version 1.89, 2009/07/16 13:17:51 |
version 1.258, 2016/08/10 11:03:43 |
|
|
/* $Id$ */ |
/* $Id$ */ |
/* |
/* |
* Copyright (c) 2008, 2009 Kristaps Dzonsons <kristaps@kth.se> |
* Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv> |
|
* Copyright (c) 2010-2016 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 "mandoc.h" |
|
#include "mandoc_aux.h" |
|
#include "out.h" |
#include "term.h" |
#include "term.h" |
#include "man.h" |
#include "main.h" |
#include "mdoc.h" |
|
|
|
extern int man_run(struct termp *, |
static size_t cond_width(const struct termp *, int, int *); |
const struct man *); |
static void adjbuf(struct termp *p, size_t); |
extern int mdoc_run(struct termp *, |
static void bufferc(struct termp *, char); |
const struct mdoc *); |
static void encode(struct termp *, const char *, size_t); |
|
static void encode1(struct termp *, int); |
|
|
static struct termp *term_alloc(enum termenc); |
|
static void term_free(struct termp *); |
|
static void term_pescape(struct termp *, const char **); |
|
static void term_nescape(struct termp *, |
|
const char *, size_t); |
|
static void term_chara(struct termp *, char); |
|
static void term_encodea(struct termp *, char); |
|
static int term_isopendelim(const char *); |
|
static int term_isclosedelim(const char *); |
|
|
|
|
|
void * |
|
ascii_alloc(void) |
|
{ |
|
|
|
return(term_alloc(TERMENC_ASCII)); |
|
} |
|
|
|
|
|
int |
|
terminal_man(void *arg, const struct man *man) |
|
{ |
|
struct termp *p; |
|
|
|
p = (struct termp *)arg; |
|
if (NULL == p->symtab) |
|
p->symtab = term_ascii2htab(); |
|
|
|
return(man_run(p, man)); |
|
} |
|
|
|
|
|
int |
|
terminal_mdoc(void *arg, const struct mdoc *mdoc) |
|
{ |
|
struct termp *p; |
|
|
|
p = (struct termp *)arg; |
|
if (NULL == p->symtab) |
|
p->symtab = term_ascii2htab(); |
|
|
|
return(mdoc_run(p, mdoc)); |
|
} |
|
|
|
|
|
void |
void |
terminal_free(void *arg) |
|
{ |
|
|
|
term_free((struct termp *)arg); |
|
} |
|
|
|
|
|
static void |
|
term_free(struct termp *p) |
term_free(struct termp *p) |
{ |
{ |
|
|
if (p->buf) |
free(p->buf); |
free(p->buf); |
free(p->fontq); |
if (TERMENC_ASCII == p->enc && p->symtab) |
|
term_asciifree(p->symtab); |
|
|
|
free(p); |
free(p); |
} |
} |
|
|
|
void |
static struct termp * |
term_begin(struct termp *p, term_margin head, |
term_alloc(enum termenc enc) |
term_margin foot, const struct roff_meta *arg) |
{ |
{ |
struct termp *p; |
|
|
|
if (NULL == (p = malloc(sizeof(struct termp)))) |
p->headf = head; |
err(1, "malloc"); |
p->footf = foot; |
bzero(p, sizeof(struct termp)); |
p->argf = arg; |
p->maxrmargin = 78; |
(*p->begin)(p); |
p->enc = enc; |
|
return(p); |
|
} |
} |
|
|
|
void |
static int |
term_end(struct termp *p) |
term_isclosedelim(const char *p) |
|
{ |
{ |
|
|
if ( ! (*p && 0 == *(p + 1))) |
(*p->end)(p); |
return(0); |
|
|
|
switch (*p) { |
|
case('.'): |
|
/* FALLTHROUGH */ |
|
case(','): |
|
/* FALLTHROUGH */ |
|
case(';'): |
|
/* FALLTHROUGH */ |
|
case(':'): |
|
/* FALLTHROUGH */ |
|
case('?'): |
|
/* FALLTHROUGH */ |
|
case('!'): |
|
/* FALLTHROUGH */ |
|
