version 1.109, 2009/10/22 18:19:36 |
version 1.270, 2017/06/14 01:31:27 |
|
|
/* $Id$ */ |
/* $Id$ */ |
/* |
/* |
* Copyright (c) 2008, 2009 Kristaps Dzonsons <kristaps@kth.se> |
* Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv> |
|
* Copyright (c) 2010-2017 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 |
* copyright notice and this permission notice appear in all copies. |
* copyright notice and this permission notice appear in all copies. |
* |
* |
* THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES |
* THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHORS DISCLAIM ALL WARRANTIES |
* WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF |
* WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF |
* MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR |
* MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR |
* ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES |
* ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES |
* WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN |
* WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN |
* ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF |
* ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF |
* OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE. |
* OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE. |
*/ |
*/ |
|
#include "config.h" |
|
|
|
#include <sys/types.h> |
|
|
#include <assert.h> |
#include <assert.h> |
#include <err.h> |
#include <ctype.h> |
#include <stdio.h> |
#include <stdio.h> |
#include <stdlib.h> |
#include <stdlib.h> |
#include <string.h> |
#include <string.h> |
|
|
#include "chars.h" |
#include "mandoc.h" |
|
#include "mandoc_aux.h" |
#include "out.h" |
#include "out.h" |
#include "term.h" |
#include "term.h" |
#include "man.h" |
|
#include "mdoc.h" |
|
#include "main.h" |
#include "main.h" |
|
|
/* FIXME: accomodate non-breaking, non-collapsing white-space. */ |
static size_t cond_width(const struct termp *, int, int *); |
/* FIXME: accomodate non-breaking, collapsing white-space. */ |
static void adjbuf(struct termp_col *, size_t); |
|
static void bufferc(struct termp *, char); |
|
static void encode(struct termp *, const char *, size_t); |
|
static void encode1(struct termp *, int); |
|
static void endline(struct termp *); |
|
|
static struct termp *term_alloc(enum termenc); |
|
static void term_free(struct termp *); |
|
|
|
static void do_escaped(struct termp *, const char **); |
void |
static void do_special(struct termp *, |
term_setcol(struct termp *p, size_t maxtcol) |
const char *, size_t); |
|
static void do_reserved(struct termp *, |
|
const char *, size_t); |
|
static void buffer(struct termp *, char); |
|
static void encode(struct termp *, char); |
|
|
|
|
|
void * |
|
ascii_alloc(void) |
|
{ |
{ |
|
if (maxtcol > p->maxtcol) { |
return(term_alloc(TERMENC_ASCII)); |
p->tcols = mandoc_recallocarray(p->tcols, |
|
p->maxtcol, maxtcol, sizeof(*p->tcols)); |
|
p->maxtcol = maxtcol; |
|
} |
|
p->lasttcol = maxtcol - 1; |
|
p->tcol = p->tcols; |
} |
} |
|
|
|
|
void |
void |
terminal_free(void *arg) |
term_free(struct termp *p) |
{ |
{ |
|
for (p->tcol = p->tcols; p->tcol < p->tcols + p->maxtcol; p->tcol++) |
term_free((struct termp *)arg); |
free(p->tcol->buf); |
|
free(p->tcols); |
|
free(p->fontq); |
|
free(p); |
} |
} |
|
|
|
void |
static void |
term_begin(struct termp *p, term_margin head, |
term_free(struct termp *p) |
term_margin foot, const struct roff_meta *arg) |
{ |
{ |
|
|
if (p->buf) |
p->headf = head; |
free(p->buf); |
p->footf = foot; |
if (p->symtab) |
p->argf = arg; |
chars_free(p->symtab); |
(*p->begin)(p); |
|
|
free(p); |
|
} |
} |
|
|
|
void |
static struct termp * |
term_end(struct termp *p) |
term_alloc(enum termenc enc) |
|
{ |
{ |
struct termp *p; |
|
|
|
if (NULL == (p = malloc(sizeof(struct termp)))) |
(*p->end)(p); |
return(NULL); |
|
bzero(p, sizeof(struct termp)); |
|
p->maxrmargin = 78; |
|
p->enc = enc; |
|
return(p); |
|
} |
} |
|
|
|
|
/* |
/* |
* Flush a line of text. A "line" is loosely defined as being something |
* Flush a chunk of text. By default, break the output line each time |
