=================================================================== RCS file: /cvs/mandoc/term.c,v retrieving revision 1.219 retrieving revision 1.238 diff -u -p -r1.219 -r1.238 --- mandoc/term.c 2014/03/30 21:28:01 1.219 +++ mandoc/term.c 2014/12/19 17:12:04 1.238 @@ -1,4 +1,4 @@ -/* $Id: term.c,v 1.219 2014/03/30 21:28:01 schwarze Exp $ */ +/* $Id: term.c,v 1.238 2014/12/19 17:12:04 schwarze Exp $ */ /* * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons * Copyright (c) 2010-2014 Ingo Schwarze @@ -15,9 +15,7 @@ * 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 @@ -39,21 +37,18 @@ static void bufferc(struct termp *, char); static void encode(struct termp *, const char *, size_t); static void encode1(struct termp *, int); + void term_free(struct termp *p) { - if (p->buf) - free(p->buf); - if (p->symtab) - mchars_free(p->symtab); - + free(p->buf); + free(p->fontq); free(p); } - void -term_begin(struct termp *p, term_margin head, +term_begin(struct termp *p, term_margin head, term_margin foot, const void *arg) { @@ -63,7 +58,6 @@ term_begin(struct termp *p, term_margin head, (*p->begin)(p); } - void term_end(struct termp *p) { @@ -72,34 +66,27 @@ term_end(struct termp *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_NOBREAK: this is the most important and is used when making - * columns. In short: don't print a newline and instead expect the - * next call to do the padding up to the start of the next column. - * p->trailspace may be set to 0, 1, or 2, depending on how many - * space characters are required at the end of the column. - * - * - 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 - * overrunning, 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_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) @@ -114,7 +101,7 @@ term_flushln(struct termp *p) 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 */ - size_t mmax; /* used in calculating bp */ + size_t rmargin; /* the rightmost of the two margins */ /* * First, establish the maximum columns of "visible" content. @@ -127,13 +114,17 @@ term_flushln(struct termp *p) * is negative, it gets sign extended. Subtracting that * very large size_t effectively adds a small number to dv. */ - assert (p->rmargin >= p->offset); - dv = p->rmargin - p->offset; + rmargin = p->rmargin > p->offset ? 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; + 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; /* * Calculate the required amount of padding. @@ -178,7 +169,7 @@ term_flushln(struct termp *p) /* Regular word. */ /* Break at the hyphen point if we overrun. */ - if (vend > vis && vend < bp && + if (vend > vis && vend < bp && (ASCII_HYPH == p->buf[j] || ASCII_BREAK == p->buf[j])) jhy = j; @@ -201,9 +192,9 @@ term_flushln(struct termp *p) vend -= vis; (*p->endline)(p); p->viscol = 0; - if (TERMP_NOBREAK & p->flags) { - vbl = p->rmargin; - vend += p->rmargin - p->offset; + if (TERMP_BRIND & p->flags) { + vbl = rmargin; + vend += rmargin - p->offset; } else vbl = p->offset; @@ -230,7 +221,7 @@ term_flushln(struct termp *p) break; if (' ' == p->buf[i]) { j = i; - while (' ' == p->buf[i]) + while (i < p->col && ' ' == p->buf[i]) i++; dv = (i - j) * (*p->width)(p, ' '); vbl += dv; @@ -258,7 +249,7 @@ term_flushln(struct termp *p) (*p->letter)(p, p->buf[i]); if (8 == p->buf[i]) p->viscol -= (*p->width)(p, p->buf[i-1]); - else + else p->viscol += (*p->width)(p, p->buf[i]); } vis = vend; @@ -268,8 +259,10 @@ term_flushln(struct termp *p) * If there was trailing white space, it was not printed; * so reset the cursor position accordingly. */ - if (vis) + if (vis > vbl) vis -= vbl; + else + vis = 0; p->col = 0; p->overstep = 0; @@ -282,7 +275,7 @@ term_flushln(struct termp *p) if (TERMP_HANG & p->flags) { p->overstep = (int)(vis - maxvis + - p->trailspace * (*p->width)(p, ' ')); + p->trailspace * (*p->width)(p, ' ')); /* * If we have overstepped the margin, temporarily move @@ -305,8 +298,7 @@ term_flushln(struct termp *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. @@ -320,7 +312,6 @@ term_newln(struct termp *p) 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. @@ -339,6 +330,7 @@ term_vspace(struct termp *p) (*p->endline)(p); } +/* Swap current and previous font; for \fP and .ft P */ void term_fontlast(struct termp *p) { @@ -349,7 +341,7 @@ 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) { @@ -358,43 +350,39 @@ 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(enum termfont *)); + } + p->fontq[p->fonti] = f; } - -const void * +/* Retrieve pointer to current font. */ +const enum termfont * 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, const enum termfont *key) { - while (p->fonti >= 0 && key < (void *)(p->fontq + p->fonti)) + while (p->fonti >= 0 && key < p->fontq + p->fonti) p->fonti--; assert(p->fonti >= 0); } - +/* Pop one font off the stack. */ void term_fontpop(struct termp *p) { @@ -413,7 +401,6 @@ term_word(struct termp *p, const char *word) { const char nbrsp[2] = { ASCII_NBRSP, 0 }; const char *seq, *cp; - char c; int sz, uc; size_t ssz; enum mandoc_esc esc; @@ -434,7 +421,7 @@ term_word(struct termp *p, const char *word) else p->flags |= TERMP_NOSPACE; - p->flags &= ~TERMP_SENTENCE; + p->flags &= ~(TERMP_SENTENCE | TERMP_NONEWLINE); while ('\0' != *word) { if ('\\' != *word) { @@ -460,71 +447,73 @@ term_word(struct termp *p, const char *word) word++; esc = mandoc_escape(&word, &seq, &sz); if (ESCAPE_ERROR == esc) - break; + continue; - if (TERMENC_ASCII != p->enc) - switch (esc) { - case (ESCAPE_UNICODE): - uc = mchars_num2uc(seq + 1, sz - 1); - if ('\0' == uc) - break; - encode1(p, uc); - continue; - case (ESCAPE_SPECIAL): - uc = mchars_spec2cp(p->symtab, seq, sz); - if (uc <= 0) - break; - encode1(p, uc); - continue; - default: - break; - } - switch (esc) { - case (ESCAPE_UNICODE): - encode1(p, '?'); + case ESCAPE_UNICODE: + uc = mchars_num2uc(seq + 1, sz - 1); break; - case (ESCAPE_NUMBERED): - c = mchars_num2char(seq, sz); - if ('\0' != c) - encode(p, &c, 1); + case ESCAPE_NUMBERED: + uc = mchars_num2char(seq, sz); + if (uc < 0) + continue; 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_SPECIAL: + if (p->enc == TERMENC_ASCII) { + cp = mchars_spec2str(p->symtab, + seq, sz, &ssz); + if (cp != NULL) + encode(p, cp, ssz); + } else { + uc = mchars_spec2cp(p->symtab, seq, sz); + if (uc > 0) + encode1(p, uc); + } + continue; + case ESCAPE_FONTBOLD: term_fontrepl(p, TERMFONT_BOLD); - break; - case (ESCAPE_FONTITALIC): + continue; + case ESCAPE_FONTITALIC: term_fontrepl(p, TERMFONT_UNDER); - break; - case (ESCAPE_FONTBI): + continue; + case ESCAPE_FONTBI: term_fontrepl(p, TERMFONT_BI); - break; - case (ESCAPE_FONT): + continue; + case ESCAPE_FONT: /* FALLTHROUGH */ - case (ESCAPE_FONTROMAN): + case ESCAPE_FONTROMAN: term_fontrepl(p, TERMFONT_NONE); - break; - case (ESCAPE_FONTPREV): + continue; + case ESCAPE_FONTPREV: term_fontlast(p); - break; - case (ESCAPE_NOSPACE): + continue; + case ESCAPE_NOSPACE: if (TERMP_SKIPCHAR & p->flags) p->flags &= ~TERMP_SKIPCHAR; else if ('\0' == *word) - p->flags |= TERMP_NOSPACE; - break; - case (ESCAPE_SKIPCHAR): + p->flags |= (TERMP_NOSPACE | TERMP_NONEWLINE); + continue; + case ESCAPE_SKIPCHAR: p->flags |= TERMP_SKIPCHAR; - break; + 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; } @@ -538,7 +527,7 @@ adjbuf(struct termp *p, size_t sz) while (sz >= p->maxcols) p->maxcols <<= 2; - p->buf = mandoc_realloc(p->buf, sizeof(int) * p->maxcols); + p->buf = mandoc_reallocarray(p->buf, p->maxcols, sizeof(int)); } static void @@ -569,7 +558,7 @@ encode1(struct termp *p, int c) if (p->col + 6 >= p->maxcols) adjbuf(p, p->col + 6); - f = term_fonttop(p); + f = *term_fontq(p); if (TERMFONT_UNDER == f || TERMFONT_BI == f) { p->buf[p->col++] = '_'; @@ -601,8 +590,8 @@ encode(struct termp *p, const char *word, size_t sz) * character by character. */ - if (TERMFONT_NONE == term_fonttop(p)) { - if (p->col + sz >= p->maxcols) + if (*term_fontq(p) == TERMFONT_NONE) { + if (p->col + sz >= p->maxcols) adjbuf(p, p->col + sz); for (i = 0; i < sz; i++) p->buf[p->col++] = word[i]; @@ -630,26 +619,26 @@ term_setwidth(struct termp *p, const char *wstr) size_t width; int iop; + iop = 0; + width = 0; if (NULL != wstr) { switch (*wstr) { - case ('+'): + case '+': iop = 1; wstr++; break; - case ('-'): + case '-': iop = -1; wstr++; break; default: - iop = 0; break; } - if ( ! a2roffsu(wstr, &su, SCALE_MAX)) { - wstr = NULL; + if (a2roffsu(wstr, &su, SCALE_MAX)) + width = term_hspan(p, &su); + else iop = 0; - } } - width = (NULL != wstr) ? term_hspan(p, &su) : 0; (*p->setwidth)(p, iop, width); } @@ -675,7 +664,7 @@ size_t term_strlen(const struct termp *p, const char *cp) { size_t sz, rsz, i; - int ssz, skip, c; + int ssz, skip, uc; const char *seq, *rhs; enum mandoc_esc esc; static const char rej[] = { '\\', ASCII_NBRSP, ASCII_HYPH, @@ -695,80 +684,83 @@ term_strlen(const struct termp *p, const char *cp) sz += cond_width(p, *cp++, &skip); switch (*cp) { - case ('\\'): + case '\\': cp++; esc = mandoc_escape(&cp, &seq, &ssz); if (ESCAPE_ERROR == esc) - return(sz); + continue; - if (TERMENC_ASCII != p->enc) - switch (esc) { - case (ESCAPE_UNICODE): - c = mchars_num2uc - (seq + 1, ssz - 1); - if ('\0' == c) - break; - sz += cond_width(p, c, &skip); - continue; - case (ESCAPE_SPECIAL): - c = mchars_spec2cp - (p->symtab, seq, ssz); - if (c <= 0) - break; - sz += cond_width(p, c, &skip); - continue; - default: - break; - } - rhs = NULL; switch (esc) { - case (ESCAPE_UNICODE): - sz += cond_width(p, '?', &skip); + case ESCAPE_UNICODE: + uc = mchars_num2uc(seq + 1, ssz - 1); break; - case (ESCAPE_NUMBERED): - c = mchars_num2char(seq, ssz); - if ('\0' != c) - sz += cond_width(p, c, &skip); + 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; - break; - case (ESCAPE_SKIPCHAR): + case ESCAPE_SPECIAL: + if (p->enc == TERMENC_ASCII) { + rhs = mchars_spec2str(p->symtab, + seq, ssz, &rsz); + if (rhs != NULL) + break; + } else { + uc = mchars_spec2cp(p->symtab, + seq, ssz); + if (uc > 0) + sz += cond_width(p, uc, &skip); + } + continue; + case ESCAPE_SKIPCHAR: skip = 1; - break; + continue; default: - break; + continue; } - if (NULL == rhs) - break; + /* + * 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): + case ASCII_NBRSP: sz += cond_width(p, ' ', &skip); cp++; break; - case (ASCII_HYPH): + case ASCII_HYPH: sz += cond_width(p, '-', &skip); cp++; /* FALLTHROUGH */ - case (ASCII_BREAK): + case ASCII_BREAK: break; default: break; @@ -778,40 +770,38 @@ term_strlen(const struct termp *p, const char *cp) return(sz); } -/* ARGSUSED */ size_t term_vspan(const struct termp *p, const struct roffsu *su) { double r; switch (su->unit) { - case (SCALE_CM): - r = su->scale * 2; + case SCALE_CM: + r = su->scale * 2.0; break; - case (SCALE_IN): - r = su->scale * 6; + case SCALE_IN: + r = su->scale * 6.0; break; - case (SCALE_PC): + case SCALE_PC: r = su->scale; break; - case (SCALE_PT): - r = su->scale / 8; + case SCALE_PT: + r = su->scale / 8.0; break; - case (SCALE_MM): - r = su->scale / 1000; + case SCALE_MM: + r = su->scale / 1000.0; break; - case (SCALE_VS): + case SCALE_VS: r = su->scale; break; default: - r = su->scale - 1; + r = su->scale - 1.0; break; } if (r < 0.0) r = 0.0; - return(/* LINTED */(size_t) - r); + return((size_t)(r + 0.0005)); } size_t @@ -819,9 +809,8 @@ term_hspan(const struct termp *p, const struct roffsu { double v; - v = ((*p->hspan)(p, su)); + v = (*p->hspan)(p, su); if (v < 0.0) v = 0.0; - return((size_t) /* LINTED */ - v); + return((size_t)(v + 0.0005)); }