UNDER == f || TERMFONT_BI == f) { |
if (f == TERMFONT_UNDER || f == TERMFONT_BI) { |
p->buf[p->col++] = '_'; |
p->tcol->buf[p->col++] = '_'; |
p->buf[p->col++] = 8; |
p->tcol->buf[p->col++] = '\b'; |
} |
} |
if (TERMFONT_BOLD == f || TERMFONT_BI == f) { |
if (f == TERMFONT_BOLD || f == TERMFONT_BI) { |
if (ASCII_HYPH == c) |
if (c == ASCII_HYPH) |
p->buf[p->col++] = '-'; |
p->tcol->buf[p->col++] = '-'; |
else |
else |
p->buf[p->col++] = c; |
p->tcol->buf[p->col++] = c; |
p->buf[p->col++] = 8; |
p->tcol->buf[p->col++] = '\b'; |
} |
} |
p->buf[p->col++] = c; |
if (p->tcol->lastcol <= p->col || (c != ' ' && c != ASCII_NBRSP)) |
|
p->tcol->buf[p->col] = c; |
|
if (p->tcol->lastcol < ++p->col) |
|
p->tcol->lastcol = p->col; |
if (p->flags & TERMP_BACKAFTER) { |
if (p->flags & TERMP_BACKAFTER) { |
p->flags |= TERMP_BACKBEFORE; |
p->flags |= TERMP_BACKBEFORE; |
p->flags &= ~TERMP_BACKAFTER; |
p->flags &= ~TERMP_BACKAFTER; |
Line 587 encode(struct termp *p, const char *word, size_t sz) |
|
Line 833 encode(struct termp *p, const char *word, size_t sz) |
|
{ |
{ |
size_t i; |
size_t i; |
|
|
if (p->col + 2 + (sz * 5) >= p->maxcols) |
if (p->flags & TERMP_NOBUF) { |
adjbuf(p, p->col + 2 + (sz * 5)); |
for (i = 0; i < sz; i++) |
|
(*p->letter)(p, word[i]); |
|
return; |
|
} |
|
|
|
if (p->col + 2 + (sz * 5) >= p->tcol->maxcols) |
|
adjbuf(p->tcol, p->col + 2 + (sz * 5)); |
|
|
for (i = 0; i < sz; i++) { |
for (i = 0; i < sz; i++) { |
if (ASCII_HYPH == word[i] || |
if (ASCII_HYPH == word[i] || |
isgraph((unsigned char)word[i])) |
isgraph((unsigned char)word[i])) |
encode1(p, word[i]); |
encode1(p, word[i]); |
else |
else { |
p->buf[p->col++] = word[i]; |
if (p->tcol->lastcol <= p->col || |
|
(word[i] != ' ' && word[i] != ASCII_NBRSP)) |
|
p->tcol->buf[p->col] = word[i]; |
|
p->col++; |
|
|
|
/* |
|
* Postpone the effect of \z while handling |
|
* an overstrike sequence from ascii_uc2str(). |
|
*/ |
|
|
|
if (word[i] == '\b' && |
|
(p->flags & TERMP_BACKBEFORE)) { |
|
p->flags &= ~TERMP_BACKBEFORE; |
|
p->flags |= TERMP_BACKAFTER; |
|
} |
|
} |
} |
} |
|
if (p->tcol->lastcol < p->col) |
|
p->tcol->lastcol = p->col; |
} |
} |
|
|
void |
void |
Line 620 term_setwidth(struct termp *p, const char *wstr) |
|
Line 889 term_setwidth(struct termp *p, const char *wstr) |
|
default: |
default: |
break; |
break; |
} |
} |
if (a2roffsu(wstr, &su, SCALE_MAX)) |
if (a2roffsu(wstr, &su, SCALE_MAX) != NULL) |
width = term_hspan(p, &su); |
width = term_hspan(p, &su); |
else |
else |
iop = 0; |
iop = 0; |
|
|
term_len(const struct termp *p, size_t sz) |
term_len(const struct termp *p, size_t sz) |
{ |
{ |
|
|
return((*p->width)(p, ' ') * sz); |
return (*p->width)(p, ' ') * sz; |
} |
} |
|
|
static size_t |
static size_t |
Line 641 cond_width(const struct termp *p, int c, int *skip) |
|
Line 910 cond_width(const struct termp *p, int c, int *skip) |
|
|
|
if (*skip) { |
if (*skip) { |
(*skip) = 0; |
(*skip) = 0; |
return(0); |
return 0; |
} else |
} else |
return((*p->width)(p, c)); |
return (*p->width)(p, c); |
} |
} |
|
|
size_t |
size_t |
Line 672 term_strlen(const struct termp *p, const char *cp) |
|
Line 941 term_strlen(const struct termp *p, const char *cp) |
|
switch (*cp) { |
switch (*cp) { |
case '\\': |
