=================================================================== RCS file: /cvs/mandoc/term.c,v retrieving revision 1.116 retrieving revision 1.191 diff -u -p -r1.116 -r1.191 --- mandoc/term.c 2009/10/28 06:54:12 1.116 +++ mandoc/term.c 2011/05/15 22:29:50 1.191 @@ -1,6 +1,7 @@ -/* $Id: term.c,v 1.116 2009/10/28 06:54:12 kristaps Exp $ */ +/* $Id: term.c,v 1.191 2011/05/15 22:29:50 kristaps Exp $ */ /* - * Copyright (c) 2008, 2009 Kristaps Dzonsons + * Copyright (c) 2008, 2009, 2010 Kristaps Dzonsons + * Copyright (c) 2010, 2011 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 @@ -14,72 +15,67 @@ * 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 + #include -#include +#include +#include #include #include #include -#include -#include "chars.h" +#include "mandoc.h" #include "out.h" #include "term.h" -#include "man.h" -#include "mdoc.h" #include "main.h" -/* FIXME: accomodate non-breaking, non-collapsing white-space. */ -/* FIXME: accomodate non-breaking, collapsing white-space. */ +static void adjbuf(struct termp *p, int); +static void bufferc(struct termp *, char); +static void encode(struct termp *, const char *, size_t); -static struct termp *term_alloc(enum termenc); -static void term_free(struct termp *); - -static void do_escaped(struct termp *, const char **); -static void do_special(struct termp *, - const char *, size_t); -static void do_reserved(struct termp *, - const char *, size_t); -static void buffer(struct termp *, char); -static void encode(struct termp *, char); - - -void * -ascii_alloc(void) +void +term_free(struct termp *p) { - return(term_alloc(TERMENC_ASCII)); + if (p->buf) + free(p->buf); + if (p->symtab) + mchars_free(p->symtab); + + free(p); } void -terminal_free(void *arg) +term_begin(struct termp *p, term_margin head, + term_margin foot, const void *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 * +struct termp * term_alloc(enum termenc enc) { - struct termp *p; + struct termp *p; - if (NULL == (p = calloc(1, sizeof(struct termp)))) - return(NULL); - p->maxrmargin = 78; + p = mandoc_calloc(1, sizeof(struct termp)); p->enc = enc; return(p); } @@ -89,13 +85,11 @@ term_alloc(enum termenc enc) * 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. + * fragment of a columnar list (`Bl -tag' or `Bl -column'), which does + * not have a trailing newline. * - * Specifically, a line is whatever's in p->buf of length p->col, which - * is zeroed after this function returns. + * The following flags may be specified: * - * The usage of termp:flags is as follows: - * * - 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. @@ -128,11 +122,13 @@ 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 vsz; /* visual characters to write 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; - static int overstep = 0; + 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 */ /* * First, establish the maximum columns of "visible" content. @@ -140,108 +136,143 @@ term_flushln(struct termp *p) * an indentation, but can be, for tagged lists or columns, a * small set of values. */ + 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; - assert(p->offset < p->rmargin); - - maxvis = (int)(p->rmargin - p->offset) - overstep < 0 ? - 0 : p->rmargin - p->offset - overstep; - mmax = (int)(p->maxrmargin - p->offset) - overstep < 0 ? - 0 : p->maxrmargin - p->offset - overstep; - bp = TERMP_NOBREAK & p->flags ? mmax : maxvis; - /* - * FIXME: if bp is zero, we still output the first word before - * breaking the line. - */ - - vis = 0; - /* - * If in the standard case (left-justified), then begin with our - * indentation, otherwise (columns, etc.) just start spitting - * out text. + * Indent the first line of a paragraph. */ + vbl = p->flags & TERMP_NOLPAD ? (size_t)0 : p->offset; - if ( ! (p->flags & TERMP_NOLPAD)) - /* LINTED */ - for (j = 0; j < (int)p->offset; j++) - putchar(' '); + vis = vend = 0; + i = 0; - for (i = 0; i < (int)p->col; i++) { + while (i < p->col) { /* + * Handle