=================================================================== RCS file: /cvs/mandoc/term.c,v retrieving revision 1.137 retrieving revision 1.262 diff -u -p -r1.137 -r1.262 --- mandoc/term.c 2010/05/17 22:11:42 1.137 +++ mandoc/term.c 2017/06/02 19:21:23 1.262 @@ -1,22 +1,21 @@ -/* $Id: term.c,v 1.137 2010/05/17 22:11:42 kristaps Exp $ */ +/* $Id: term.c,v 1.262 2017/06/02 19:21:23 schwarze Exp $ */ /* - * Copyright (c) 2008, 2009 Kristaps Dzonsons + * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons + * Copyright (c) 2010-2017 Ingo Schwarze * * Permission to use, copy, modify, and distribute this software for any * purpose with or without fee is hereby granted, provided that the above * 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 - * 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 * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE. */ -#ifdef HAVE_CONFIG_H #include "config.h" -#endif #include @@ -25,150 +24,130 @@ #include #include #include -#include #include "mandoc.h" -#include "chars.h" +#include "mandoc_aux.h" #include "out.h" #include "term.h" -#include "man.h" -#include "mdoc.h" #include "main.h" -static struct termp *term_alloc(enum termenc, size_t); -static void term_free(struct termp *); -static void spec(struct termp *, const char *, size_t); -static void res(struct termp *, const char *, size_t); -static void buffera(struct termp *, const char *, size_t); -static void bufferc(struct termp *, char); -static void adjbuf(struct termp *p, size_t); -static void encode(struct termp *, const char *, size_t); +static size_t cond_width(const struct termp *, int, int *); +static void adjbuf(struct termp *p, size_t); +static void bufferc(struct termp *, char); +static void encode(struct termp *, const char *, size_t); +static void encode1(struct termp *, int); -void * -ascii_alloc(size_t width) +void +term_free(struct termp *p) { - return(term_alloc(TERMENC_ASCII, width)); + free(p->buf); + free(p->fontq); + free(p); } - void -terminal_free(void *arg) +term_begin(struct termp *p, term_margin head, + term_margin foot, const struct roff_meta *arg) { - term_free((struct termp *)arg); + p->headf = head; + p->footf = foot; + p->argf = arg; + (*p->begin)(p); } - -static void -term_free(struct termp *p) +void +term_end(struct termp *p) { - if (p->buf) - free(p->buf); - if (p->symtab) - chars_free(p->symtab); - - free(p); + (*p->end)(p); } - -static struct termp * -term_alloc(enum termenc enc, size_t width) -{ - struct termp *p; - - p = calloc(1, sizeof(struct termp)); - if (NULL == p) { - perror(NULL); - exit(EXIT_FAILURE); - } - p->enc = enc; - /* Enforce some lower boundary. */ - if (width < 60) - width = 60; - p->defrmargin = width - 2; - return(p); -} - - /* - * Flush a line of text. A "line" is loosely defined as being something - * that should be followed by a newline, regardless of whether it's - * broken apart by newlines getting there. A line can also be a - * fragment of a columnar list (`Bl -tag' or `Bl -column'), which does - * not have a trailing newline. - * + * 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 + * at the same offset as the chunk itself. By default, also break the + * output line at the end of the chunk. * The following flags may be specified: * - * - TERMP_NOLPAD: when beginning to write the line, don't left-pad the - * offset value. This is useful when doing columnar lists where the - * prior column has right-padded. - * - * - TERMP_NOBREAK: this is the most important and is used when making - * columns. In short: don't print a newline and instead pad to the - * right margin. Used in conjunction with TERMP_NOLPAD. - * - * - TERMP_TWOSPACE: when padding, make sure there are at least two - * space characters of padding. Otherwise, rather break the line. - * - * - TERMP_DANGLE: don't newline when TERMP_NOBREAK is specified and - * 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. + * - 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_BRTRSP: Consider trailing whitespace significant + * when deciding whether the chunk fits or not. + * - 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 term_flushln(struct termp *p) { - int i; /* current input position in p->buf */ + size_t i; /* current input position in p->buf */ + int ntab; /* number of tabs to prepend */ 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 */ - int j; /* temporary loop index */ - size_t maxvis, mmax; + 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. * This is usually the difference between the right-margin and * 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) { + dv = p->maxrmargin > p->offset ? + p->maxrmargin - p->offset : 0; + bp = (int)dv > p->overstep ? + dv - (size_t)p->overstep : 0; + } else + bp = maxvis; - maxvis = (int)(p->rmargin - p->offset) - p->overstep < 0 ? - /* LINTED */ - 0 : p->rmargin - p->offset - p->overstep; - mmax = (int)(p->maxrmargin - p->offset) - p->overstep < 0 ? - /* LINTED */ - 0 : p->maxrmargin - p->offset - p->overstep; - - bp = TERMP_NOBREAK & p->flags ? mmax : maxvis; - /* - * Indent the first line of a paragraph. + * Calculate the required amount of padding. */ - vbl = p->flags & TERMP_NOLPAD ? 0 : p->offset; + vbl = p->offset + p->overstep > p->viscol ? + p->offset + p->overstep - p->viscol : 0; - /* - * FIXME: if bp is zero, we still output the first word before - * breaking the line. - */ + vis = vend = 0; + i = 0; - vis = vend = i = 0; - while (i < (int)p->col) { + 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 && p->buf[i] == '\t') { + vend = term_tab_next(vis); + vbl += vend - vis; + vis = vend; + ntab++; + i++; + } /* * Count up visible word characters. Control sequences @@ -177,51 +156,88 @@ term_flushln(struct termp *p) * space is printed according to regular spacing rules). */ - /* LINTED */ - for (j = i; j < (int)p->col; j++) { - if (j && ' ' == p->buf[j]) + for (j = i, jhy = 0; j < p->col; j++) { + if (' ' == p->buf[j] || '\t' == p->buf[j]) break; - if (8 == p->buf[j]) - vend--; - else - vend++; + + /* Back over the last printed character. */ + if (8 == p->buf[j]) { + 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]); } /* * Find out whether we would exceed the right margin. * If so, break to the next line. */ - if (vend > bp && vis > 0) { + if (vend > bp && 0 == jhy && vis > 0) { vend -= vis; - putchar('\n'); - if (TERMP_NOBREAK & p->flags) { - for (j = 0; j < (int)p->rmargin; j++) - putchar(' '); + (*p->endline)(p); + p->viscol = 0; + + /* Use pending tabs on the new line. */ + + vbl = 0; + while (ntab--) + vbl = term_tab_next(vbl); + + /* Re-establish indentation. */ + + if (p->flags & TERMP_BRIND) { + vbl += p->rmargin; vend += p->rmargin - p->offset; - } else { - vbl = p->offset; - } + } else + vbl += p->offset; - /* Remove the p->overstep width. */ + /* + * Remove the p->overstep width. + * Again, if p->overstep is negative, + * sign extension does the right thing. + */ - bp += (int)/* LINTED */ - p->overstep; + bp += (size_t)p->overstep; p->overstep = 0; } /* Write out the [remaining] word. */ - for ( ; i < (int)p->col; i++) { + for ( ; i < p->col; i++) { + if (vend > bp && jhy > 0 && i > jhy) + break; + if ('\t' == p->buf[i]) + break; if (' ' == p->buf[i]) { - while (' ' == p->buf[i]) { - vbl++; + 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++; + vbl += (*p->width)(p, ' '); continue; } + if (ASCII_BREAK == p->buf[i]) + continue; /* * Now we definitely know there will be @@ -229,63 +245,69 @@ term_flushln(struct termp *p) * so write preceding white space now. */ if (vbl) { - for (j = 0; j < (int)vbl; j++) - putchar(' '); + (*p->advance)(p, vbl); + p->viscol += vbl; vbl = 0; } - putchar(p->buf[i]); + + (*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]); } - vend += vbl; vis = vend; } + /* + * If there was trailing white space, it was not printed; + * so reset the cursor position accordingly. + */ + if (vis > vbl) + vis -= vbl; + else + vis = 0; + p->col = 0; p->overstep = 0; + p->flags &= ~(TERMP_BACKAFTER | TERMP_BACKBEFORE); if ( ! (TERMP_NOBREAK & p->flags)) { - putchar('\n'); + p->viscol = 0; + (*p->endline)(p); return; } if (TERMP_HANG & p->flags) { - /* We need one blank after the tag. */ - p->overstep = /* LINTED */ - vis - maxvis + 1; + p->overstep += (int)(p->offset + vis - p->rmargin + + p->trailspace * (*p->width)(p, ' ')); /* - * Behave exactly the same way as groff: * If we have overstepped the margin, temporarily move * it to the right and flag the rest of the line to be * 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 there is a request to keep the columns together, + * allow negative overstep when the column is not full. */ - if (p->overstep >= -1) { - assert((int)maxvis + p->overstep >= 0); - /* LINTED */ - maxvis += p->overstep; - } else + if (p->trailspace && p->overstep < 0) p->overstep = 0; + return; } else if (TERMP_DANGLE & p->flags) return; - /* Right-pad. */ - if (maxvis > vis + /* LINTED */ - ((TERMP_TWOSPACE & p->flags) ? 1 : 0)) - for ( ; vis < maxvis; vis++) - putchar(' '); - else { /* ...or newline break. */ - putchar('\n'); - for (i = 0; i < (int)p->rmargin; i++) - putchar(' '); + /* 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 * space. All data in the output buffer is flushed prior to the newline * assertion. @@ -295,15 +317,10 @@ term_newln(struct termp *p) { p->flags |= TERMP_NOSPACE; - if (0 == p->col) { - p->flags &= ~TERMP_NOLPAD; - return; - } - term_flushln(p); - p->flags &= ~TERMP_NOLPAD; + if (p->col || p->viscol) + term_flushln(p); } - /* * 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. @@ -315,34 +332,14 @@ term_vspace(struct termp *p) { term_newln(p); - putchar('\n'); + p->viscol = 0; + if (0 < p->skipvsp) + p->skipvsp--; + else + (*p->endline)(p); } - -static void -spec(struct termp *p, const char *word, size_t len) -{ - const char *rhs; - size_t sz; - - rhs = chars_a2ascii(p->symtab, word, len, &sz); - if (rhs) - encode(p, rhs, sz); -} - - -static void -res(struct termp *p, const char *word, size_t len) -{ - const char *rhs; - size_t sz; - - rhs = chars_a2res(p->symtab, word, len, &sz); - if (rhs) - encode(p, rhs, sz); -} - - +/* Swap current and previous font; for \fP and .ft P */ void term_fontlast(struct termp *p) { @@ -353,7 +350,7 @@ term_fontlast(struct termp *p) 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) { @@ -362,43 +359,31 @@ term_fontrepl(struct termp *p, enum termfont f) p->fontq[p->fonti] = f; } - +/* Set font, save previous. */ void term_fontpush(struct termp *p, enum termfont f) { - assert(p->fonti + 1 < 10); p->fontl = p->fontq[p->fonti]; - p->fontq[++p->fonti] = f; + if (++p->fonti == p->fontsz) { + p->fontsz += 8; + p->fontq = mandoc_reallocarray(p->fontq, + p->fontsz, sizeof(*p->fontq)); + } + p->fontq[p->fonti] = f; } - -const void * -term_fontq(struct termp *p) -{ - - return(&p->fontq[p->fonti]); -} - - -enum termfont -term_fonttop(struct termp *p) -{ - - return(p->fontq[p->fonti]); -} - - +/* Flush to make the saved pointer current again. */ void -term_fontpopq(struct termp *p, const void *key) +term_fontpopq(struct termp *p, int i) { - while (p->fonti >= 0 && key != &p->fontq[p->fonti]) - p->fonti--; - assert(p->fonti >= 0); + assert(i >= 0); + if (p->fonti > i) + p->fonti = i; } - +/* Pop one font off the stack. */ void term_fontpop(struct termp *p) { @@ -407,7 +392,6 @@ term_fontpop(struct termp *p) p->fonti--; } - /* * Handle pwords, partial words, which may be either a single word or a * phrase that cannot be broken down (such as a literal string). This @@ -416,105 +400,218 @@ term_fontpop(struct termp *p) void