=================================================================== RCS file: /cvs/mandoc/term.c,v retrieving revision 1.209 retrieving revision 1.223 diff -u -p -r1.209 -r1.223 --- mandoc/term.c 2013/08/08 20:07:47 1.209 +++ mandoc/term.c 2014/04/23 21:06:41 1.223 @@ -1,7 +1,7 @@ -/* $Id: term.c,v 1.209 2013/08/08 20:07:47 schwarze Exp $ */ +/* $Id: term.c,v 1.223 2014/04/23 21:06:41 schwarze Exp $ */ /* * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons - * Copyright (c) 2010, 2011, 2012, 2013 Ingo Schwarze + * Copyright (c) 2010-2014 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 @@ -23,22 +23,23 @@ #include #include -#include #include #include #include #include "mandoc.h" +#include "mandoc_aux.h" #include "out.h" #include "term.h" #include "main.h" static size_t cond_width(const struct termp *, int, int *); -static void adjbuf(struct termp *p, 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 term_free(struct termp *p) { @@ -51,9 +52,8 @@ term_free(struct termp *p) 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 +63,6 @@ term_begin(struct termp *p, term_margin head, (*p->begin)(p); } - void term_end(struct termp *p) { @@ -72,48 +71,40 @@ 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. - * - * - TERMP_TWOSPACE: make sure there is room for 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 - * 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) { - 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 */ 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 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 */ @@ -121,7 +112,12 @@ term_flushln(struct termp *p) * 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. */ assert (p->rmargin >= p->offset); dv = p->rmargin - p->offset; @@ -174,10 +170,18 @@ term_flushln(struct termp *p) /* Regular word. */ /* Break at the hyphen point if we overrun. */ - if (vend > vis && vend < bp && - ASCII_HYPH == p->buf[j]) + 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]); } @@ -189,7 +193,7 @@ term_flushln(struct termp *p) vend -= vis; (*p->endline)(p); p->viscol = 0; - if (TERMP_NOBREAK & p->flags) { + if (TERMP_BRIND & p->flags) { vbl = p->rmargin; vend += p->rmargin - p->offset; } else @@ -200,7 +204,11 @@ term_flushln(struct termp *p) if (0 < ntab) vbl += ntab * p->tabwidth; - /* Remove the p->overstep width. */ + /* + * 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; @@ -216,7 +224,7 @@ term_flushln(struct termp *p) j = i; while (' ' == p->buf[i]) i++; - dv = (size_t)(i - j) * (*p->width)(p, ' '); + dv = (i - j) * (*p->width)(p, ' '); vbl += dv; vend += dv; break; @@ -225,6 +233,8 @@ term_flushln(struct termp *p) vbl += (*p->width)(p, ' '); continue; } + if (ASCII_BREAK == p->buf[i]) + continue; /* * Now we definitely know there will be @@ -237,16 +247,10 @@ term_flushln(struct termp *p) vbl = 0; } - if (ASCII_HYPH == p->buf[i]) { - (*p->letter)(p, '-'); - p->viscol += (*p->width)(p, '-'); - continue; - } - (*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; @@ -269,15 +273,17 @@ term_flushln(struct termp *p) } if (TERMP_HANG & p->flags) { - /* We need one blank after the tag. */ - p->overstep = (int)(vis - maxvis + (*p->width)(p, ' ')); + p->overstep = (int)(vis - maxvis + + p->trailspace * (*p->width)(p, ' ')); /* * If we have overstepped the margin, temporarily move * it to the right and flag the rest of the line to be * shorter. + * If there is a request to keep the columns together, + * allow negative overstep when the column is not full. */ - if (p->overstep < 0) + if (p->trailspace && p->overstep < 0) p->overstep = 0; return; @@ -285,15 +291,13 @@ term_flushln(struct termp *p) return; /* If the column was overrun, break the line. */ - if (maxvis <= vis + - ((TERMP_TWOSPACE & p->flags) ? (*p->width)(p, ' ') : 0)) { + 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. @@ -307,7 +311,6 @@ 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. @@ -336,7 +339,6 @@ term_fontlast(struct termp *p) p->fontq[p->fonti] = f; } - void term_fontrepl(struct termp *p, enum termfont f) { @@ -345,7 +347,6 @@ term_fontrepl(struct termp *p, enum termfont f) p->fontq[p->fonti] = f; } - void term_fontpush(struct termp *p, enum termfont f) { @@ -355,7 +356,6 @@ term_fontpush(struct termp *p, enum termfont f) p->fontq[++p->fonti] = f; } - const void * term_fontq(struct termp *p) { @@ -363,7 +363,6 @@ term_fontq(struct termp *p) return(&p->fontq[p->fonti]); } - enum termfont term_fonttop(struct termp *p) { @@ -371,7 +370,6 @@ term_fonttop(struct termp *p) return(p->fontq[p->fonti]); } - void term_fontpopq(struct termp *p, const void *key) { @@ -381,7 +379,6 @@ term_fontpopq(struct termp *p, const void *key) assert(p->fonti >= 0); } - void term_fontpop(struct termp *p) { @@ -398,6 +395,7 @@ term_fontpop(struct termp *p) void term_word(struct termp *p, const char *word) { + const char nbrsp[2] = { ASCII_NBRSP, 0 }; const char *seq, *cp; char c; int sz, uc; @@ -420,7 +418,7 @@ term_word(struct termp *p, const char *word) else p->flags |= TERMP_NOSPACE; - p->flags &= ~(TERMP_SENTENCE | TERMP_IGNDELIM); + p->flags &= ~TERMP_SENTENCE; while ('\0' != *word) { if ('\\' != *word) { @@ -429,7 +427,15 @@ term_word(struct termp *p, const char *word) word++; continue; } - ssz = strcspn(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; @@ -442,13 +448,13 @@ term_word(struct termp *p, const char *word) if (TERMENC_ASCII != p->enc) switch (esc) { - case (ESCAPE_UNICODE): + case ESCAPE_UNICODE: uc = mchars_num2uc(seq + 1, sz - 1); if ('\0' == uc) break; encode1(p, uc); continue; - case (ESCAPE_SPECIAL): + case ESCAPE_SPECIAL: uc = mchars_spec2cp(p->symtab, seq, sz); if (uc <= 0) break; @@ -459,55 +465,56 @@ term_word(struct termp *p, const char *word) } switch (esc) { - case (ESCAPE_UNICODE): + case ESCAPE_UNICODE: encode1(p, '?'); break; - case (ESCAPE_NUMBERED): + case ESCAPE_NUMBERED: c = mchars_num2char(seq, sz); if ('\0' != c) encode(p, &c, 1); break; - case (ESCAPE_SPECIAL): + case ESCAPE_SPECIAL: cp = mchars_spec2str(p->symtab, seq, sz, &ssz); - if (NULL != cp) + if (NULL != cp) encode(p, cp, ssz); else if (1 == ssz) encode(p, seq, sz); break; - case (ESCAPE_FONTBOLD): + case ESCAPE_FONTBOLD: term_fontrepl(p, TERMFONT_BOLD); break; - case (ESCAPE_FONTITALIC): + case ESCAPE_FONTITALIC: term_fontrepl(p, TERMFONT_UNDER); break; - case (ESCAPE_FONTBI): + case ESCAPE_FONTBI: term_fontrepl(p, TERMFONT_BI); break; - case (ESCAPE_FONT): + case ESCAPE_FONT: /* FALLTHROUGH */ - case (ESCAPE_FONTROMAN): + case ESCAPE_FONTROMAN: term_fontrepl(p, TERMFONT_NONE); break; - case (ESCAPE_FONTPREV): + case ESCAPE_FONTPREV: term_fontlast(p); break; - case (ESCAPE_NOSPACE): + case ESCAPE_NOSPACE: if (TERMP_SKIPCHAR & p->flags) p->flags &= ~TERMP_SKIPCHAR; else if ('\0' == *word) p->flags |= TERMP_NOSPACE; break; - case (ESCAPE_SKIPCHAR): + case ESCAPE_SKIPCHAR: p->flags |= TERMP_SKIPCHAR; break; default: break; } } + p->flags &= ~TERMP_NBRWORD; } static void -adjbuf(struct termp *p, int sz) +adjbuf(struct termp *p, size_t sz) { if (0 == p->maxcols) @@ -515,8 +522,7 @@ adjbuf(struct termp *p, int sz) while (sz >= p->maxcols) p->maxcols <<= 2; - p->buf = mandoc_realloc - (p->buf, sizeof(int) * (size_t)p->maxcols); + p->buf = mandoc_reallocarray(p->buf, p->maxcols, sizeof(int)); } static void @@ -566,16 +572,13 @@ encode1(struct termp *p, int c) static void encode(struct termp *p, const char *word, size_t sz) { - int i, len; + size_t