* that should be followed by a newline, regardless of whether it's |
* the right margin is reached, and continue output on the next line |
* broken apart by newlines getting there. A line can also be a |
* at the same offset as the chunk itself. By default, also break the |
* fragment of a columnar list. |
* output line at the end of the chunk. |
|
* The following flags may be specified: |
* |
* |
* Specifically, a line is whatever's in p->buf of length p->col, which |
* - TERMP_NOBREAK: Do not break the output line at the right margin, |
* is zeroed after this function returns. |
* 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 usage of termp:flags is as follows: |
* the next column. However, if less than p->trailspace blanks, |
* |
* which can be 0, 1, or 2, remain to the right margin, the line |
* - TERMP_NOLPAD: when beginning to write the line, don't left-pad the |
* will be broken. |
* offset value. This is useful when doing columnar lists where the |
* - TERMP_BRTRSP: Consider trailing whitespace significant |
* prior column has right-padded. |
* when deciding whether the chunk fits or not. |
* |
* - TERMP_BRIND: If the chunk does not fit and the output line has |
* - TERMP_NOBREAK: this is the most important and is used when making |
* to be broken, start the next line at the right margin instead |
* columns. In short: don't print a newline and instead pad to the |
* of at the offset. Used together with TERMP_NOBREAK for the tags |
* right margin. Used in conjunction with TERMP_NOLPAD. |
* in various kinds of tagged lists. |
* |
* - TERMP_HANG: Do not break the output line at the right margin, |
* - TERMP_TWOSPACE: when padding, make sure there are at least two |
* append the next chunk after it even if this one is too long. |
* space characters of padding. Otherwise, rather break the line. |
* To be used together with TERMP_NOBREAK. |
* |
* - TERMP_NOPAD: Start writing at the current position, |
* - TERMP_DANGLE: don't newline when TERMP_NOBREAK is specified and |
* do not pad with blank characters up to the offset. |
* the line is overrun, and don't pad-right if it's underrun. |
|
* |
|
* - TERMP_HANG: like TERMP_DANGLE, but doesn't newline when |
|
* overruning, instead save the position and continue at that point |
|
* when the next invocation. |
|
* |
|
* In-line line breaking: |
|
* |
|
* If TERMP_NOBREAK is specified and the line overruns the right |
|
* margin, it will break and pad-right to the right margin after |
|
* writing. If maxrmargin is violated, it will break and continue |
|
* writing from the right-margin, which will lead to the above |
|
* scenario upon exit. |
|
* |
|
* Otherwise, the line will break at the right margin. Extremely long |
|
* lines will cause the system to emit a warning (TODO: hyphenate, if |
|
* possible). |
|
*/ |
*/ |
void |
void |
term_flushln(struct termp *p) |
term_flushln(struct termp *p) |
{ |
{ |
int i, j; |
size_t vis; /* current visual position on output */ |
size_t vbl, vsz, vis, maxvis, mmax, bp, os; |
size_t vbl; /* number of blanks to prepend to output */ |
static int overstep = 0; |
size_t vend; /* end of word visual position on output */ |
|
size_t bp; /* visual right border position */ |
|
size_t dv; /* temporary for visual pos calculations */ |
|
size_t j; /* temporary loop index for p->tcol->buf */ |
|
size_t jhy; /* last hyph before overflow w/r/t j */ |
|
size_t maxvis; /* output position of visible boundary */ |
|
int ntab; /* number of tabs to prepend */ |
|
int breakline; /* after this word */ |
|
|
/* |
vbl = (p->flags & TERMP_NOPAD) || p->tcol->offset < p->viscol ? |
* First, establish the maximum columns of "visible" content. |
0 : p->tcol->offset - p->viscol; |
* This is usually the difference between the right-margin and |
if (p->minbl && vbl < p->minbl) |
* an indentation, but can be, for tagged lists or columns, a |
vbl = p->minbl; |
* small set of values. |
maxvis = p->tcol->rmargin > p->viscol + vbl ? |
*/ |
p->tcol->rmargin - p->viscol - vbl : 0; |
|
bp = !(p->flags & TERMP_NOBREAK) ? maxvis : |
|
p->maxrmargin > p->viscol + vbl ? |
|
p->maxrmargin - p->viscol - vbl : 0; |
|
vis = vend = 0; |
|
|
assert(p->offset < p->rmargin); |