term_word(struct termp *p, const char *word) { - const char *sv, *seq; - int sz; - size_t ssz; - enum roffdeco deco; + struct roffsu su; + const char nbrsp[2] = { ASCII_NBRSP, 0 }; + const char *seq, *cp; + int sz, uc; + size_t csz, lsz, ssz; + enum mandoc_esc esc; - sv = word; - - if (word[0] && '\0' == word[1]) - switch (word[0]) { - 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)) { - bufferc(p, ' '); - if (TERMP_SENTENCE & p->flags) + if ( ! (TERMP_KEEP & p->flags)) { bufferc(p, ' '); + if (TERMP_SENTENCE & p->flags) + bufferc(p, ' '); + } else + bufferc(p, ASCII_NBRSP); } + 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; + p->flags &= ~(TERMP_SENTENCE | TERMP_NONEWLINE); + p->skipvsp = 0; - /* FIXME: use strcspn. */ - - while (*word) { + while ('\0' != *word) { if ('\\' != *word) { - encode(p, word, 1); - word++; + if (TERMP_NBRWORD & p->flags) { + if (' ' == *word) { + encode(p, nbrsp, 1); + word++; + continue; + } + ssz = strcspn(word, "\\ "); + } else + ssz = strcspn(word, "\\"); + encode(p, word, ssz); + word += (int)ssz; continue; } - seq = ++word; - sz = a2roffdeco(&deco, &seq, &ssz); + word++; + esc = mandoc_escape(&word, &seq, &sz); + if (ESCAPE_ERROR == esc) + continue; - switch (deco) { - case (DECO_RESERVED): - res(p, seq, ssz); + switch (esc) { + case ESCAPE_UNICODE: + uc = mchars_num2uc(seq + 1, sz - 1); break; - case (DECO_SPECIAL): - spec(p, seq, ssz); + case ESCAPE_NUMBERED: + uc = mchars_num2char(seq, sz); + if (uc < 0) + continue; break; - case (DECO_BOLD): + 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); - break; - case (DECO_ITALIC): + continue; + case ESCAPE_FONTITALIC: term_fontrepl(p, TERMFONT_UNDER); - break; - case (DECO_ROMAN): + continue; + case ESCAPE_FONTBI: + term_fontrepl(p, TERMFONT_BI); + continue; + case ESCAPE_FONT: + case ESCAPE_FONTROMAN: term_fontrepl(p, TERMFONT_NONE); - break; - case (DECO_PREVIOUS): + continue; + case ESCAPE_FONTPREV: term_fontlast(p); - break; + 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) == 0) + 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->offset > (size_t)(-uc)) { + p->ti += uc; + p->offset += uc; + } else { + p->ti -= p->offset; + p->offset = 0; + } + } + continue; + case ESCAPE_HLINE: + if (a2roffsu(seq, &su, SCALE_EM) == 0) + continue; + uc = term_hspan(p, &su) / 24; + if (uc <= 0) { + if (p->rmargin <= p->offset) + continue; + lsz = p->rmargin - p->offset; + } else + lsz = uc; + while (sz && + strchr(" %&()*+-./0123456789:<=>", *seq)) { + seq++; + sz--; + } + if (sz && strchr("cifMmnPpuv", *seq)) { + seq++; + sz--; + } + if (sz == 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->col > 2 && + (p->buf[p->col - 1] == ' ' || + p->buf[p->col - 1] == '\t')) + p->col -= 2; + continue; default: - break; + continue; } - word += sz; - if (DECO_NOSPACE == deco && '\0' == *word) - p->flags |= TERMP_NOSPACE; - } + /* + * Common handling for Unicode and numbered + * character escape sequences. + */ - /* - * Note that we don't process the pipe: the parser sees it as - * punctuation, but we don't in terms of typography. - */ - if (sv[0] && 0 == sv[1]) - switch (sv[0]) { - case('('): - /* FALLTHROUGH */ - case('['): - p->flags |= TERMP_NOSPACE; - break; - default: - break; + if (p->enc == TERMENC_ASCII) { + cp = ascii_uc2str(uc); + encode(p, cp, strlen(cp)); + } else { + if ((uc < 0x20 && uc != 0x09) || + (uc > 0x7E && uc < 0xA0)) + uc = 0xFFFD; + encode1(p, uc); } + } + p->flags &= ~TERMP_NBRWORD; } - static void adjbuf(struct termp *p, size_t sz) { @@ -524,143 +621,304 @@ adjbuf(struct termp *p, size_t sz) while (sz >= p->maxcols) p->maxcols <<= 2; - p->buf = realloc(p->buf, p->maxcols); - if (NULL == p->buf) { - perror(NULL); - exit(EXIT_FAILURE); - } + p->buf = mandoc_reallocarray(p->buf, p->maxcols, sizeof(int)); } - static void -buffera(struct termp *p, const char *word, size_t