case(')'): |
|
/* FALLTHROUGH */ |
|
case(']'): |
|
/* FALLTHROUGH */ |
|
case('}'): |
|
return(1); |
|
default: |
|
break; |
|
} |
|
|
|
return(0); |
|
} |
} |
|
|
|
|
static int |
|
term_isopendelim(const char *p) |
|
{ |
|
|
|
if ( ! (*p && 0 == *(p + 1))) |
|
return(0); |
|
|
|
switch (*p) { |
|
case('('): |
|
/* FALLTHROUGH */ |
|
case('['): |
|
/* FALLTHROUGH */ |
|
case('{'): |
|
return(1); |
|
default: |
|
break; |
|
} |
|
|
|
return(0); |
|
} |
|
|
|
|
|
/* |
/* |
* 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_DANGLE: Do not break the output line at the right margin, |
* - TERMP_DANGLE: don't newline when TERMP_NOBREAK is specified and |
* append the next chunk after it even if this one is too long. |
* the line is overrun, and don't pad-right if it's underrun. |
* To be used together with TERMP_NOBREAK. |
* |
* - TERMP_HANG: Like TERMP_DANGLE, and also suppress padding before |
* - TERMP_HANG: like TERMP_DANGLE, but doesn't newline when |
* the next chunk if this column is not full. |
* 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). |
|
* |
|
* FIXME: newline breaks occur (in groff) also occur when a single |
|
* space follows a NOBREAK! |
|
*/ |
*/ |
void |
void |
term_flushln(struct termp *p) |
term_flushln(struct termp *p) |
{ |
{ |
int i, j; |
size_t i; /* current input position in p->buf */ |
size_t vbl, vsz, vis, maxvis, mmax, bp; |
int ntab; /* number of tabs to prepend */ |
static int sv = -1; |
size_t vis; /* current visual position on output */ |
|
size_t vbl; /* number of blanks to prepend to output */ |
|
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->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. |
* First, establish the maximum columns of "visible" content. |
* This is usually the difference between the right-margin and |
* This is usually the difference between the right-margin and |
* an indentation, but can be, for tagged lists or columns, a |
* an indentation, but can be, for tagged lists or columns, a |
* small set of values. |
* 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; |
|
|
assert(p->offset < p->rmargin); |
if (p->flags & TERMP_NOBREAK) { |
maxvis = p->rmargin - p->offset; |
dv = p->maxrmargin > p->offset ? |
mmax = p->maxrmargin - p->offset; |
p->maxrmargin - p->offset : 0; |
bp = TERMP_NOBREAK & p->flags ? mmax : maxvis; |
bp = (int)dv > p->overstep ? |
vis = 0; |
dv - (size_t)p->overstep : 0; |
|
} else |
|
bp = maxvis; |
|
|
if (sv >= 0) { |
|
vis = (size_t)sv; |
|
sv = -1; |
|
} |
|
|
|
/* |
/* |
* If in the standard case (left-justified), then begin with our |
* Calculate the required amount of padding. |
* indentation, otherwise (columns, etc.) just start spitting |
|
* out text. |
|
*/ |
*/ |
|
vbl = p->offset + p->overstep > p->viscol ? |
|
p->offset + p->overstep - p->viscol : 0; |
|
|
if ( ! (p->flags & TERMP_NOLPAD)) |
vis = vend = 0; |
/* LINTED */ |
i = 0; |
for (j = 0; j < (int)p->offset; j++) |
|
putchar(' '); |
|
|
|
for (i = 0; i < (int)p->col; i++) { |
while (i < p->col) { |
/* |
/* |
|
* Handle literal tab characters: collapse all |