if ((p->flags && TERMP_MULTICOL) == 0) |
assert((int)(p->rmargin - p->offset) - overstep > 0); |
p->tcol->col = 0; |
|
while (p->tcol->col < p->tcol->lastcol) { |
|
|
/* Save the overstep. */ |
/* |
os = (size_t)overstep; |
* Handle literal tab characters: collapse all |
|
* subsequent tabs into a single huge set of spaces. |
|
*/ |
|
|
maxvis = /* LINTED */ |
ntab = 0; |
p->rmargin - p->offset - overstep; |
while (p->tcol->col < p->tcol->lastcol && |
mmax = /* LINTED */ |
p->tcol->buf[p->tcol->col] == '\t') { |
p->maxrmargin - p->offset - overstep; |
vend = term_tab_next(vis); |
|
vbl += vend - vis; |
|
vis = vend; |
|
ntab++; |
|
p->tcol->col++; |
|
} |
|
|
bp = TERMP_NOBREAK & p->flags ? mmax : maxvis; |
|
vis = 0; |
|
overstep = 0; |
|
|
|
/* |
|
* If in the standard case (left-justified), then begin with our |
|
* indentation, otherwise (columns, etc.) just start spitting |
|
* out text. |
|
*/ |
|
|
|
if ( ! (p->flags & TERMP_NOLPAD)) |
|
/* LINTED */ |
|
for (j = 0; j < (int)p->offset; j++) |
|
putchar(' '); |
|
|
|
for (i = 0; i < (int)p->col; i++) { |
|
/* |
/* |
* Count up visible word characters. Control sequences |
* Count up visible word characters. Control sequences |
* (starting with the CSI) aren't counted. A space |
* (starting with the CSI) aren't counted. A space |
Line 174 term_flushln(struct termp *p) |
|
Line 155 term_flushln(struct termp *p) |
|
* space is printed according to regular spacing rules). |
* space is printed according to regular spacing rules). |
*/ |
*/ |
|
|
/* LINTED */ |
jhy = 0; |
for (j = i, vsz = 0; j < (int)p->col; j++) { |
breakline = 0; |
if (j && ' ' == p->buf[j]) |
for (j = p->tcol->col; j < p->tcol->lastcol; j++) { |
|
if (p->tcol->buf[j] == '\n') { |
|
if ((p->flags & TERMP_BRIND) == 0) |
|
breakline = 1; |
|
continue; |
|
} |
|
if (p->tcol->buf[j] == ' ' || p->tcol->buf[j] == '\t') |
break; |
break; |
else if (8 == p->buf[j]) |
|
vsz--; |
/* Back over the last printed character. */ |
else |
if (p->tcol->buf[j] == '\b') { |
vsz++; |
assert(j); |
|
vend -= (*p->width)(p, p->tcol->buf[j - 1]); |
|
continue; |
|
} |
|
|
|
/* Regular word. */ |
|
/* Break at the hyphen point if we overrun. */ |
|
if (vend > vis && vend < bp && |
|
(p->tcol->buf[j] == ASCII_HYPH|| |
|
p->tcol->buf[j] == ASCII_BREAK)) |
|
jhy = j; |
|
|
|
/* |
|
* Hyphenation now decided, put back a real |
|
* hyphen such that we get the correct width. |
|
*/ |
|
if (p->tcol->buf[j] == ASCII_HYPH) |
|
p->tcol->buf[j] = '-'; |
|
|
|
vend += (*p->width)(p, p->tcol->buf[j]); |
} |
} |
|
|
/* |
/* |
* Choose the number of blanks to prepend: no blank at the |
|
* beginning of a line, one between words -- but do not |
|
* actually write them yet. |
|
*/ |
|
vbl = (size_t)(0 == vis ? 0 : 1); |
|
|
|
/* |
|
* Find out whether we would exceed the right margin. |
* Find out whether we would exceed the right margin. |
* If so, break to the next line. (TODO: hyphenate) |
* If so, break to the next line. |
* Otherwise, write the chosen number of blanks now. |
|
*/ |
*/ |
if (vis && vis + vbl + vsz > bp) { |
|
putchar('\n'); |
if (vend > bp && jhy == 0 && vis > 0 && |
if (TERMP_NOBREAK & p->flags) { |
(p->flags & TERMP_BRNEVER) == 0) { |
for (j = 0; j < (int)p->rmargin; j++) |
if (p->flags & TERMP_MULTICOL) |
putchar(' '); |
return; |
vis = p->rmargin - p->offset; |
|
} else { |
endline(p); |
for (j = 0; j < (int)p->offset; j++) |
vend -= vis; |
putchar(' '); |
|
vis = 0; |
/* Use pending tabs on the new line. */ |
} |
|
/* Remove the overstep width. */ |
vbl = 0; |
bp += os; |
while (ntab--) |
os = 0; |
vbl = term_tab_next(vbl); |
} else { |
|
for (j = 0; j < (int)vbl; j++) |
/* Re-establish indentation. */ |
putchar(' '); |
|
vis += vbl; |
if (p->flags & TERMP_BRIND) |
|
vbl += p->tcol->rmargin; |
|
else |
|
vbl += p->tcol->offset; |
|
maxvis = p->tcol->rmargin > vbl ? |
|
p->tcol->rmargin - vbl : 0; |
|
bp = !(p->flags & TERMP_NOBREAK) ? maxvis : |
|
p->maxrmargin > vbl ? p->maxrmargin - vbl : 0; |
} |
} |
|
|
/* |
/* |
* Finally, write out the word. |
* Write out the rest of the word. |
*/ |
*/ |
for ( ; i < (int)p->col; i++) { |
|
if (' ' == p->buf[i]) |
for ( ; p->tcol->col < p->tcol->lastcol; p->tcol->col++) { |
