=================================================================== RCS file: /cvs/mandoc/term.c,v retrieving revision 1.191 retrieving revision 1.290 diff -u -p -r1.191 -r1.290 --- mandoc/term.c 2011/05/15 22:29:50 1.191 +++ mandoc/term.c 2022/08/16 17:45:55 1.290 @@ -1,23 +1,21 @@ -/* $Id: term.c,v 1.191 2011/05/15 22:29:50 kristaps Exp $ */ +/* $Id: term.c,v 1.290 2022/08/16 17:45:55 schwarze Exp $ */ /* - * Copyright (c) 2008, 2009, 2010 Kristaps Dzonsons - * Copyright (c) 2010, 2011 Ingo Schwarze + * Copyright (c) 2010-2022 Ingo Schwarze + * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons * * 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 @@ -29,30 +27,48 @@ #include #include "mandoc.h" +#include "mandoc_aux.h" #include "out.h" #include "term.h" #include "main.h" -static void adjbuf(struct termp *p, int); +static size_t cond_width(const struct termp *, int, int *); +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 void term_field(struct termp *, size_t, size_t); +static void term_fill(struct termp *, size_t *, size_t *, + size_t); + void -term_free(struct termp *p) +term_setcol(struct termp *p, size_t maxtcol) { + if (maxtcol > p->maxtcol) { + p->tcols = mandoc_recallocarray(p->tcols, + p->maxtcol, maxtcol, sizeof(*p->tcols)); + p->maxtcol = maxtcol; + } + p->lasttcol = maxtcol - 1; + p->tcol = p->tcols; +} - if (p->buf) - free(p->buf); - if (p->symtab) - mchars_free(p->symtab); - +void +term_free(struct termp *p) +{ + term_tab_free(); + for (p->tcol = p->tcols; p->tcol < p->tcols + p->maxtcol; p->tcol++) + free(p->tcol->buf); + free(p->tcols); + free(p->fontq); free(p); } - void -term_begin(struct termp *p, term_margin head, - term_margin foot, const void *arg) +term_begin(struct termp *p, term_margin head, + term_margin foot, const struct roff_meta *arg) { p->headf = head; @@ -61,7 +77,6 @@ term_begin(struct termp *p, term_margin head, (*p->begin)(p); } - void term_end(struct termp *p) { @@ -69,245 +84,375 @@ term_end(struct termp *p) (*p->end)(p); } - -struct termp * -term_alloc(enum termenc enc) -{ - struct termp *p; - - p = mandoc_calloc(1, sizeof(struct termp)); - p->enc = enc; - 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. - * - * 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. + * 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. There are many flags modifying + * this behaviour, see the comments in the body of the function. */ void term_flushln(struct termp *p) { - int i; /* current input position in p->buf */ - 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 */ - int j; /* temporary loop index for p->buf */ - int jhy; /* last hyph before overflow w/r/t j */ - size_t maxvis; /* output position of visible boundary */ - size_t mmax; /* used in calculating bp */ + size_t vbl; /* Number of blanks to prepend to the output. */ + size_t vbr; /* Actual visual position of the end of field. */ + size_t vfield; /* Desired visual field width. */ + size_t vtarget; /* Desired visual position of the right margin. */ + size_t ic; /* Character position in the input buffer. */ + size_t nbr; /* Number of characters to print in this field. */ /* - * 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. + * Normally, start writing at the left margin, but with the + * NOPAD flag, start writing at the current position instead. */ - assert (p->rmargin >= p->offset); - dv = p->rmargin - p->offset; - maxvis = (int)dv > p->overstep ? dv - (size_t)p->overstep : 0; - dv = p->maxrmargin - p->offset; - mmax = (int)dv > p->overstep ? dv - (size_t)p->overstep : 0; - bp = TERMP_NOBREAK & p->flags ? mmax : maxvis; + vbl = (p->flags & TERMP_NOPAD) || p->tcol->offset < p->viscol ? + 0 : p->tcol->offset - p->viscol; + if (p->minbl && vbl < p->minbl) + vbl = p->minbl; - /* - * Indent the first line of a paragraph. - */ - vbl = p->flags & TERMP_NOLPAD ? (size_t)0 : p->offset; + if ((p->flags & TERMP_MULTICOL) == 0) + p->tcol->col = 0; - vis = vend = 0; - i = 0; + /* Loop over output lines. */ - while (i < p->col) { + for (;;) { + vfield = p->tcol->rmargin > p->viscol + vbl ? + p->tcol->rmargin - p->viscol - vbl : 0; + /* - * Handle literal tab characters: collapse all - * subsequent tabs into a single huge set of spaces. + * Normally, break the line at the the right margin + * of the field, but with the NOBREAK flag, only + * break it at the max right margin of the screen, + * and with the BRNEVER flag, never break it at all. */ - while (i < p->col && '\t' == p->buf[i]) { - vend = (vis / p->tabwidth + 1) * p->tabwidth; - vbl += vend - vis; - vis = vend; - i++; - } + vtarget = (p->flags & TERMP_NOBREAK) == 0 ? vfield : + p->maxrmargin > p->viscol + vbl ? + p->maxrmargin - p->viscol - vbl : 0; + /* - * Count up visible word characters. Control sequences - * (starting with the CSI) aren't counted. A space - * generates a non-printing word, which is valid (the - * space is printed according to regular spacing rules). + * Figure out how much text will fit in the field. + * If there is whitespace only, print nothing. */ - for (j = i, jhy = 0; j < p->col; j++) { - if ((j && ' ' == p->buf[j]) || '\t' == p->buf[j]) - break; + term_fill(p, &nbr, &vbr, + p->flags & TERMP_BRNEVER ? SIZE_MAX : vtarget); + if (nbr == 0) + break; - /* Back over the the last printed character. */ - if (8 == p->buf[j]) { - assert(j); - vend -= (*p->width)(p, p->buf[j - 1]); + /* + * With the CENTER or RIGHT flag, increase the indentation + * to center the text between the left and right margins + * or to adjust it to the right margin, respectively. + */ + + if (vbr < vtarget) { + if (p->flags & TERMP_CENTER) + vbl += (vtarget - vbr) / 2; + else if (p->flags & TERMP_RIGHT) + vbl += vtarget - vbr; + } + + /* Finally, print the field content. */ + + term_field(p, vbl, nbr); + if (vbr < vtarget) + p->tcol->taboff += vbr; + else + p->tcol->taboff += vtarget; + p->tcol->taboff += (*p->width)(p, ' '); + + /* + * If there is no text left in the field, exit the loop. + * If the BRTRSP flag is set, consider trailing + * whitespace significant when deciding whether + * the field fits or not. + */ + + for (ic = p->tcol->col; ic < p->tcol->lastcol; ic++) { + switch (p->tcol->buf[ic]) { + case '\t': + if (p->flags & TERMP_BRTRSP) + vbr = term_tab_next(vbr); continue; + case ' ': + if (p->flags & TERMP_BRTRSP) + vbr += (*p->width)(p, ' '); + continue; + case '\n': + case ASCII_NBRZW: + case ASCII_BREAK: + case ASCII_TABREF: + continue; + default: + break; } + break; + } + if (ic == p->tcol->lastcol) + break; - /* Regular word. */ - /* Break at the hyphen point if we overrun. */ - if (vend > vis && vend < bp && - ASCII_HYPH == p->buf[j]) - jhy = j; + /* + * At the location of an automtic line break, input + * space characters are consumed by the line break. + */ - vend += (*p->width)(p, p->buf[j]); - } + while (p->tcol->col < p->tcol->lastcol && + p->tcol->buf[p->tcol->col] == ' ') + p->tcol->col++; /* - * Find out whether we would exceed the right margin. - * If so, break to the next line. + * In multi-column mode, leave the rest of the text + * in the buffer to be handled by a subsequent + * invocation, such that the other columns of the + * table can be handled first. + * In single-column mode, simply break the line. */ - if (vend > bp && 0 == jhy && vis > 0) { - vend -= vis; - (*p->endline)(p); - if (TERMP_NOBREAK & p->flags) { - p->viscol = p->rmargin; - (*p->advance)(p, p->rmargin); - vend += p->rmargin - p->offset; - } else { - p->viscol = 0; - vbl = p->offset; - } - /* Remove the p->overstep width. */ + if (p->flags & TERMP_MULTICOL) + return; - bp += (size_t)p->overstep; - p->overstep = 0; - } + endline(p); - /* Write out the [remaining] word. */ - for ( ; i < p->col; i++) { - if (vend > bp && jhy > 0 && i > jhy) + /* + * Normally, start the next line at the same indentation + * as this one, but with the BRIND flag, start it at the + * right margin instead. This is used together with + * NOBREAK for the tags in various kinds of tagged lists. + */ + + vbl = p->flags & TERMP_BRIND ? + p->tcol->rmargin : p->tcol->offset; + } + + /* Reset output state in preparation for the next field. */ + + 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; + + /* + * The HANG flag means that the next field + * always follows on the same line. + * The NOBREAK flag means that the next field + * follows on the same line unless the field was overrun. + * Normally, break the line at the end of each field. + */ + + if ((p->flags & TERMP_HANG) == 0 && + ((p->flags & TERMP_NOBREAK) == 0 || + vbr + term_len(p, p->trailspace) > vfield)) + endline(p); +} + +/* + * Store the number of input characters to print in this field in *nbr + * and their total visual width to print in *vbr. + * If there is only whitespace in the field, both remain zero. + * The desired visual width of the field is provided by vtarget. + * If the first word is longer, the field will be overrun. + */ +static void +term_fill(struct termp *p, size_t *nbr, size_t *vbr, size_t vtarget) +{ + size_t ic; /* Character position in the input buffer. */ + size_t vis; /* Visual position of the current character. */ + size_t vn; /* Visual position of the next character. */ + int breakline; /* Break at the end of this word. */ + int graph; /* Last character was non-blank. */ + int taboff; /* Temporary offset for literal tabs. */ + + *nbr = *vbr = vis = 0; + breakline = graph = 0; + taboff = p->tcol->taboff; + for (ic = p->tcol->col; ic < p->tcol->lastcol; ic++) { + switch (p->tcol->buf[ic]) { + case '\b': /* Escape \o (overstrike) or backspace markup. */ + assert(ic > 0); + vis -= (*p->width)(p, p->tcol->buf[ic - 1]); + continue; + + case ' ': + case ASCII_BREAK: /* Escape \: (breakpoint). */ + vn = vis; + if (p->tcol->buf[ic] == ' ') + vn += (*p->width)(p, ' '); + /* Can break at the end of a word. */ + if (breakline || vn > vtarget) break; - if ('\t' == p->buf[i]) - break; - if (' ' == p->buf[i]) { - j = i; - while (' ' == p->buf[i]) - i++; - dv = (size_t)(i - j) * (*p->width)(p, ' '); - vbl += dv; - vend += dv; - break; + if (graph) { + *nbr = ic; + *vbr = vis; + graph = 0; } - if (ASCII_NBRSP == p->buf[i]) { - vbl += (*p->width)(p, ' '); - continue; - } + vis = vn; + continue; + case '\n': /* Escape \p (break at the end of the word). */ + breakline = 1; + continue; + + case ASCII_HYPH: /* Breakable hyphen. */ + graph = 1; /* - * Now we definitely know there will be - * printable characters to output, - * so write preceding white space now. + * We are about to decide whether to break the + * line or not, so we no longer need this hyphen + * to be marked as breakable. Put back a real + * hyphen such that we get the correct width. */ - if (vbl) { - (*p->advance)(p, vbl); - p->viscol += vbl; - vbl = 0; + p->tcol->buf[ic] = '-'; + vis += (*p->width)(p, '-'); + if (vis > vtarget) { + ic++; + break; } + *nbr = ic + 1; + *vbr = vis; + continue; - if (ASCII_HYPH == p->buf[i]) { - (*p->letter)(p, '-'); - p->viscol += (*p->width)(p, '-'); - } else { - (*p->letter)(p, p->buf[i]); - p->viscol += (*p->width)(p, p->buf[i]); + case ASCII_TABREF: + taboff = -vis - (*p->width)(p, ' '); + continue; + + default: + switch (p->tcol->buf[ic]) { + case '\t': + if (taboff < 0 && (size_t)-taboff > vis) + vis = 0; + else + vis += taboff; + vis = term_tab_next(vis); + vis -= taboff; + break; + case ASCII_NBRZW: /* Non-breakable zero-width. */ + break; + case ASCII_NBRSP: /* Non-breakable space. */ + p->tcol->buf[ic] = ' '; + /* FALLTHROUGH */ + default: /* Printable character. */ + vis += (*p->width)(p, p->tcol->buf[ic]); + break; } + graph = 1; + if (vis > vtarget && *nbr > 0) + return; + continue; } - vis = vend; + break; } /* - * If there was trailing white space, it was not printed; - * so reset the cursor position accordingly. + * If the last word extends to the end of the field without any + * trailing whitespace, the loop could not check yet whether it + * can remain on this line. So do the check now. */ - vis -= vbl; - p->col = 0; - p->overstep = 0; - - if ( ! (TERMP_NOBREAK & p->flags)) { - p->viscol = 0; - (*p->endline)(p); - return; + if (graph && (vis <= vtarget || *nbr == 0)) { + *nbr = ic; + *vbr = vis; } +} - if (TERMP_HANG & p->flags) { - /* We need one blank after the tag. */ - p->overstep = (int)(vis - maxvis + (*p->width)(p, ' ')); +/* + * Print the contents of one field + * with an indentation of vbl visual columns, + * and an input string length of nbr characters. + */ +static void +term_field(struct termp *p, size_t vbl, size_t nbr) +{ + size_t ic; /* Character position in the input buffer. */ + size_t vis; /* Visual position of the current character. */ + size_t vt; /* Visual position including tab offset. */ + size_t dv; /* Visual width of the current character. */ + int taboff; /* Temporary offset for literal tabs. */ + vis = 0; + taboff = p->tcol->taboff; + for (ic = p->tcol->col; ic < nbr; ic++) { + /* - * 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. + * To avoid the printing of trailing whitespace, + * do not print whitespace right away, only count it. */ - if (p->overstep >= -1) { - assert((int)maxvis + p->overstep >= 0); - maxvis += (size_t)p->overstep; - } else - p->overstep = 0; - } else if (TERMP_DANGLE & p->flags) - return; + switch (p->tcol->buf[ic]) { + case '\n': + case ASCII_BREAK: + case ASCII_NBRZW: + continue; + case ASCII_TABREF: + taboff = -vis - (*p->width)(p, ' '); + continue; + case '\t': + case ' ': + case ASCII_NBRSP: + if (p->tcol->buf[ic] == '\t') { + if (taboff < 0 && (size_t)-taboff > vis) + vt = 0; + else + vt = vis + taboff; + dv = term_tab_next(vt) - vt; + } else + dv = (*p->width)(p, ' '); + vbl += dv; + vis += dv; + continue; + default: + break; + } - /* Right-pad. */ - if (maxvis > vis + - ((TERMP_TWOSPACE & p->flags) ? (*p->width)(p, ' ') : 0)) { - p->viscol += maxvis - vis; - (*p->advance)(p, maxvis - vis); - vis += (maxvis - vis); - } else { /* ...or newline break. */ - (*p->endline)(p); - p->viscol = p->rmargin; - (*p->advance)(p, p->rmargin); + /* + * We found a non-blank character to print, + * so write preceding white space now. + */ + + if (vbl > 0) { + (*p->advance)(p, vbl); + p->viscol += vbl; + vbl = 0; + } + + /* Print the character and adjust the visual position. */ + + (*p->letter)(p, p->tcol->buf[ic]); + if (p->tcol->buf[ic] == '\b') { + dv = (*p->width)(p, p->tcol->buf[ic - 1]); + p->viscol -= dv; + vis -= dv; + } else { + dv = (*p->width)(p, p->tcol->buf[ic]); + p->viscol += dv; + vis += dv; + } } + p->tcol->col = nbr; } +static void +endline(struct termp *p) +{ + if ((p->flags & (TERMP_NEWMC | TERMP_ENDMC)) == TERMP_ENDMC) { + p->mc = NULL; + p->flags &= ~TERMP_ENDMC; + } + if (p->mc != NULL) { + if (p->viscol && p->maxrmargin >= p->viscol) + (*p->advance)(p, p->maxrmargin - p->viscol + 1); + p->flags |= TERMP_NOBUF | TERMP_NOSPACE; + term_word(p, p->mc); + p->flags &= ~(TERMP_NOBUF | TERMP_NEWMC); + } + p->viscol = 0; + p->minbl = 0; + (*p->endline)(p); +} -/* +/* * A newline only breaks an existing line; it won't assert vertical * space. All data in the output buffer is flushed prior to the newline * assertion. @@ -315,17 +460,12 @@ term_flushln(struct termp *p) void term_newln(struct termp *p) { - p->flags |= TERMP_NOSPACE; - if (0 == p->col && 0 == p->viscol) { - p->flags &= ~TERMP_NOLPAD; - return; - } - term_flushln(p); - p->flags &= ~TERMP_NOLPAD; + if (p->tcol->lastcol || p->viscol) + term_flushln(p); + p->tcol->taboff = 0; } - /* * 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. @@ -338,9 +478,14 @@ term_vspace(struct termp *p) term_newln(p); p->viscol = 0; - (*p->endline)(p); + 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) { @@ -351,7 +496,7 @@ 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) { @@ -360,43 +505,31 @@ 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) { @@ -413,164 +546,438 @@ term_fontpop(struct termp *p) void term_word(struct termp *p, const char *word) { + struct roffsu su; + const char nbrsp[2] = { ASCII_NBRSP, 0 }; const char *seq, *cp; - char c; - int sz; - size_t ssz; + int sz, uc; + size_t csz, lsz, ssz; enum mandoc_esc esc; - if ( ! (TERMP_NOSPACE & p->flags)) { - if ( ! (TERMP_KEEP & p->flags)) { - if (TERMP_PREKEEP & p->flags) - p->flags |= TERMP_KEEP; - bufferc(p, ' '); - if (TERMP_SENTENCE & p->flags) + if ((p->flags & TERMP_NOBUF) == 0) { + if ((p->flags & TERMP_NOSPACE) == 0) { + if ((p->flags & TERMP_KEEP) == 0) { bufferc(p, ' '); - } else - bufferc(p, ASCII_NBRSP); + if (p->flags & TERMP_SENTENCE) + 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 ( ! (p->flags & TERMP_NONOSPACE)) - p->flags &= ~TERMP_NOSPACE; - else - p->flags |= TERMP_NOSPACE; - - p->flags &= ~(TERMP_SENTENCE | TERMP_IGNDELIM); - while ('\0' != *word) { - if ((ssz = strcspn(word, "\\")) > 0) + if ('\\' != *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; - if ('\\' != *word) + word += (int)ssz; continue; + } word++; esc = mandoc_escape(&word, &seq, &sz); - if (ESCAPE_ERROR == esc) - break; - switch (esc) { - case (ESCAPE_NUMBERED): - if ('\0' != (c = mchars_num2char(seq, sz))) - encode(p, &c, 1); + case ESCAPE_UNICODE: + uc = mchars_num2uc(seq + 1, sz - 1); break; - case (ESCAPE_PREDEF): - cp = mchars_res2str(p->symtab, seq, sz, &ssz); - if (NULL != cp) - encode(p, cp, ssz); + case ESCAPE_NUMBERED: + uc = mchars_num2char(seq, sz); + if (uc >= 0) + break; + bufferc(p, ASCII_NBRZW); + continue; + case ESCAPE_SPECIAL: + if (p->enc == TERMENC_ASCII) { + cp = mchars_spec2str(seq, sz, &ssz); + if (cp != NULL) + encode(p, cp, ssz); + else + bufferc(p, ASCII_NBRZW); + } else { + uc = mchars_spec2cp(seq, sz); + if (uc > 0) + encode1(p, uc); + else + bufferc(p, ASCII_NBRZW); + } + continue; + case ESCAPE_UNDEF: + uc = *seq; break; - case (ESCAPE_SPECIAL): - cp = mchars_spec2str(p->symtab, seq, sz, &ssz); - if (NULL != cp) - encode(p, cp, ssz); - else if (1 == ssz) - encode(p, seq, sz); - break; - case (ESCAPE_FONTBOLD): + case ESCAPE_FONTBOLD: + case ESCAPE_FONTCB: term_fontrepl(p, TERMFONT_BOLD); - break; - case (ESCAPE_FONTITALIC): + continue; + case ESCAPE_FONTITALIC: + case ESCAPE_FONTCI: term_fontrepl(p, TERMFONT_UNDER); - break; - case (ESCAPE_FONTROMAN): + continue; + case ESCAPE_FONTBI: + term_fontrepl(p, TERMFONT_BI); + continue; + case ESCAPE_FONT: + case ESCAPE_FONTCR: + case ESCAPE_FONTROMAN: term_fontrepl(p, TERMFONT_NONE); - break; - case (ESCAPE_FONTPREV): + continue; + case ESCAPE_FONTPREV: term_fontlast(p); - break; - case (ESCAPE_NOSPACE): - if ('\0' == *word) - p->flags |= TERMP_NOSPACE; - break; + 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_DEVICE: + if (p->type == TERMTYPE_PDF) + encode(p, "pdf", 3); + else if (p->type == TERMTYPE_PS) + encode(p, "ps", 2); + else if (p->enc == TERMENC_ASCII) + encode(p, "ascii", 5); + else + encode(p, "utf8", 4); + continue; + case ESCAPE_HORIZ: + if (p->flags & TERMP_BACKAFTER) { + p->flags &= ~TERMP_BACKAFTER; + continue; + } + if (*seq == '|') { + seq++; + uc = -p->col; + } else + uc = 0; + if (a2roffsu(seq, &su, SCALE_EM) == NULL) + continue; + uc += term_hen(p, &su); + if (uc >= 0) { + while (uc > 0) { + uc -= term_len(p, 1); + if (p->flags & TERMP_BACKBEFORE) + p->flags &= ~TERMP_BACKBEFORE; + else + bufferc(p, ASCII_NBRSP); + } + continue; + } + if (p->flags & TERMP_BACKBEFORE) { + p->flags &= ~TERMP_BACKBEFORE; + assert(p->col > 0); + p->col--; + } + if (p->col >= (size_t)(-uc)) { + p->col += uc; + } else { + uc += p->col; + p->col = 0; + if (p->tcol->offset > (size_t)(-uc)) { + p->ti += uc; + p->tcol->offset += uc; + } else { + p->ti -= p->tcol->offset; + p->tcol->offset = 0; + } + } + continue; + case ESCAPE_HLINE: + if ((cp = a2roffsu(seq, &su, SCALE_EM)) == NULL) + continue; + uc = term_hen(p, &su); + if (uc <= 0) { + if (p->tcol->rmargin <= p->tcol->offset) + continue; + lsz = p->tcol->rmargin - p->tcol->offset; + } else + lsz = uc; + if (*cp == seq[-1]) + uc = -1; + else if (*cp == '\\') { + seq = cp + 1; + esc = mandoc_escape(&seq, &cp, &sz); + switch (esc) { + case ESCAPE_UNICODE: + uc = mchars_num2uc(cp + 1, sz - 1); + break; + case ESCAPE_NUMBERED: + uc = mchars_num2char(cp, sz); + break; + case ESCAPE_SPECIAL: + uc = mchars_spec2cp(cp, sz); + break; + case ESCAPE_UNDEF: + uc = *seq; + break; + default: + uc = -1; + break; + } + } else + uc = *cp; + if (uc < 0x20 || (uc > 0x7E && uc < 0xA0)) + uc = '_'; + if (p->enc == TERMENC_ASCII) { + cp = ascii_uc2str(uc); + csz = term_strlen(p, cp); + ssz = strlen(cp); + } else + csz = (*p->width)(p, uc); + while (lsz >= csz) { + if (p->enc == TERMENC_ASCII) + encode(p, cp, ssz); + else + encode1(p, uc); + lsz -= csz; + } + continue; + case ESCAPE_SKIPCHAR: + 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; + case ESCAPE_IGNORE: + bufferc(p, ASCII_NBRZW); + continue; default: - break; + continue; } + + /* + * Common handling for Unicode and numbered + * character escape sequences. + */ + + 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, int sz) +adjbuf(struct termp_col *c, size_t sz) { - - if (0 == p->maxcols) - p->maxcols = 1024; - while (sz >= p->maxcols) - p->maxcols <<= 2; - - p->buf = mandoc_realloc - (p->buf, sizeof(int) * (size_t)p->maxcols); + 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)); } static void bufferc(struct termp *p, char c) { + if (p->flags & TERMP_NOBUF) { + (*p->letter)(p, c); + 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; +} - if (p->col + 1 >= p->maxcols) - adjbuf(p, p->col + 1); - - p->buf[p->col++] = c; +void +term_tab_ref(struct termp *p) +{ + if (p->tcol->lastcol && p->tcol->lastcol <= p->col && + (p->flags & TERMP_NOBUF) == 0) + bufferc(p, ASCII_TABREF); } +/* + * See encode(). + * Do this for a single (probably unicode) value. + * Does not check for non-decorated glyphs. + */ static void -encode(struct termp *p, const char *word, size_t sz) +encode1(struct termp *p, int c) { enum termfont f; - int i, len; - /* LINTED */ - len = sz; + if (p->flags & TERMP_NOBUF) { + (*p->letter)(p, c); + return; + } - /* - * 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 + 7 >= p->tcol->maxcols) + adjbuf(p->tcol, p->col + 7); - if (TERMFONT_NONE == (f = term_fonttop(p))) { - if (p->col + len >= p->maxcols) - adjbuf(p, p->col + len); - for (i = 0; i < len; i++) - p->buf[p->col++] = word[i]; + f = (c == ASCII_HYPH || c > 127 || isgraph(c)) ? + p->fontq[p->fonti] : TERMFONT_NONE; + + if (p->flags & TERMP_BACKBEFORE) { + if (p->tcol->buf[p->col - 1] == ' ' || + p->tcol->buf[p->col - 1] == '\t') + p->col--; + else + p->tcol->buf[p->col++] = '\b'; + p->flags &= ~TERMP_BACKBEFORE; + } + 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) +{ + size_t i; + + if (p->flags & TERMP_NOBUF) { + for (i = 0; i < sz; i++) + (*p->letter)(p, word[i]); return; } - /* Pre-buffer, assuming worst-case. */ + if (p->col + 2 + (sz * 5) >= p->tcol->maxcols) + adjbuf(p->tcol, p->col + 2 + (sz * 5)); - if (p->col + 1 + (len * 3) >= p->maxcols) - adjbuf(p, p->col + 1 + (len * 3)); + for (i = 0; i < sz; i++) { + 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++; - for (i = 0; i < len; i++) { - if ( ! isgraph((unsigned char)word[i])) { - p->buf[p->col++] = 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 (p->tcol->lastcol < p->col) + p->tcol->lastcol = p->col; +} - if (TERMFONT_UNDER == f) - p->buf[p->col++] = '_'; - else - p->buf[p->col++] = word[i]; +void +term_setwidth(struct termp *p, const char *wstr) +{ + struct roffsu su; + int iop, width; - p->buf[p->col++] = 8; - p->buf[p->col++] = 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) != 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) { - return((*p->width)(p, ' ') * 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_strlen(const struct termp *p, const char *cp) { size_t sz, rsz, i; - int ssz, c; + int