|
* subsequent tabs into a single huge set of spaces. |
|
*/ |
|
ntab = 0; |
|
while (i < p->col && '\t' == p->buf[i]) { |
|
vend = (vis / p->tabwidth + 1) * p->tabwidth; |
|
vbl += vend - vis; |
|
vis = vend; |
|
ntab++; |
|
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 |
* generates a non-printing word, which is valid (the |
* generates a non-printing word, which is valid (the |
* space is printed according to regular spacing rules). |
* space is printed according to regular spacing rules). |
*/ |
*/ |
|
|
/* LINTED */ |
for (j = i, jhy = 0; j < p->col; j++) { |
for (j = i, vsz = 0; j < (int)p->col; j++) { |
if (' ' == p->buf[j] || '\t' == p->buf[j]) |
if (' ' == p->buf[j]) |
|
break; |
break; |
else if (8 == p->buf[j]) |
|
vsz--; |
/* Back over the last printed character. */ |
else |
if (8 == p->buf[j]) { |
vsz++; |
assert(j); |
|
vend -= (*p->width)(p, p->buf[j - 1]); |
|
continue; |
|
} |
|
|
|
/* Regular word. */ |
|
/* Break at the hyphen point if we overrun. */ |
|
if (vend > vis && vend < bp && |
|
(ASCII_HYPH == p->buf[j] || |
|
ASCII_BREAK == p->buf[j])) |
|
jhy = j; |
|
|
|
/* |
|
* Hyphenation now decided, put back a real |
|
* hyphen such that we get the correct width. |
|
*/ |
|
if (ASCII_HYPH == p->buf[j]) |
|
p->buf[j] = '-'; |
|
|
|
vend += (*p->width)(p, p->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) { |
if (vend > bp && 0 == jhy && vis > 0) { |
putchar('\n'); |
vend -= vis; |
if (TERMP_NOBREAK & p->flags) { |
(*p->endline)(p); |
for (j = 0; j < (int)p->rmargin; j++) |
p->viscol = 0; |
putchar(' '); |
if (TERMP_BRIND & p->flags) { |
vis = p->rmargin - p->offset; |
vbl = p->rmargin; |
} else { |
vend += p->rmargin; |
for (j = 0; j < (int)p->offset; j++) |
vend -= p->offset; |
putchar(' '); |
} else |
vis = 0; |
vbl = p->offset; |
} |
|
} else { |
/* use pending tabs on the new line */ |
for (j = 0; j < (int)vbl; j++) |
|
putchar(' '); |
if (0 < ntab) |
vis += vbl; |
vbl += ntab * p->tabwidth; |
|
|
|
/* |
|
* Remove the p->overstep width. |
|
* Again, if p->overstep is negative, |
|
* sign extension does the right thing. |
|
*/ |
|
|
|
bp += (size_t)p->overstep; |
|
p->overstep = 0; |
} |
} |
|
|
/* |
/* Write out the [remaining] word. */ |
* Finally, write out the word. |
for ( ; i < p->col; i++) { |
*/ |
if (vend > bp && jhy > 0 && i > jhy) |
for ( ; i < (int)p->col; i++) { |
|
if (' ' == p->buf[i]) |
|
break; |
break; |
putchar(p->buf[i]); |
if ('\t' == p->buf[i]) |
|
break; |
|
if (' ' == p->buf[i]) { |
|
j = i; |
|
while (i < p->col && ' ' == p->buf[i]) |
|
i++; |
|
dv = (i - j) * (*p->width)(p, ' '); |
|
vbl += dv; |
|
vend += dv; |
|
break; |
|
} |
|
if (ASCII_NBRSP == p->buf[i]) { |
|
vbl += (*p->width)(p, ' '); |
|
continue; |
|
} |
|
if (ASCII_BREAK == p->buf[i]) |
|
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->buf[i]); |
|
if (8 == p->buf[i]) |
|
p->viscol -= (*p->width)(p, p->buf[i-1]); |
|
else |
|
p->viscol += (*p->width)(p, p->buf[i]); |
} |
} |
vis += vsz; |
vis = vend; |
} |
} |
|
|
/* |
/* |
* If we've overstepped our maximum visible no-break space, then |