|
if (vend > bp && jhy > 0 && p->tcol->col > jhy) |
break; |
break; |
putchar(p->buf[i]); |
if (p->tcol->buf[p->tcol->col] == '\n') |
|
continue; |
|
if (p->tcol->buf[p->tcol->col] == '\t') |
|
break; |
|
if (p->tcol->buf[p->tcol->col] == ' ') { |
|
j = p->tcol->col; |
|
while (p->tcol->col < p->tcol->lastcol && |
|
p->tcol->buf[p->tcol->col] == ' ') |
|
p->tcol->col++; |
|
dv = (p->tcol->col - j) * (*p->width)(p, ' '); |
|
vbl += dv; |
|
vend += dv; |
|
break; |
|
} |
|
if (p->tcol->buf[p->tcol->col] == ASCII_NBRSP) { |
|
vbl += (*p->width)(p, ' '); |
|
continue; |
|
} |
|
if (p->tcol->buf[p->tcol->col] == ASCII_BREAK) |
|
continue; |
|
|
|
/* |
|
* Now we definitely know there will be |
|
* printable characters to output, |
|
* so write preceding white space now. |
|
*/ |
|
if (vbl) { |
|
(*p->advance)(p, vbl); |
|
p->viscol += vbl; |
|
vbl = 0; |
|
} |
|
|
|
(*p->letter)(p, p->tcol->buf[p->tcol->col]); |
|
if (p->tcol->buf[p->tcol->col] == '\b') |
|
p->viscol -= (*p->width)(p, |
|
p->tcol->buf[p->tcol->col - 1]); |
|
else |
|
p->viscol += (*p->width)(p, |
|
p->tcol->buf[p->tcol->col]); |
} |
} |
vis += vsz; |
vis = vend; |
} |
|
p->col = 0; |
|
|
|
if ( ! (TERMP_NOBREAK & p->flags)) { |
if (breakline == 0) |
putchar('\n'); |
continue; |
return; |
|
|
/* Explicitly requested output line break. */ |
|
|
|
if (p->flags & TERMP_MULTICOL) |
|
return; |
|
|
|
endline(p); |
|
breakline = 0; |
|
vis = vend = 0; |
|
|
|
/* Re-establish indentation. */ |
|
|
|
vbl = p->tcol->offset; |
|
maxvis = p->tcol->rmargin > vbl ? |
|
p->tcol->rmargin - vbl : 0; |
|
bp = !(p->flags & TERMP_NOBREAK) ? maxvis : |
|
p->maxrmargin > vbl ? p->maxrmargin - vbl : 0; |
} |
} |
|
|
if (TERMP_HANG & p->flags) { |
/* |
/* We need one blank after the tag. */ |
* If there was trailing white space, it was not printed; |
overstep = /* LINTED */ |
* so reset the cursor position accordingly. |
vis - maxvis + 1; |
*/ |
|
|
/* |
if (vis > vbl) |
* Behave exactly the same way as groff: |
vis -= vbl; |
* If we have overstepped the margin, temporarily move |
else |
* it to the right and flag the rest of the line to be |
vis = 0; |
* shorter. |
|
* If we landed right at the margin, be happy. |
|
* If we are one step before the margin, temporarily |
|
* move it one step LEFT and flag the rest of the line |
|
* to be longer. |
|
*/ |
|
if (overstep >= -1) { |
|
assert((int)maxvis + overstep >= 0); |
|
/* LINTED */ |
|
maxvis += overstep; |
|
} else |
|
overstep = 0; |
|
|
|
} else if (TERMP_DANGLE & p->flags) |
p->col = p->tcol->col = p->tcol->lastcol = 0; |
|
p->minbl = p->trailspace; |
|
p->flags &= ~(TERMP_BACKAFTER | TERMP_BACKBEFORE | TERMP_NOPAD); |
|
|
|
if (p->flags & TERMP_MULTICOL) |
return; |
return; |
|
|
/* Right-pad. */ |
/* Trailing whitespace is significant in some columns. */ |
if (maxvis > vis + /* LINTED */ |
|
((TERMP_TWOSPACE & p->flags) ? 1 : 0)) |
if (vis && vbl && (TERMP_BRTRSP & p->flags)) |
for ( ; vis < maxvis; vis++) |
vis += vbl; |
putchar(' '); |
|
else { /* ...or newline break. */ |
/* If the column was overrun, break the line. */ |
putchar('\n'); |
if ((p->flags & TERMP_NOBREAK) == 0 || |
for (i = 0; i < (int)p->rmargin; i++) |
((p->flags & TERMP_HANG) == 0 && |
putchar(' '); |
vis + p->trailspace * (*p->width)(p, ' ') > maxvis)) |
|
endline(p); |
|
} |
|
|
|
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 |
* assertion. |
* assertion. |
Line 281 term_newln(struct termp *p) |
|
Line 349 term_newln(struct termp *p) |
|
{ |
{ |
|
|
p->flags |= TERMP_NOSPACE; |
p->flags |= TERMP_NOSPACE; |
if (0 == p->col) { |
if (p->tcol->lastcol || p->viscol) |
p->flags &= ~TERMP_NOLPAD; |
term_flushln(p); |
return; |
|
} |
|
term_flushln(p); |
|
p->flags &= ~TERMP_NOLPAD; |
|
} |
} |
|
|
|
|
/* |
/* |
* Asserts a vertical space (a full, empty line-break between lines). |
* Asserts a vertical space (a full, empty line-break between lines). |
* Note that if used twice, this will cause two blank spaces and so on. |
* Note that if used twice, this will cause two blank spaces and so on. |
Line 301 term_vspace(struct termp *p) |
|
Line 364 term_vspace(struct termp *p) |
|
{ |
{ |
|
|
term_newln(p); |