sz) +bufferc(struct termp *p, char c) { - if (p->col + sz >= p->maxcols) - adjbuf(p, p->col + sz); + if (p->col + 1 >= p->maxcols) + adjbuf(p, p->col + 1); - memcpy(&p->buf[(int)p->col], word, sz); - p->col += sz; + p->buf[p->col++] = c; } - +/* + * See encode(). + * Do this for a single (probably unicode) value. + * Does not check for non-decorated glyphs. + */ static void -bufferc(struct termp *p, char c) +encode1(struct termp *p, int c) { + enum termfont f; - if (p->col + 1 >= p->maxcols) - adjbuf(p, p->col + 1); + if (p->col + 7 >= p->maxcols) + adjbuf(p, p->col + 7); - p->buf[(int)p->col++] = c; + f = (c == ASCII_HYPH || c > 127 || isgraph(c)) ? + p->fontq[p->fonti] : TERMFONT_NONE; + + if (p->flags & TERMP_BACKBEFORE) { + 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 encode(struct termp *p, const char *word, size_t sz) { - enum termfont f; - int i; + size_t i; - /* - * Encode and buffer a string of characters. If the current - * font mode is unset, buffer directly, else encode then buffer - * character by character. - */ + if (p->col + 2 + (sz * 5) >= p->maxcols) + adjbuf(p, p->col + 2 + (sz * 5)); - if (TERMFONT_NONE == (f = term_fonttop(p))) { - buffera(p, word, sz); - return; - } + for (i = 0; i < sz; i++) { + if (ASCII_HYPH == word[i] || + isgraph((unsigned char)word[i])) + encode1(p, word[i]); + else { + p->buf[p->col++] = word[i]; - for (i = 0; i < (int)sz; i++) { - if ( ! isgraph((u_char)word[i])) { - bufferc(p, word[i]); - continue; + /* + * 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 (TERMFONT_UNDER == f) - bufferc(p, '_'); - else - bufferc(p, word[i]); +void +term_setwidth(struct termp *p, const char *wstr) +{ + struct roffsu su; + int iop, width; - bufferc(p, 8); - bufferc(p, word[i]); + iop = 0; + width = 0; + if (NULL != wstr) { + switch (*wstr) { + case '+': + iop = 1; + wstr++; + break; + case '-': + iop = -1; + wstr++; + break; + default: + break; + } + if (a2roffsu(wstr, &su, SCALE_MAX)) + width = term_hspan(p, &su); + else + iop = 0; } + (*p->setwidth)(p, iop, width); } +size_t +term_len(const struct termp *p, size_t sz) +{ + return (*p->width)(p, ' ') * sz; +} + +static size_t +cond_width(const struct termp *p, int c, int *skip) +{ + + if (*skip) { + (*skip) = 0; + return 0; + } else + return (*p->width)(p, c); +} + 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): - r = su->scale * 2; - break; - case (SCALE_IN): - r = su->scale * 6; - break; - case (SCALE_PC): - r = su->scale; - break; - case (SCALE_PT): - r = su->scale / 8; - break; - case (SCALE_MM): - r = su->scale / 1000; - break; - case (SCALE_VS): - r = su->scale; - break; - default: - r = su->scale - 1; - break; + /* + * 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 (r < 0.0) - r = 0.0; - return(/* LINTED */(size_t) - r); + return sz; } - -size_t -term_hspan(const struct roffsu *su) +int +term_vspan(const struct termp *p, const struct roffsu *su) { double r; + int ri; - /* XXX: CM, IN, and PT are approximations. */ - switch (su->unit) { - case (SCALE_CM): - r = 4 * su->scale; + case SCALE_BU: + r = su->scale / 40.0; break; - case (SCALE_IN): - /* XXX: this is an approximation. */ - r = 10 * su->scale; + case SCALE_CM: + r = su->scale * 6.0 / 2.54; break; - case (SCALE_PC): - r = (10 * su->scale) / 6; + case SCALE_FS: + r = su->scale * 65536.0 / 40.0; break; - case (SCALE_PT): - r = (10 * su->scale) / 72; + case SCALE_IN: + r = su->scale * 6.0; break; - case (SCALE_MM): - r = su->scale / 1000; /* FIXME: double-check. */ + case SCALE_MM: + r = su->scale * 0.006; break; - case (SCALE_VS): - r = su->scale * 2 - 1; /* FIXME: double-check. */ + case SCALE_PC: + r = su->scale; 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; break; + default: + abort(); } - - if (r < 0.0) - r = 0.0; - return((size_t)/* LINTED */ - r); + 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); +}