* If there was trailing white space, it was not printed; |
* cause a newline and offset at the right margin. |
* so reset the cursor position accordingly. |
*/ |
*/ |
|
if (vis > vbl) |
|
vis -= vbl; |
|
else |
|
vis = 0; |
|
|
if ((TERMP_NOBREAK & p->flags) && vis >= maxvis) { |
p->col = 0; |
if ( ! (TERMP_DANGLE & p->flags) && |
p->overstep = 0; |
! (TERMP_HANG & p->flags)) { |
p->flags &= ~(TERMP_BACKAFTER | TERMP_BACKBEFORE); |
putchar('\n'); |
|
for (i = 0; i < (int)p->rmargin; i++) |
if ( ! (TERMP_NOBREAK & p->flags)) { |
putchar(' '); |
p->viscol = 0; |
} |
(*p->endline)(p); |
if (TERMP_HANG & p->flags) |
|
sv = (int)(vis - maxvis); |
|
p->col = 0; |
|
return; |
return; |
} |
} |
|
|
/* |
if (TERMP_HANG & p->flags) { |
* If we're not to right-marginalise it (newline), then instead |
p->overstep += (int)(p->offset + vis - p->rmargin + |
* pad to the right margin and stay off. |
p->trailspace * (*p->width)(p, ' ')); |
*/ |
|
|
|
if (p->flags & TERMP_NOBREAK) { |
/* |
if ( ! (TERMP_DANGLE & p->flags)) |
* If we have overstepped the margin, temporarily move |
for ( ; vis < maxvis; vis++) |
* it to the right and flag the rest of the line to be |
putchar(' '); |
* shorter. |
} else |
* If there is a request to keep the columns together, |
putchar('\n'); |
* allow negative overstep when the column is not full. |
|
*/ |
|
if (p->trailspace && p->overstep < 0) |
|
p->overstep = 0; |
|
return; |
|
|
p->col = 0; |
} else if (TERMP_DANGLE & p->flags) |
} |
return; |
|
|
|
/* Trailing whitespace is significant in some columns. */ |
|
if (vis && vbl && (TERMP_BRTRSP & p->flags)) |
|
vis += vbl; |
|
|
/* |
/* If the column was overrun, break the line. */ |
|
if (maxvis < vis + p->trailspace * (*p->width)(p, ' ')) { |
|
(*p->endline)(p); |
|
p->viscol = 0; |
|
} |
|
} |
|
|
|
/* |
* 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 344 term_newln(struct termp *p) |
|
Line 314 term_newln(struct termp *p) |
|
{ |
{ |
|
|
p->flags |= TERMP_NOSPACE; |
p->flags |= TERMP_NOSPACE; |
if (0 == p->col) { |
if (p->col || 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 364 term_vspace(struct termp *p) |
|
Line 329 term_vspace(struct termp *p) |
|
{ |
{ |
|
|
term_newln(p); |
term_newln(p); |
putchar('\n'); |
p->viscol = 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; |
|
|
/* |
f = p->fontl; |
* Determine the symbol indicated by an escape sequences, that is, one |
p->fontl = p->fontq[p->fonti]; |
* starting with a backslash. Once done, we pass this value into the |
p->fontq[p->fonti] = f; |
* output buffer by way of the symbol table. |
} |
*/ |
|
static void |
/* Set font, save current, discard previous; for \f, .ft, .B etc. */ |
term_nescape(struct termp *p, const char *word, size_t len) |
void |
|
term_fontrepl(struct termp *p, enum termfont f) |
{ |
{ |
const char *rhs; |
|
size_t sz; |
|
int i; |
|
|
|
rhs = term_a2ascii(p->symtab, word, len, &sz); |
p->fontl = p->fontq[p->fonti]; |
|
p->fontq[p->fonti] = f; |
|
} |
|
|
if (rhs) |
/* Set font, save previous. */ |
for (i = 0; i < (int)sz; i++) |
void |