case '\\': |
cp++; |
cp++; |
esc = mandoc_escape(&cp, &seq, &ssz); |
|
if (ESCAPE_ERROR == esc) |
|
continue; |
|
|
|
rhs = NULL; |
rhs = NULL; |
|
esc = mandoc_escape(&cp, &seq, &ssz); |
switch (esc) { |
switch (esc) { |
case ESCAPE_UNICODE: |
case ESCAPE_UNICODE: |
uc = mchars_num2uc(seq + 1, ssz - 1); |
uc = mchars_num2uc(seq + 1, ssz - 1); |
Line 689 term_strlen(const struct termp *p, const char *cp) |
|
Line 954 term_strlen(const struct termp *p, const char *cp) |
|
break; |
break; |
case ESCAPE_SPECIAL: |
case ESCAPE_SPECIAL: |
if (p->enc == TERMENC_ASCII) { |
if (p->enc == TERMENC_ASCII) { |
rhs = mchars_spec2str(p->symtab, |
rhs = mchars_spec2str(seq, ssz, &rsz); |
seq, ssz, &rsz); |
|
if (rhs != NULL) |
if (rhs != NULL) |
break; |
break; |
} else { |
} else { |
uc = mchars_spec2cp(p->symtab, |
uc = mchars_spec2cp(seq, ssz); |
seq, ssz); |
|
if (uc > 0) |
if (uc > 0) |
sz += cond_width(p, uc, &skip); |
sz += cond_width(p, uc, &skip); |
} |
} |
continue; |
continue; |
|
case ESCAPE_UNDEF: |
|
uc = *seq; |
|
break; |
|
case ESCAPE_DEVICE: |
|
if (p->type == TERMTYPE_PDF) { |
|
rhs = "pdf"; |
|
rsz = 3; |
|
} else if (p->type == TERMTYPE_PS) { |
|
rhs = "ps"; |
|
rsz = 2; |
|
} else if (p->enc == TERMENC_ASCII) { |
|
rhs = "ascii"; |
|
rsz = 5; |
|
} else { |
|
rhs = "utf8"; |
|
rsz = 4; |
|
} |
|
break; |
case ESCAPE_SKIPCHAR: |
case ESCAPE_SKIPCHAR: |
skip = 1; |
skip = 1; |
continue; |
continue; |
Line 759 term_strlen(const struct termp *p, const char *cp) |
|
Line 1040 term_strlen(const struct termp *p, const char *cp) |
|
case ASCII_HYPH: |
case ASCII_HYPH: |
sz += cond_width(p, '-', &skip); |
sz += cond_width(p, '-', &skip); |
cp++; |
cp++; |
/* FALLTHROUGH */ |
|
case ASCII_BREAK: |
|
break; |
break; |
default: |
default: |
break; |
break; |
} |
} |
} |
} |
|
|
return(sz); |
return sz; |
} |
} |
|
|
int |
int |
Line 799 term_vspan(const struct termp *p, const struct roffsu |
|
Line 1078 term_vspan(const struct termp *p, const struct roffsu |
|
r = su->scale / 12.0; |
r = su->scale / 12.0; |
break; |
break; |
case SCALE_EN: |
case SCALE_EN: |
/* FALLTHROUGH */ |
|
case SCALE_EM: |
case SCALE_EM: |
r = su->scale * 0.6; |
r = su->scale * 0.6; |
break; |
break; |
Line 808 term_vspan(const struct termp *p, const struct roffsu |
|
Line 1086 term_vspan(const struct termp *p, const struct roffsu |
|
break; |
break; |
default: |
default: |
abort(); |
abort(); |
/* NOTREACHED */ |
|
} |
} |
ri = r > 0.0 ? r + 0.4995 : r - 0.4995; |
ri = r > 0.0 ? r + 0.4995 : r - 0.4995; |
return(ri < 66 ? ri : 1); |
return ri < 66 ? ri : 1; |
} |
} |
|
|
/* |
/* |
* Convert a scaling width to basic units, rounding down. |
* Convert a scaling width to basic units, rounding towards 0. |
*/ |
*/ |
int |
int |
term_hspan(const struct termp *p, const struct roffsu *su) |
term_hspan(const struct termp *p, const struct roffsu *su) |
{ |
{ |
|
|
return((*p->hspan)(p, su)); |
return (*p->hspan)(p, su); |
|
} |
|
|
|
/* |
|
* Convert a scaling width to basic units, rounding to closest. |
|
*/ |
|
int |
|
term_hen(const struct termp *p, const struct roffsu *su) |
|
{ |
|
int bu; |
|
|
|
if ((bu = (*p->hspan)(p, su)) >= 0) |
|
return (bu + 11) / 24; |
|
else |
|
return -((-bu + 11) / 24); |
} |
} |