term_newln(p); |
putchar('\n'); |
p->viscol = 0; |
|
p->minbl = 0; |
|
if (0 < p->skipvsp) |
|
p->skipvsp--; |
|
else |
|
(*p->endline)(p); |
} |
} |
|
|
|
/* Swap current and previous font; for \fP and .ft P */ |
|
void |
|
term_fontlast(struct termp *p) |
|
{ |
|
enum termfont f; |
|
|
static void |
f = p->fontl; |
do_special(struct termp *p, const char *word, size_t len) |
p->fontl = p->fontq[p->fonti]; |
|
p->fontq[p->fonti] = f; |
|
} |
|
|
|
/* Set font, save current, discard previous; for \f, .ft, .B etc. */ |
|
void |
|
term_fontrepl(struct termp *p, enum termfont f) |
{ |
{ |
const char *rhs; |
|
size_t sz; |
|
int i; |
|
|
|
rhs = chars_a2ascii(p->symtab, word, len, &sz); |
p->fontl = p->fontq[p->fonti]; |
|
p->fontq[p->fonti] = f; |
|
} |
|
|
if (NULL == rhs) { |
/* Set font, save previous. */ |
#if 0 |
void |
fputs("Unknown special character: ", stderr); |
term_fontpush(struct termp *p, enum termfont f) |
for (i = 0; i < (int)len; i++) |
{ |
fputc(word[i], stderr); |
|
fputc('\n', stderr); |
p->fontl = p->fontq[p->fonti]; |
#endif |
if (++p->fonti == p->fontsz) { |
return; |
p->fontsz += 8; |
|
p->fontq = mandoc_reallocarray(p->fontq, |
|
p->fontsz, sizeof(*p->fontq)); |
} |
} |
for (i = 0; i < (int)sz; i++) |
p->fontq[p->fonti] = f; |
encode(p, rhs[i]); |
|
} |
} |
|
|
|
/* Flush to make the saved pointer current again. */ |
static void |
void |
do_reserved(struct termp *p, const char *word, size_t len) |
term_fontpopq(struct termp *p, int i) |
{ |
{ |
const char *rhs; |
|
size_t sz; |
|
int i; |
|
|
|
rhs = chars_a2res(p->symtab, word, len, &sz); |
assert(i >= 0); |
|
if (p->fonti > i) |
if (NULL == rhs) { |
p->fonti = i; |
#if 0 |
|
fputs("Unknown reserved word: ", stderr); |
|
for (i = 0; i < (int)len; i++) |
|
fputc(word[i], stderr); |
|
fputc('\n', stderr); |
|
#endif |
|
return; |
|
} |
|
for (i = 0; i < (int)sz; i++) |
|
encode(p, rhs[i]); |
|
} |
} |
|
|
|
/* Pop one font off the stack. */ |
|
void |
|
term_fontpop(struct termp *p) |
|
{ |
|
|
|
assert(p->fonti); |
|
p->fonti--; |
|
} |
|
|
/* |
/* |
* Handle an escape sequence: determine its length and pass it to the |
* Handle pwords, partial words, which may be either a single word or a |
* escape-symbol look table. Note that we assume mdoc(3) has validated |
* phrase that cannot be broken down (such as a literal string). This |
* the escape sequence (we assert upon badly-formed escape sequences). |
* handles word styling. |
*/ |
*/ |
static void |
void |
do_escaped(struct termp *p, const char **word) |
term_word(struct termp *p, const char *word) |
{ |
{ |
int j, type; |
struct roffsu su; |
const char *wp; |
const char nbrsp[2] = { ASCII_NBRSP, 0 }; |
|
const char *seq, *cp; |
|
int sz, uc; |
|
size_t csz, lsz, ssz; |
|
enum mandoc_esc esc; |
|
|
wp = *word; |
if ((p->flags & TERMP_NOBUF) == 0) { |
type = 1; |
if ((p->flags & TERMP_NOSPACE) == 0) { |
|
if ((p->flags & TERMP_KEEP) == 0) { |
if (0 == *(++wp)) { |
bufferc(p, ' '); |
*word = wp; |
if (p->flags & TERMP_SENTENCE) |
return; |
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 ('(' == *wp) { |
while ('\0' != *word) { |
wp++; |
if ('\\' != *word) { |
if (0 == *wp || 0 == *(wp + 1)) { |
if (TERMP_NBRWORD & p->flags) { |
*word = 0 == *wp ? wp : wp + 1; |
if (' ' == *word) { |
return; |
encode(p, nbrsp, 1); |
|
word++; |
|
continue; |
|
} |
|
ssz = strcspn(word, "\\ "); |
|
} else |
|
ssz = strcspn(word, "\\"); |
|
encode(p, word, ssz); |
|
word += (int)ssz; |
|
continue; |
} |
} |
|
|
do_special(p, wp, 2); |
word++; |
*word = ++wp; |
esc = mandoc_escape(&word, &seq, &sz); |
return; |
if (ESCAPE_ERROR == esc) |
|
continue; |
|
|
} else if ('*' == *wp) { |
switch (esc) { |
if (0 == *(++wp)) { |
case ESCAPE_UNICODE: |
*word = wp; |
uc = mchars_num2uc(seq + 1, sz - 1); |
return; |
|
} |
|
|
|
switch (*wp) { |
|
case ('('): |
|
wp++; |
|
if (0 == *wp || 0 == *(wp + 1)) { |
|
*word = 0 == *wp ? wp : wp + 1; |
|
return; |
|
} |
|
|
|
do_reserved(p, wp, 2); |
|
*word = ++wp; |
|
return; |
|
case ('['): |
|
type = 0; |
|
break; |
break; |
|
case ESCAPE_NUMBERED: |
|
uc = mchars_num2char(seq, sz); |
|
if (uc < 0) |
|
continue; |
|
break; |
|
case ESCAPE_SPECIAL: |
|