i; if (TERMP_SKIPCHAR & p->flags) { p->flags &= ~TERMP_SKIPCHAR; return; } - /* LINTED */ - len = sz; - /* * Encode and buffer a string of characters. If the current * font mode is unset, buffer directly, else encode then buffer @@ -583,19 +586,19 @@ encode(struct termp *p, const char *word, size_t sz) */ if (TERMFONT_NONE == term_fonttop(p)) { - if (p->col + len >= p->maxcols) - adjbuf(p, p->col + len); - for (i = 0; i < len; i++) + if (p->col + sz >= p->maxcols) + adjbuf(p, p->col + sz); + for (i = 0; i < sz; i++) p->buf[p->col++] = word[i]; return; } /* Pre-buffer, assuming worst-case. */ - if (p->col + 1 + (len * 5) >= p->maxcols) - adjbuf(p, p->col + 1 + (len * 5)); + if (p->col + 1 + (sz * 5) >= p->maxcols) + adjbuf(p, p->col + 1 + (sz * 5)); - for (i = 0; i < len; i++) { + for (i = 0; i < sz; i++) { if (ASCII_HYPH == word[i] || isgraph((unsigned char)word[i])) encode1(p, word[i]); @@ -604,6 +607,36 @@ encode(struct termp *p, const char *word, size_t sz) } } +void +term_setwidth(struct termp *p, const char *wstr) +{ + struct roffsu su; + size_t width; + int iop; + + 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) { @@ -629,7 +662,8 @@ term_strlen(const struct termp *p, const char *cp) int ssz, skip, c; const char *seq, *rhs; enum mandoc_esc esc; - static const char rej[] = { '\\', ASCII_HYPH, ASCII_NBRSP, '\0' }; + static const char rej[] = { '\\', ASCII_NBRSP, ASCII_HYPH, + ASCII_BREAK, '\0' }; /* * Account for escaped sequences within string length @@ -644,9 +678,8 @@ term_strlen(const struct termp *p, const char *cp) for (i = 0; i < rsz; i++) sz += cond_width(p, *cp++, &skip); - c = 0; switch (*cp) { - case ('\\'): + case '\\': cp++; esc = mandoc_escape(&cp, &seq, &ssz); if (ESCAPE_ERROR == esc) @@ -654,16 +687,16 @@ term_strlen(const struct termp *p, const char *cp) if (TERMENC_ASCII != p->enc) switch (esc) { - case (ESCAPE_UNICODE): - c = mchars_num2uc - (seq + 1, ssz - 1); + 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); + case ESCAPE_SPECIAL: + c = mchars_spec2cp(p->symtab, + seq, ssz); if (c <= 0) break; sz += cond_width(p, c, &skip); @@ -675,17 +708,17 @@ term_strlen(const struct termp *p, const char *cp) rhs = NULL; switch (esc) { - case (ESCAPE_UNICODE): + case ESCAPE_UNICODE: sz += cond_width(p, '?', &skip); break; - case (ESCAPE_NUMBERED): + case ESCAPE_NUMBERED: c = mchars_num2char(seq, ssz); if ('\0' != c) sz += cond_width(p, c, &skip); break; - case (ESCAPE_SPECIAL): - rhs = mchars_spec2str - (p->symtab, seq, ssz, &rsz); + case ESCAPE_SPECIAL: + rhs = mchars_spec2str(p->symtab, + seq, ssz, &rsz); if (ssz != 1 || rhs) break; @@ -693,7 +726,7 @@ term_strlen(const struct termp *p, const char *cp) rhs = seq; rsz = ssz; break; - case (ESCAPE_SKIPCHAR): + case ESCAPE_SKIPCHAR: skip = 1; break; default: @@ -711,13 +744,15 @@ term_strlen(const struct termp *p, const char *cp) 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: break; default: break; @@ -727,29 +762,28 @@ 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): + case SCALE_CM: r = su->scale * 2; break; - case (SCALE_IN): + case SCALE_IN: r = su->scale * 6; break; - case (SCALE_PC): + case SCALE_PC: r = su->scale; break; - case (SCALE_PT): + case SCALE_PT: r = su->scale / 8; break; - case (SCALE_MM): + case SCALE_MM: r = su->scale / 1000; break; - case (SCALE_VS): + case SCALE_VS: r = su->scale; break; default: @@ -759,8 +793,7 @@ term_vspan(const struct termp *p, const struct roffsu if (r < 0.0) r = 0.0; - return(/* LINTED */(size_t) - r); + return((size_t)r); } size_t @@ -771,6 +804,5 @@ term_hspan(const struct termp *p, const struct roffsu v = ((*p->hspan)(p, su)); if (v < 0.0) v = 0.0; - return((size_t) /* LINTED */ - v); + return((size_t)v); }