ssz, skip, uc; const char *seq, *rhs; - static const char rej[] = { '\\', ASCII_HYPH, ASCII_NBRSP, '\0' }; + enum mandoc_esc esc; + static const char rej[] = { '\\', ASCII_NBRSP, ASCII_NBRZW, + ASCII_BREAK, ASCII_HYPH, ASCII_TABREF, '\0' }; /* * Account for escaped sequences within string length @@ -579,53 +986,113 @@ term_strlen(const struct termp *p, const char *cp) */ sz = 0; + skip = 0; while ('\0' != *cp) { rsz = strcspn(cp, rej); for (i = 0; i < rsz; i++) - sz += (*p->width)(p, *cp++); + sz += cond_width(p, *cp++, &skip); switch (*cp) { - case ('\\'): + case '\\': cp++; rhs = NULL; - switch (mandoc_escape(&cp, &seq, &ssz)) { - case (ESCAPE_ERROR): - return(sz); - case (ESCAPE_NUMBERED): - c = mchars_num2char(seq, ssz); - if ('\0' != c) - sz += (*p->width)(p, c); + esc = mandoc_escape(&cp, &seq, &ssz); + switch (esc) { + case ESCAPE_UNICODE: + uc = mchars_num2uc(seq + 1, ssz - 1); break; - case (ESCAPE_PREDEF): - rhs = mchars_res2str - (p->symtab, seq, ssz, &rsz); + case ESCAPE_NUMBERED: + uc = mchars_num2char(seq, ssz); + if (uc < 0) + continue; break; - case (ESCAPE_SPECIAL): - rhs = mchars_spec2str - (p->symtab, seq, ssz, &rsz); - - if (ssz != 1 || rhs) - break; - - rhs = seq; - rsz = ssz; + 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_UNDEF: + uc = *seq; break; - default: + case ESCAPE_DEVICE: + if (p->type == TERMTYPE_PDF) { + rhs = "pdf"; + rsz = 3; + } else if (p->type == TERMTYPE_PS) { + rhs = "ps"; + rsz = 2; + } else if (p->enc == TERMENC_ASCII) { + rhs = "ascii"; + rsz = 5; + } else { + rhs = "utf8"; + rsz = 4; + } break; + case ESCAPE_SKIPCHAR: + 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; } - if (NULL == rhs) + /* + * 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 += (*p->width)(p, ' '); + case ASCII_NBRSP: + sz += cond_width(p, ' ', &skip); cp++; break; - case (ASCII_HYPH): - sz += (*p->width)(p, '-'); + case ASCII_HYPH: + sz += cond_width(p, '-', &skip); cp++; break; default: @@ -633,53 +1100,71 @@ term_strlen(const struct termp *p, const char *cp) } } - return(sz); + return sz; } -/* ARGSUSED */ -size_t +int term_vspan(const struct termp *p, const struct roffsu *su) { double r; + int ri; switch (su->unit) { - case (SCALE_CM): - r = su->scale * 2; + case SCALE_BU: + r = su->scale / 40.0; break; - case (SCALE_IN): - r = su->scale * 6; + case SCALE_CM: + r = su->scale * 6.0 / 2.54; break; - case (SCALE_PC): + 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 / 8; + case SCALE_PT: + r = su->scale / 12.0; break; - case (SCALE_MM): - r = su->scale / 1000; + case SCALE_EN: + case SCALE_EM: + r = su->scale * 0.6; break; - case (SCALE_VS): + case SCALE_VS: r = su->scale; break; default: - r = su->scale - 1; - break; + abort(); } - - if (r < 0.0) - r = 0.0; - return(/* LINTED */(size_t) - r); + ri = r > 0.0 ? r + 0.4995 : r - 0.4995; + return ri < 66 ? ri : 1; } -size_t +/* + * Convert a scaling width to basic units, rounding towards 0. + */ +int term_hspan(const struct termp *p, const struct roffsu *su) { - double v; - v = ((*p->hspan)(p, su)); - if (v < 0.0) - v = 0.0; - return((size_t) /* LINTED */ - v); + return (*p->hspan)(p, su); +} + +/* + * Convert a scaling width to basic units, rounding to closest. + */ +int +term_hen(const struct termp *p, const struct roffsu *su) +{ + int bu; + + if ((bu = (*p->hspan)(p, su)) >= 0) + return (bu + 11) / 24; + else + return -((-bu + 11) / 24); }