if (p->enc == TERMENC_ASCII) { |
|
cp = mchars_spec2str(seq, sz, &ssz); |
|
if (cp != NULL) |
|
encode(p, cp, ssz); |
|
} else { |
|
uc = mchars_spec2cp(seq, sz); |
|
if (uc > 0) |
|
encode1(p, uc); |
|
} |
|
continue; |
|
case ESCAPE_FONTBOLD: |
|
term_fontrepl(p, TERMFONT_BOLD); |
|
continue; |
|
case ESCAPE_FONTITALIC: |
|
term_fontrepl(p, TERMFONT_UNDER); |
|
continue; |
|
case ESCAPE_FONTBI: |
|
term_fontrepl(p, TERMFONT_BI); |
|
continue; |
|
case ESCAPE_FONT: |
|
case ESCAPE_FONTROMAN: |
|
term_fontrepl(p, TERMFONT_NONE); |
|
continue; |
|
case ESCAPE_FONTPREV: |
|
term_fontlast(p); |
|
continue; |
|
case ESCAPE_BREAK: |
|
bufferc(p, '\n'); |
|
continue; |
|
case ESCAPE_NOSPACE: |
|
if (p->flags & TERMP_BACKAFTER) |
|
p->flags &= ~TERMP_BACKAFTER; |
|
else if (*word == '\0') |
|
p->flags |= (TERMP_NOSPACE | TERMP_NONEWLINE); |
|
continue; |
|
case ESCAPE_HORIZ: |
|
if (a2roffsu(seq, &su, SCALE_EM) == NULL) |
|
continue; |
|
uc = term_hspan(p, &su) / 24; |
|
if (uc > 0) |
|
while (uc-- > 0) |
|
bufferc(p, ASCII_NBRSP); |
|
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 ((seq = a2roffsu(seq, &su, SCALE_EM)) == NULL) |
|
continue; |
|
uc = term_hspan(p, &su) / 24; |
|
if (uc <= 0) { |
|
if (p->tcol->rmargin <= p->tcol->offset) |
|
continue; |
|
lsz = p->tcol->rmargin - p->tcol->offset; |
|
} else |
|
lsz = uc; |
|
if (*seq == '\0') |
|
uc = -1; |
|
else if (*seq == '\\') { |
|
seq++; |
|
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; |
|
default: |
|
uc = -1; |
|
break; |
|
} |
|
} else |
|
uc = *seq; |
|
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: |
|
p->flags |= TERMP_BACKAFTER; |
|
continue; |
|
case ESCAPE_OVERSTRIKE: |
|
cp = seq + sz; |
|
while (seq < cp) { |
|
if (*seq == '\\') { |
|
mandoc_escape(&seq, NULL, NULL); |
|
continue; |
|
} |
|
encode1(p, *seq++); |
|
if (seq < cp) { |
|
if (p->flags & TERMP_BACKBEFORE) |
|
p->flags |= TERMP_BACKAFTER; |
|
else |
|
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; |
default: |
default: |
do_reserved(p, wp, 1); |
continue; |
*word = wp; |
|
return; |
|
} |
} |
|
|
} else if ('f' == *wp) { |
|
if (0 == *(++wp)) { |
|
*word = wp; |
|
return; |
|
} |
|
|
|
switch (*wp) { |
/* |
case ('B'): |
* Common handling for Unicode and numbered |
p->bold++; |
* character escape sequences. |
break; |
*/ |
case ('I'): |
|
p->under++; |
if (p->enc == TERMENC_ASCII) { |
break; |
cp = ascii_uc2str(uc); |
case ('P'): |
encode(p, cp, strlen(cp)); |
/* FALLTHROUGH */ |
} else { |
case ('R'): |
if ((uc < 0x20 && uc != 0x09) || |
p->bold = p->under = 0; |
(uc > 0x7E && uc < 0xA0)) |
break; |
uc = 0xFFFD; |
default: |
encode1(p, uc); |
break; |
|
} |
} |
|
} |
|
p->flags &= ~TERMP_NBRWORD; |
|
} |
|
|
*word = wp; |
static void |
return; |
adjbuf(struct termp_col *c, size_t sz) |
|
{ |
|
if (c->maxcols == 0) |
|
c->maxcols = 1024; |
|
while (c->maxcols <= sz) |
|
c->maxcols <<= 2; |
|
c->buf = mandoc_reallocarray(c->buf, c->maxcols, sizeof(*c->buf)); |
|
} |
|
|
} else if ('[' != *wp) { |
static void |
do_special(p, wp, 1); |
bufferc(struct termp *p, char c) |
*word = wp; |
{ |
|
if (p->flags & TERMP_NOBUF) { |
|
(*p->letter)(p, c); |
return; |
return; |
} |
} |
|
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; |
|
} |
|
|
wp++; |
/* |
for (j = 0; *wp && ']' != *wp; wp++, j++) |
* See encode(). |
/* Loop... */ ; |
* Do this for a single (probably unicode) value. |
|
* Does not check for non-decorated glyphs. |
|
*/ |
|
static void |
|
encode1(struct termp *p, int c) |
|
{ |
|
enum termfont f; |
|
|
if (0 == *wp) { |
if (p->flags & TERMP_NOBUF) { |
*word = wp; |
(*p->letter)(p, c); |
return; |
return; |
} |
} |
|
|
if (type) |
if (p->col + 7 >= p->tcol->maxcols) |
do_special(p, wp - j, (size_t)j); |
adjbuf(p->tcol, p->col + 7); |
else |
|
do_reserved(p, wp - j, (size_t)j); |
|
*word = wp; |
|
} |
|
|
|
|
f = (c == ASCII_HYPH || c > 127 || isgraph(c)) ? |
|
p->fontq[p->fonti] : TERMFONT_NONE; |
|
|
/* |
if (p->flags & TERMP_BACKBEFORE) { |
* Handle pwords, partial words, which may be either a single word or a |
if (p->tcol->buf[p->col - 1] == ' ' || |
* phrase that cannot be broken down (such as a literal string). This |
