=================================================================== RCS file: /cvs/mandoc/term.c,v retrieving revision 1.207 retrieving revision 1.221 diff -u -p -r1.207 -r1.221 --- mandoc/term.c 2013/05/29 15:17:52 1.207 +++ mandoc/term.c 2014/04/08 07:13:12 1.221 @@ -1,7 +1,7 @@ -/* $Id: term.c,v 1.207 2013/05/29 15:17:52 schwarze Exp $ */ +/* $Id: term.c,v 1.221 2014/04/08 07:13:12 schwarze Exp $ */ /* * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons - * Copyright (c) 2010, 2011, 2012 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 @@ -23,18 +23,18 @@ #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); @@ -72,48 +72,40 @@ 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 +113,12 @@ 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; @@ -162,7 +159,7 @@ term_flushln(struct termp *p) */ for (j = i, jhy = 0; j < p->col; j++) { - if ((j && ' ' == p->buf[j]) || '\t' == p->buf[j]) + if (' ' == p->buf[j] || '\t' == p->buf[j]) break; /* Back over the the last printed character. */ @@ -175,9 +172,17 @@ 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]) + (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 +194,7 @@ 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 +205,11 @@ 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 +225,7 @@ 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 +234,8 @@ term_flushln(struct termp *p) vbl += (*p->width)(p, ' '); continue; } + if (ASCII_BREAK == p->buf[i]) + continue; /* * Now we definitely know there will be @@ -237,12 +248,6 @@ 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]); @@ -269,15 +274,17 @@ 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; @@ -285,8 +292,7 @@ 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; } @@ -398,6 +404,7 @@ 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 +427,7 @@ 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 +436,15 @@ 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; @@ -480,6 +495,9 @@ term_word(struct termp *p, const char *word) case (ESCAPE_FONTITALIC): term_fontrepl(p, TERMFONT_UNDER); break; + case (ESCAPE_FONTBI): + term_fontrepl(p, TERMFONT_BI); + break; case (ESCAPE_FONT): /* FALLTHROUGH */ case (ESCAPE_FONTROMAN): @@ -501,10 +519,11 @@ term_word(struct termp *p, const char *word) break; } } + p->flags &= ~TERMP_NBRWORD; } static void -adjbuf(struct termp *p, int sz) +adjbuf(struct termp *p, size_t sz) { if (0 == p->maxcols) @@ -512,8 +531,7 @@ 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_realloc(p->buf, sizeof(int) * p->maxcols); } static void @@ -541,73 +559,91 @@ encode1(struct termp *p, int c) return; } - if (p->col + 4 >= p->maxcols) - adjbuf(p, p->col + 4); + if (p->col + 6 >= p->maxcols) + adjbuf(p, p->col + 6); f = term_fonttop(p); - if (TERMFONT_NONE == f) { - p->buf[p->col++] = c; - return; - } else if (TERMFONT_UNDER == f) { + if (TERMFONT_UNDER == f || TERMFONT_BI == f) { p->buf[p->col++] = '_'; - } else - p->buf[p->col++] = c; - - p->buf[p->col++] = 8; + 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; } static void encode(struct termp *p, const char *word, size_t sz) { - enum termfont f; - 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 * character by character. */ - if (TERMFONT_NONE == (f = term_fonttop(p))) { - if (p->col + len >= p->maxcols) - adjbuf(p, p->col + len); - for (i = 0; i < len; i++) + if (TERMFONT_NONE == term_fonttop(p)) { + 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 * 3) >= p->maxcols) - adjbuf(p, p->col + 1 + (len * 3)); + if (p->col + 1 + (sz * 5) >= p->maxcols) + adjbuf(p, p->col + 1 + (sz * 5)); - for (i = 0; i < len; i++) { - if (ASCII_HYPH != word[i] && - ! isgraph((unsigned char)word[i])) { - p->buf[p->col++] = word[i]; - continue; - } - - if (TERMFONT_UNDER == f) - p->buf[p->col++] = '_'; - else if (ASCII_HYPH == word[i]) - p->buf[p->col++] = '-'; + 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]; + } +} - p->buf[p->col++] = 8; - p->buf[p->col++] = word[i]; +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 @@ -635,7 +671,8 @@ 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 @@ -650,7 +687,6 @@ 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 ('\\'): cp++; @@ -724,6 +760,8 @@ term_strlen(const struct termp *p, const char *cp) case (ASCII_HYPH): sz += cond_width(p, '-', &skip); cp++; + /* FALLTHROUGH */ + case (ASCII_BREAK): break; default: break;