term_encodea(p, rhs[i]); |
term_fontpush(struct termp *p, enum termfont f) |
|
{ |
|
|
|
p->fontl = p->fontq[p->fonti]; |
|
if (++p->fonti == p->fontsz) { |
|
p->fontsz += 8; |
|
p->fontq = mandoc_reallocarray(p->fontq, |
|
p->fontsz, sizeof(*p->fontq)); |
|
} |
|
p->fontq[p->fonti] = f; |
} |
} |
|
|
|
/* Flush to make the saved pointer current again. */ |
|
void |
|
term_fontpopq(struct termp *p, int i) |
|
{ |
|
|
|
assert(i >= 0); |
|
if (p->fonti > i) |
|
p->fonti = 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 |
term_pescape(struct termp *p, const char **word) |
term_word(struct termp *p, const char *word) |
{ |
{ |
int j; |
const char nbrsp[2] = { ASCII_NBRSP, 0 }; |
const char *wp; |
const char *seq, *cp; |
|
int sz, uc; |
|
size_t ssz; |
|
enum mandoc_esc esc; |
|
|
wp = *word; |
if ( ! (TERMP_NOSPACE & p->flags)) { |
|
if ( ! (TERMP_KEEP & p->flags)) { |
if (0 == *(++wp)) { |
bufferc(p, ' '); |
*word = wp; |
if (TERMP_SENTENCE & p->flags) |
return; |
bufferc(p, ' '); |
|
} else |
|
bufferc(p, ASCII_NBRSP); |
} |
} |
|
if (TERMP_PREKEEP & p->flags) |
|
p->flags |= TERMP_KEEP; |
|
|
if ('(' == *wp) { |
if ( ! (p->flags & TERMP_NONOSPACE)) |
wp++; |
p->flags &= ~TERMP_NOSPACE; |
if (0 == *wp || 0 == *(wp + 1)) { |
else |
*word = 0 == *wp ? wp : wp + 1; |
p->flags |= TERMP_NOSPACE; |
return; |
|
} |
|
|
|
term_nescape(p, wp, 2); |
p->flags &= ~(TERMP_SENTENCE | TERMP_NONEWLINE); |
*word = ++wp; |
p->skipvsp = 0; |
return; |
|
|
|
} else if ('*' == *wp) { |
while ('\0' != *word) { |
if (0 == *(++wp)) { |
if ('\\' != *word) { |
*word = wp; |
if (TERMP_NBRWORD & p->flags) { |
return; |
if (' ' == *word) { |
|
encode(p, nbrsp, 1); |
|
word++; |
|
continue; |
|
} |
|
ssz = strcspn(word, "\\ "); |
|
} else |
|
ssz = strcspn(word, "\\"); |
|
encode(p, word, ssz); |
|
word += (int)ssz; |
|
continue; |
} |
} |
|
|
switch (*wp) { |
word++; |
case ('('): |
esc = mandoc_escape(&word, &seq, &sz); |
wp++; |
if (ESCAPE_ERROR == esc) |
if (0 == *wp || 0 == *(wp + 1)) { |
continue; |
*word = 0 == *wp ? wp : wp + 1; |
|
return; |
|
} |
|
|
|
term_nescape(p, wp, 2); |
switch (esc) { |
*word = ++wp; |
case ESCAPE_UNICODE: |
return; |
uc = mchars_num2uc(seq + 1, sz - 1); |
case ('['): |
|
break; |
break; |
default: |
case ESCAPE_NUMBERED: |
term_nescape(p, wp, 1); |
uc = mchars_num2char(seq, sz); |
*word = wp; |
if (uc < 0) |
return; |
continue; |
} |
|
|
|
} else if ('f' == *wp) { |
|
if (0 == *(++wp)) { |
|
*word = wp; |
|
return; |
|
} |
|
|
|
switch (*wp) { |
|
case ('B'): |
|
p->flags |= TERMP_BOLD; |
|
break; |
break; |
case ('I'): |
case ESCAPE_SPECIAL: |
p->flags |= TERMP_UNDER; |
if (p->enc == TERMENC_ASCII) { |
break; |
cp = mchars_spec2str(seq, sz, &ssz); |
case ('P'): |
if (cp != NULL) |
/* FALLTHROUGH */ |
encode(p, cp, ssz); |
case ('R'): |
} else { |
p->flags &= ~TERMP_STYLE; |
uc = mchars_spec2cp(seq, sz); |
break; |
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_NOSPACE: |
|
if (p->flags & TERMP_BACKAFTER) |
|
p->flags &= ~TERMP_BACKAFTER; |
|
else if (*word == '\0') |
|
p->flags |= (TERMP_NOSPACE | TERMP_NONEWLINE); |