p->tcol->buf[p->col - 1] == '\t') |
* handles word styling. |
p->col--; |
*/ |
else |
void |
p->tcol->buf[p->col++] = '\b'; |
term_word(struct termp *p, const char *word) |
p->flags &= ~TERMP_BACKBEFORE; |
|
} |
|
if (f == TERMFONT_UNDER || f == TERMFONT_BI) { |
|
p->tcol->buf[p->col++] = '_'; |
|
p->tcol->buf[p->col++] = '\b'; |
|
} |
|
if (f == TERMFONT_BOLD || f == TERMFONT_BI) { |
|
if (c == ASCII_HYPH) |
|
p->tcol->buf[p->col++] = '-'; |
|
else |
|
p->tcol->buf[p->col++] = c; |
|
p->tcol->buf[p->col++] = '\b'; |
|
} |
|
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) { |
|
p->flags |= TERMP_BACKBEFORE; |
|
p->flags &= ~TERMP_BACKAFTER; |
|
} |
|
} |
|
|
|
static void |
|
encode(struct termp *p, const char *word, size_t sz) |
{ |
{ |
const char *sv; |
size_t i; |
|
|
sv = word; |
if (p->flags & TERMP_NOBUF) { |
|
for (i = 0; i < sz; i++) |
|
(*p->letter)(p, word[i]); |
|
return; |
|
} |
|
|
if (word[0] && 0 == word[1]) |
if (p->col + 2 + (sz * 5) >= p->tcol->maxcols) |
switch (word[0]) { |
adjbuf(p->tcol, p->col + 2 + (sz * 5)); |
case('.'): |
|
/* FALLTHROUGH */ |
|
case(','): |
|
/* FALLTHROUGH */ |
|
case(';'): |
|
/* FALLTHROUGH */ |
|
case(':'): |
|
/* FALLTHROUGH */ |
|
case('?'): |
|
/* FALLTHROUGH */ |
|
case('!'): |
|
/* FALLTHROUGH */ |
|
case(')'): |
|
/* FALLTHROUGH */ |
|
case(']'): |
|
/* FALLTHROUGH */ |
|
case('}'): |
|
if ( ! (TERMP_IGNDELIM & p->flags)) |
|
p->flags |= TERMP_NOSPACE; |
|
break; |
|
default: |
|
break; |
|
} |
|
|
|
if ( ! (TERMP_NOSPACE & p->flags)) |
for (i = 0; i < sz; i++) { |
buffer(p, ' '); |
if (ASCII_HYPH == word[i] || |
|
isgraph((unsigned char)word[i])) |
|
encode1(p, word[i]); |
|
else { |
|
if (p->tcol->lastcol <= p->col || |
|
(word[i] != ' ' && word[i] != ASCII_NBRSP)) |
|
p->tcol->buf[p->col] = word[i]; |
|
p->col++; |
|
|
if ( ! (p->flags & TERMP_NONOSPACE)) |
/* |
p->flags &= ~TERMP_NOSPACE; |
* Postpone the effect of \z while handling |
|
* an overstrike sequence from ascii_uc2str(). |
|
*/ |
|
|
for ( ; *word; word++) |
if (word[i] == '\b' && |
if ('\\' != *word) |
(p->flags & TERMP_BACKBEFORE)) { |
encode(p, *word); |
p->flags &= ~TERMP_BACKBEFORE; |
else |
p->flags |= TERMP_BACKAFTER; |
do_escaped(p, &word); |
} |
|
} |
|
} |
|
if (p->tcol->lastcol < p->col) |
|
p->tcol->lastcol = p->col; |
|
} |
|
|
if (sv[0] && 0 == sv[1]) |
void |
switch (sv[0]) { |
term_setwidth(struct termp *p, const char *wstr) |
case('('): |
{ |
/* FALLTHROUGH */ |
struct roffsu su; |
case('['): |
int iop, width; |
/* FALLTHROUGH */ |
|
case('{'): |
iop = 0; |
p->flags |= TERMP_NOSPACE; |
width = 0; |
|
if (NULL != wstr) { |
|
switch (*wstr) { |
|
case '+': |
|
iop = 1; |
|
wstr++; |
break; |
break; |
|
case '-': |
|
iop = -1; |
|
wstr++; |
|
break; |
default: |
default: |
break; |
break; |
} |
} |
|
if (a2roffsu(wstr, &su, SCALE_MAX) != NULL) |
|
width = term_hspan(p, &su); |
|
else |
|
iop = 0; |
|
} |
|
(*p->setwidth)(p, iop, width); |
} |
} |
|
|
|
size_t |
/* |
term_len(const struct termp *p, size_t sz) |
* Insert a single character into the line-buffer. If the buffer's |
|
* space is exceeded, then allocate more space by doubling the buffer |
|
* size. |
|
*/ |
|
static void |
|
buffer(struct termp *p, char c) |
|
{ |
{ |
size_t s; |
|
|
|
if (p->col + 1 >= p->maxcols) { |
return (*p->width)(p, ' ') * sz; |
if (0 == p->maxcols) |
|
p->maxcols = 256; |
|
s = p->maxcols * 2; |
|
p->buf = realloc(p->buf, s); |
|
if (NULL == p->buf) |
|
err(1, "realloc"); /* FIXME: shouldn't be here! */ |
|
p->maxcols = s; |
|
} |
|
p->buf[(int)(p->col)++] = c; |
|
} |
} |
|
|
|
static size_t |
static void |
cond_width(const struct termp *p, int c, int *skip) |
encode(struct termp *p, char c) |
|
{ |
{ |
|
|
if (' ' != c) { |
if (*skip) { |
if (p->under) { |
(*skip) = 0; |
buffer(p, '_'); |
return 0; |
buffer(p, 8); |
} else |
} |
return (*p->width)(p, c); |
if (p->bold) { |
|
buffer(p, c); |
|
buffer(p, 8); |
|
} |
|
} |
|
buffer(p, c); |
|
} |
} |
|
|
|
|
size_t |
size_t |
term_vspan(const struct roffsu *su) |
term_strlen(const struct termp *p, const char *cp) |
{ |
{ |
double r; |
size_t sz, rsz, i; |
|
int ssz, skip, uc; |
|
const char *seq, *rhs; |
|
enum mandoc_esc esc; |
|