literal tab characters: collapse all + * subsequent tabs into a single huge set of spaces. + */ + while (i < p->col && '\t' == p->buf[i]) { + vend = (vis / p->tabwidth + 1) * p->tabwidth; + vbl += vend - vis; + vis = vend; + i++; + } + + /* * 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). */ - /* LINTED */ - for (j = i, vsz = 0; j < (int)p->col; j++) { - if (j && ' ' == p->buf[j]) + for (j = i, jhy = 0; j < p->col; j++) { + if ((j && ' ' == p->buf[j]) || '\t' == p->buf[j]) break; - else if (8 == p->buf[j]) - vsz--; - else - vsz++; + + /* Back over the 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]) + jhy = j; + + vend += (*p->width)(p, p->buf[j]); } /* - * Choose the number of blanks to prepend: no blank at the - * beginning of a line, one between words -- but do not - * actually write them yet. - */ - vbl = (size_t)(0 == vis ? 0 : 1); - - /* * Find out whether we would exceed the right margin. - * If so, break to the next line. (TODO: hyphenate) - * Otherwise, write the chosen number of blanks now. + * If so, break to the next line. */ - if (vis && vis + vbl + vsz > bp) { - putchar('\n'); + if (vend > bp && 0 == jhy && vis > 0) { + vend -= vis; + (*p->endline)(p); if (TERMP_NOBREAK & p->flags) { - for (j = 0; j < (int)p->rmargin; j++) - putchar(' '); - vis = p->rmargin - p->offset; + p->viscol = p->rmargin; + (*p->advance)(p, p->rmargin); + vend += p->rmargin - p->offset; } else { - for (j = 0; j < (int)p->offset; j++) - putchar(' '); - vis = 0; + p->viscol = 0; + vbl = p->offset; } - /* Remove the overstep width. */ - bp += (int)/* LINTED */ - overstep; - overstep = 0; - } else { - for (j = 0; j < (int)vbl; j++) - putchar(' '); - vis += vbl; + + /* Remove the p->overstep width. */ + + bp += (size_t)p->overstep; + p->overstep = 0; } - /* - * Finally, write out the word. - */ - for ( ; i < (int)p->col; i++) { - if (' ' == p->buf[i]) + /* Write out the [remaining] word. */ + for ( ; i < p->col; i++) { + if (vend > bp && jhy > 0 && i > jhy) break; - putchar(p->buf[i]); + 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 (ASCII_NBRSP == p->buf[i]) { + vbl += (*p->width)(p, ' '); + continue; + } + + /* + * Now we definitely know there will be + * printable characters to output, + * so write preceding white space now. + */ + if (vbl) { + (*p->advance)(p, vbl); + p->viscol += vbl; + vbl = 0; + } + + 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]); + } } - vis += vsz; + vis = vend; } + /* + * If there was trailing white space, it was not printed; + * so reset the cursor position accordingly. + */ + vis -= vbl; + p->col = 0; - overstep = 0; + p->overstep = 0; 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. */ - overstep = /* LINTED */ - vis - maxvis + 1; + p->overstep = (int)(vis - maxvis + (*p->width)(p, ' ')); /* * Behave exactly the same way as groff: @@ -253,25 +284,25 @@ term_flushln(struct termp *p) * move it one step LEFT and flag the rest of the line * to be longer. */ - if (overstep >= -1) { - assert((int)maxvis + overstep >= 0); - /* LINTED */ - maxvis += overstep; + if (p->overstep >= -1) { + assert((int)maxvis + p->overstep >= 0); + maxvis += (size_t)p->overstep; } else - overstep = 0; + p->overstep = 0; } 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(' '); + 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); } } @@ -286,7 +317,7 @@ term_newln(struct termp *p) { p->flags |= TERMP_NOSPACE; - if (0 == p->col) { + if (0 == p->col && 0 == p->viscol) { p->flags &= ~TERMP_NOLPAD; return; } @@ -306,160 +337,74 @@ term_vspace(struct termp *p) { term_newln(p); - putchar('\n'); + p->viscol = 0; + (*p->endline)(p); } - -static void -do_special(struct termp *p, const char *word, size_t len) +void +term_fontlast(struct termp *p) { - const char *rhs; - size_t sz; - int i; + enum termfont f; - rhs = chars_a2ascii(p->symtab, word, len, &sz); - - if (NULL == rhs) { -#if 0 - fputs("Unknown special character: ", stderr); - for (i = 0; i < (int)len; i++) - fputc(word[i], stderr); - fputc('\n', stderr); -#endif - return; - } - for (i = 0; i < (int)sz; i++) - encode(p, rhs[i]); + f = p->fontl; + p->fontl = p->fontq[p->fonti]; + p->fontq[p->fonti] = f; } -static void -do_reserved(struct termp *p, const char *word, size_t len) +void +term_fontrepl(struct termp *p, enum termfont f) { - const char *rhs; - size_t sz; - int i; - rhs = chars_a2res(p->symtab, word, len, &sz); - - if (NULL == rhs) { -#if 0 - fputs("Unknown reserved word: ", stderr); - for (i = 0; i < (int)len; i++) - fputc(word[i], stderr); - fputc('\n', stderr); -#endif - return; - } - for (i = 0; i < (int)sz; i++) - encode(p, rhs[i]); + p->fontl = p->fontq[p->fonti]; + p->fontq[p->fonti] = f; } -/* - * Handle an escape sequence: determine its length and pass it to the - * escape-symbol look table. Note that we assume mdoc(3) has validated - * the escape sequence (we assert upon badly-formed escape sequences). - */ -static void -do_escaped(struct termp *p, const char **word) +void +term_fontpush(struct termp *p, enum termfont f) { - int j, type; - const char *wp; - wp = *word; - type = 1; + assert(p->fonti + 1 < 10); + p->fontl = p->fontq[p->fonti]; + p->fontq[++p->fonti] = f; +} - if (0 == *(++wp)) { - *word = wp; - return; - } - if ('(' == *wp) { - wp++; - if (0 == *wp || 0 == *(wp + 1)) { - *word = 0 == *wp ? wp : wp + 1; - return; - } +const void * +term_fontq(struct termp *p) +{ - do_special(p, wp, 2); - *word = ++wp; - return; + return(&p->fontq[p->fonti]); +} - } else if ('*' == *wp) { - if (0 == *(++wp)) { - *word = wp; - return; - } - switch (*wp) { - case ('('): - wp++; - if (0 == *wp || 0 == *(wp + 1)) { - *word = 0 == *wp ? wp : wp + 1; - return; - } +enum termfont +term_fonttop(struct termp *p) +{ - do_reserved(p, wp, 2); - *word = ++wp; - return; - case ('['): - type = 0; - break; - default: - do_reserved(p, wp, 1); - *word = wp; - return; - } - - } else if ('f' == *wp) { - if (0 == *(++wp)) { - *word = wp; - return; - } + return(p->fontq[p->fonti]); +} - switch (*wp) { - case ('B'): - p->bold++; - break; - case ('I'): - p->under++; - break; - case ('P'): - /* FALLTHROUGH */ - case ('R'): - p->bold = p->under = 0; - break; - default: - break; - } - *word = wp; - return; +void +term_fontpopq(struct termp *p, const void *key) +{ - } else if ('[' != *wp) { - do_special(p, wp, 1); - *word = wp; - return; - } + while (p->fonti >= 0 && key != &p->fontq[p->fonti]) + p->fonti--; + assert(p->fonti >= 0); +} - wp++; - for (j = 0; *wp && ']' != *wp; wp++, j++) - /* Loop... */ ; - if (0 == *wp) { - *word = wp; - return; - } +void +term_fontpop(struct termp *p) +{ - if (type) - do_special(p, wp - j, (size_t)j); - else - do_reserved(p, wp - j, (size_t)j); - *word = wp; + assert(p->fonti); + 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 @@ -468,107 +413,233 @@ do_escaped(struct termp *p, const char **word) void term_word(struct termp *p, const char *word) { - const char *sv; + const char *seq, *cp; + char c; + int sz; + size_t ssz; + enum mandoc_esc esc; - sv = word; + 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) + bufferc(p, ' '); + } else + bufferc(p, ASCII_NBRSP); + } - if (word[0] && 0 == word[1]) - switch (word[0]) { - case('.'): - /* FALLTHROUGH */ - 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)) - buffer(p, ' '); - if ( ! (p->flags & TERMP_NONOSPACE)) p->flags &= ~TERMP_NOSPACE; + else + p->flags |= TERMP_NOSPACE; - for ( ; *word; word++) + p->flags &= ~(TERMP_SENTENCE | TERMP_IGNDELIM); + + while ('\0' != *word) { + if ((ssz = strcspn(word, "\\")) > 0) + encode(p, word, ssz); + + word += (int)ssz; if ('\\' != *word) - encode(p, *word); - else - do_escaped(p, &word); + continue; - if (sv[0] && 0 == sv[1]) - switch (sv[0]) { - case('('): - /* FALLTHROUGH */ - case('['): - /* FALLTHROUGH */ - case('{'): - p->flags |= TERMP_NOSPACE; + 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); + break; + case (ESCAPE_PREDEF): + cp = mchars_res2str(p->symtab, seq, sz, &ssz); + if (NULL != cp) + encode(p, cp, ssz); + 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): + term_fontrepl(p, TERMFONT_BOLD); + break; + case (ESCAPE_FONTITALIC): + term_fontrepl(p, TERMFONT_UNDER); + break; + case (ESCAPE_FONTROMAN): + term_fontrepl(p, TERMFONT_NONE); + break; + case (ESCAPE_FONTPREV): + term_fontlast(p); + break; + case (ESCAPE_NOSPACE): + if ('\0' == *word) + p->flags |= TERMP_NOSPACE; + break; default: break; } + } } - -/* - * Insert a single character into the line-buffer. If the buffer's - * space is exceeded, then allocate more space by doubling the buffer - * size. - */ static void -buffer(struct termp *p, char c) +adjbuf(struct termp *p, int sz) { - size_t s; - if (p->col + 1 >= p->maxcols) { - if (0 == p->maxcols) - p->maxcols = 256; - s = p->maxcols * 2; - p->buf = realloc(p->buf, s); - if (NULL == p->buf) - err(1, "realloc"); /* FIXME: shouldn't be here! */ - p->maxcols = s; - } - p->buf[(int)(p->col)++] = c; + 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); } +static void +bufferc(struct termp *p, char c) +{ + if (p->col + 1 >= p->maxcols) + adjbuf(p, p->col + 1); + + p->buf[p->col++] = c; +} + static void -encode(struct termp *p, char c) +encode(struct termp *p, const char *word, size_t sz) { - - if (' ' != c) { - if (p->under) { - buffer(p, '_'); - buffer(p, 8); + enum termfont f; + int i, len; + + /* 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++) + 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)); + + for (i = 0; i < len; i++) { + if ( ! isgraph((unsigned char)word[i])) { + p->buf[p->col++] = word[i]; + continue; } - if (p->bold) { - buffer(p, c); - buffer(p, 8); - } + + if (TERMFONT_UNDER == f) + p->buf[p->col++] = '_'; + else + p->buf[p->col++] = word[i]; + + p->buf[p->col++] = 8; + p->buf[p->col++] = word[i]; } - buffer(p, c); } +size_t +term_len(const struct termp *p, size_t sz) +{ + return((*p->width)(p, ' ') * sz); +} + + size_t -term_vspan(const struct roffsu *su) +term_strlen(const struct termp *p, const char *cp) { + size_t sz, rsz, i; + int ssz, c; + const char *seq, *rhs; + static const char rej[] = { '\\', ASCII_HYPH, ASCII_NBRSP, '\0' }; + + /* + * 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; + while ('\0' != *cp) { + rsz = strcspn(cp, rej); + for (i = 0; i < rsz; i++) + sz += (*p->width)(p, *cp++); + + switch (*cp) { + 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); + break; + case (ESCAPE_PREDEF): + rhs = mchars_res2str + (p->symtab, seq, ssz, &rsz); + break; + case (ESCAPE_SPECIAL): + rhs = mchars_spec2str + (p->symtab, seq, ssz, &rsz); + + if (ssz != 1 || rhs) + break; + + rhs = seq; + rsz = ssz; + break; + default: + break; + } + + if (NULL == rhs) + break; + + for (i = 0; i < rsz; i++) + sz += (*p->width)(p, *rhs++); + break; + case (ASCII_NBRSP): + sz += (*p->width)(p, ' '); + cp++; + break; + case (ASCII_HYPH): + sz += (*p->width)(p, '-'); + cp++; + break; + default: + break; + } + } + + return(sz); +} + +/* ARGSUSED */ +size_t +term_vspan(const struct termp *p, const struct roffsu *su) +{ double r; switch (su->unit) { @@ -601,43 +672,14 @@ term_vspan(const struct roffsu *su) r); } - size_t -term_hspan(const struct roffsu *su) +term_hspan(const struct termp *p, const struct roffsu *su) { - double r; + double v; - /* XXX: CM, IN, and PT are approximations. */ - - switch (su->unit) { - case (SCALE_CM): - r = 4 * su->scale; - break; - case (SCALE_IN): - /* XXX: this is an approximation. */ - r = 10 * su->scale; - break; - case (SCALE_PC): - r = (10 * su->scale) / 6; - break; - case (SCALE_PT): - r = (10 * su->scale) / 72; - break; - case (SCALE_MM): - r = su->scale / 1000; /* FIXME: double-check. */ - break; - case (SCALE_VS): - r = su->scale * 2 - 1; /* FIXME: double-check. */ - break; - default: - r = su->scale; - break; - } - - if (r < 0.0) - r = 0.0; - return((size_t)/* LINTED */ - r); + v = ((*p->hspan)(p, su)); + if (v < 0.0) + v = 0.0; + return((size_t) /* LINTED */ + v); } - -