|
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->col > 2 && |
|
(p->buf[p->col - 1] == ' ' || |
|
p->buf[p->col - 1] == '\t')) |
|
p->col -= 2; |
|
continue; |
default: |
default: |
break; |
continue; |
} |
} |
|
|
*word = wp; |
/* |
return; |
* Common handling for Unicode and numbered |
|
* character escape sequences. |
|
*/ |
|
|
} else if ('[' != *wp) { |
if (p->enc == TERMENC_ASCII) { |
term_nescape(p, wp, 1); |
cp = ascii_uc2str(uc); |
*word = wp; |
encode(p, cp, strlen(cp)); |
return; |
} else { |
|
if ((uc < 0x20 && uc != 0x09) || |
|
(uc > 0x7E && uc < 0xA0)) |
|
uc = 0xFFFD; |
|
encode1(p, uc); |
|
} |
} |
} |
|
p->flags &= ~TERMP_NBRWORD; |
|
} |
|
|
wp++; |
static void |
for (j = 0; *wp && ']' != *wp; wp++, j++) |
adjbuf(struct termp *p, size_t sz) |
/* Loop... */ ; |
{ |
|
|
if (0 == *wp) { |
if (0 == p->maxcols) |
*word = wp; |
p->maxcols = 1024; |
return; |
while (sz >= p->maxcols) |
} |
p->maxcols <<= 2; |
|
|
term_nescape(p, wp - j, (size_t)j); |
p->buf = mandoc_reallocarray(p->buf, p->maxcols, sizeof(int)); |
*word = wp; |
|
} |
} |
|
|
|
static void |
|
bufferc(struct termp *p, char c) |
|
{ |
|
|
|
if (p->col + 1 >= p->maxcols) |
|
adjbuf(p, p->col + 1); |
|
|
|
p->buf[p->col++] = c; |
|
} |
|
|
/* |
/* |
* Handle pwords, partial words, which may be either a single word or a |
* See encode(). |
* phrase that cannot be broken down (such as a literal string). This |
* Do this for a single (probably unicode) value. |
* handles word styling. |
* Does not check for non-decorated glyphs. |
*/ |
*/ |
void |
static void |
term_word(struct termp *p, const char *word) |
encode1(struct termp *p, int c) |
{ |
{ |
const char *sv; |
enum termfont f; |
|
|
if (term_isclosedelim(word)) |
if (p->col + 7 >= p->maxcols) |
if ( ! (TERMP_IGNDELIM & p->flags)) |
adjbuf(p, p->col + 7); |
p->flags |= TERMP_NOSPACE; |
|
|
|
if ( ! (TERMP_NOSPACE & p->flags)) |
f = (c == ASCII_HYPH || c > 127 || isgraph(c)) ? |
term_chara(p, ' '); |
p->fontq[p->fonti] : TERMFONT_NONE; |
|
|
if ( ! (p->flags & TERMP_NONOSPACE)) |
if (p->flags & TERMP_BACKBEFORE) { |
p->flags &= ~TERMP_NOSPACE; |
if (p->buf[p->col - 1] == ' ' || p->buf[p->col - 1] == '\t') |
|
p->col--; |
|
else |
|
p->buf[p->col++] = 8; |
|
p->flags &= ~TERMP_BACKBEFORE; |
|
} |
|
if (TERMFONT_UNDER == f || TERMFONT_BI == f) { |
|
p->buf[p->col++] = '_'; |
|
p->buf[p->col++] = 8; |
|
} |
|
if (TERMFONT_BOLD == f || TERMFONT_BI == f) { |
|
if (ASCII_HYPH == c) |
|
p->buf[p->col++] = '-'; |
|
else |
|
p->buf[p->col++] = c; |
|
p->buf[p->col++] = 8; |
|
} |
|
p->buf[p->col++] = c; |
|
if (p->flags & TERMP_BACKAFTER) { |
|
p->flags |= TERMP_BACKBEFORE; |
|
p->flags &= ~TERMP_BACKAFTER; |
|
} |
|
} |
|
|
/* |
static void |
* If ANSI (word-length styling), then apply our style now, |
encode(struct termp *p, const char *word, size_t sz) |
* before the word. |
{ |
*/ |
size_t i; |
|
|
for (sv = word; *word; word++) |
if (p->col + 2 + (sz * 5) >= p->maxcols) |
if ('\\' != *word) |
adjbuf(p, p->col + 2 + (sz * 5)); |
term_encodea(p, *word); |
|
else |
|