static const char rej[] = { '\\', ASCII_NBRSP, ASCII_HYPH, |
|
ASCII_BREAK, '\0' }; |
|
|
switch (su->unit) { |
/* |
case (SCALE_CM): |
* Account for escaped sequences within string length |
r = su->scale * 2; |
* calculations. This follows the logic in term_word() as we |
break; |
* must calculate the width of produced strings. |
case (SCALE_IN): |
*/ |
r = su->scale * 6; |
|
break; |
sz = 0; |
case (SCALE_PC): |
skip = 0; |
r = su->scale; |
while ('\0' != *cp) { |
break; |
rsz = strcspn(cp, rej); |
case (SCALE_PT): |
for (i = 0; i < rsz; i++) |
r = su->scale / 8; |
sz += cond_width(p, *cp++, &skip); |
break; |
|
case (SCALE_MM): |
switch (*cp) { |
r = su->scale / 1000; |
case '\\': |
break; |
cp++; |
case (SCALE_VS): |
esc = mandoc_escape(&cp, &seq, &ssz); |
r = su->scale; |
if (ESCAPE_ERROR == esc) |
break; |
continue; |
default: |
|
r = su->scale - 1; |
rhs = NULL; |
break; |
|
|
switch (esc) { |
|
case ESCAPE_UNICODE: |
|
uc = mchars_num2uc(seq + 1, ssz - 1); |
|
break; |
|
case ESCAPE_NUMBERED: |
|
uc = mchars_num2char(seq, ssz); |
|
if (uc < 0) |
|
continue; |
|
break; |
|
case ESCAPE_SPECIAL: |
|
if (p->enc == TERMENC_ASCII) { |
|
rhs = mchars_spec2str(seq, ssz, &rsz); |
|
if (rhs != NULL) |
|
break; |
|
} else { |
|
uc = mchars_spec2cp(seq, ssz); |
|
if (uc > 0) |
|
sz += cond_width(p, uc, &skip); |
|
} |
|
continue; |
|
case ESCAPE_SKIPCHAR: |
|
skip = 1; |
|
continue; |
|
case ESCAPE_OVERSTRIKE: |
|
rsz = 0; |
|
rhs = seq + ssz; |
|
while (seq < rhs) { |
|
if (*seq == '\\') { |
|
mandoc_escape(&seq, NULL, NULL); |
|
continue; |
|
} |
|
i = (*p->width)(p, *seq++); |
|
if (rsz < i) |
|
rsz = i; |
|
} |
|
sz += rsz; |
|
continue; |
|
default: |
|
continue; |
|
} |
|
|
|
/* |
|
* Common handling for Unicode and numbered |
|
* character escape sequences. |
|
*/ |
|
|
|
if (rhs == NULL) { |
|
if (p->enc == TERMENC_ASCII) { |
|
rhs = ascii_uc2str(uc); |
|
rsz = strlen(rhs); |
|
} else { |
|
if ((uc < 0x20 && uc != 0x09) || |
|
(uc > 0x7E && uc < 0xA0)) |
|
uc = 0xFFFD; |
|
sz += cond_width(p, uc, &skip); |
|
continue; |
|
} |
|
} |
|
|
|
if (skip) { |
|
skip = 0; |
|
break; |
|
} |
|
|
|
/* |
|
* Common handling for all escape sequences |
|
* printing more than one character. |
|
*/ |
|
|
|
for (i = 0; i < rsz; i++) |
|
sz += (*p->width)(p, *rhs++); |
|
break; |
|
case ASCII_NBRSP: |
|
sz += cond_width(p, ' ', &skip); |
|
cp++; |
|
break; |
|
case ASCII_HYPH: |
|
sz += cond_width(p, '-', &skip); |
|
cp++; |
|
break; |
|
default: |
|
break; |
|
} |
} |
} |
|
|
if (r < 0.0) |
return sz; |
r = 0.0; |
|
return(/* LINTED */(size_t) |
|
r); |
|
} |
} |
|
|
|
int |
size_t |
term_vspan(const struct termp *p, const struct roffsu *su) |
term_hspan(const struct roffsu *su) |
|
{ |
{ |
double r; |
double r; |
|
int ri; |
|
|
/* XXX: CM, IN, and PT are approximations. */ |
|
|
|
switch (su->unit) { |
switch (su->unit) { |
case (SCALE_CM): |
case SCALE_BU: |
r = 4 * su->scale; |
r = su->scale / 40.0; |
break; |
break; |
case (SCALE_IN): |
case SCALE_CM: |
/* XXX: this is an approximation. */ |
r = su->scale * 6.0 / 2.54; |
r = 10 * su->scale; |
|
break; |
break; |
case (SCALE_PC): |
case SCALE_FS: |
r = (10 * su->scale) / 6; |
r = su->scale * 65536.0 / 40.0; |
break; |
break; |
case (SCALE_PT): |
case SCALE_IN: |
r = (10 * su->scale) / 72; |
r = su->scale * 6.0; |
break; |
break; |
case (SCALE_MM): |
case SCALE_MM: |
r = su->scale / 1000; /* FIXME: double-check. */ |
r = su->scale * 0.006; |
break; |
break; |
case (SCALE_VS): |
case SCALE_PC: |
r = su->scale * 2 - 1; /* FIXME: double-check. */ |
r = su->scale; |
break; |
break; |
default: |
case SCALE_PT: |
|
r = su->scale / 12.0; |
|
break; |
|
case SCALE_EN: |
|
case SCALE_EM: |
|
r = su->scale * 0.6; |
|
break; |
|
case SCALE_VS: |
r = su->scale; |
r = su->scale; |
break; |
break; |
|
default: |
|
abort(); |
} |
} |
|
ri = r > 0.0 ? r + 0.4995 : r - 0.4995; |
if (r < 0.0) |
return ri < 66 ? ri : 1; |
r = 0.0; |
|
return((size_t)/* LINTED */ |
|
r); |
|
} |
} |
|
|
|
/* |
|
* Convert a scaling width to basic units, rounding down. |
|
*/ |
|
int |
|
term_hspan(const struct termp *p, const struct roffsu *su) |
|
{ |
|
|
|
return (*p->hspan)(p, su); |
|
} |