term_pescape(p, &word); |
|
|
|
if (term_isopendelim(sv)) |
for (i = 0; i < sz; i++) { |
p->flags |= TERMP_NOSPACE; |
if (ASCII_HYPH == word[i] || |
|
isgraph((unsigned char)word[i])) |
|
encode1(p, word[i]); |
|
else |
|
p->buf[p->col++] = word[i]; |
|
} |
} |
} |
|
|
|
void |
/* |
term_setwidth(struct termp *p, const char *wstr) |
* 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 |
|
term_chara(struct termp *p, char c) |
|
{ |
{ |
size_t s; |
struct roffsu su; |
|
int iop, width; |
|
|
if (p->col + 1 >= p->maxcols) { |
iop = 0; |
if (0 == p->maxcols) |
width = 0; |
p->maxcols = 256; |
if (NULL != wstr) { |
s = p->maxcols * 2; |
switch (*wstr) { |
p->buf = realloc(p->buf, s); |
case '+': |
if (NULL == p->buf) |
iop = 1; |
err(1, "realloc"); |
wstr++; |
p->maxcols = s; |
break; |
|
case '-': |
|
iop = -1; |
|
wstr++; |
|
break; |
|
default: |
|
break; |
|
} |
|
if (a2roffsu(wstr, &su, SCALE_MAX)) |
|
width = term_hspan(p, &su); |
|
else |
|
iop = 0; |
} |
} |
p->buf[(int)(p->col)++] = c; |
(*p->setwidth)(p, iop, width); |
} |
} |
|
|
|
size_t |
|
term_len(const struct termp *p, size_t sz) |
|
{ |
|
|
static void |
return (*p->width)(p, ' ') * sz; |
term_encodea(struct termp *p, char c) |
} |
|
|
|
static size_t |
|
cond_width(const struct termp *p, int c, int *skip) |
{ |
{ |
|
|
if (' ' != c && TERMP_STYLE & p->flags) { |
if (*skip) { |
if (TERMP_BOLD & p->flags) { |
(*skip) = 0; |
term_chara(p, c); |
return 0; |
term_chara(p, 8); |
} else |
|
return (*p->width)(p, c); |
|
} |
|
|
|
size_t |
|
term_strlen(const struct termp *p, const char *cp) |
|
{ |
|
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' }; |
|
|
|
/* |
|
* Account for escaped sequences within string length |
|
* calculations. This follows the logic in term_word() as we |
|
* must calculate the width of produced strings. |
|
*/ |
|
|
|
sz = 0; |
|
skip = 0; |
|
while ('\0' != *cp) { |
|
rsz = strcspn(cp, rej); |
|
for (i = 0; i < rsz; i++) |
|
sz += cond_width(p, *cp++, &skip); |
|
|
|
switch (*cp) { |
|
case '\\': |
|
cp++; |
|
esc = mandoc_escape(&cp, &seq, &ssz); |
|
if (ESCAPE_ERROR == esc) |
|
continue; |
|
|
|
rhs = NULL; |
|
|
|
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 (TERMP_UNDER & p->flags) { |
|
term_chara(p, '_'); |
|
term_chara(p, 8); |
|
} |
|
} |
} |
term_chara(p, c); |
|
|
return sz; |
|
} |
|
|
|
int |
|
term_vspan(const struct termp *p, const struct roffsu *su) |
|
{ |
|
double r; |
|
int ri; |
|
|
|
switch (su->unit) { |
|
case SCALE_BU: |
|
r = su->scale / 40.0; |
|
break; |
|
case SCALE_CM: |
|
r = su->scale * 6.0 / 2.54; |
|
break; |
|
case SCALE_FS: |
|
r = su->scale * 65536.0 / 40.0; |
|
break; |
|
case SCALE_IN: |
|
r = su->scale * 6.0; |
|
break; |
|
case SCALE_MM: |
|
r = su->scale * 0.006; |
|
break; |
|
case SCALE_PC: |
|
r = su->scale; |
|
break; |
|
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; |
|
break; |
|
default: |
|
abort(); |
|
} |
|
ri = r > 0.0 ? r + 0.4995 : r - 0.4995; |
|
return ri < 66 ? ri : 1; |
|
} |
|
|
|
/* |
|
* 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); |
} |
} |