=================================================================== RCS file: /cvs/mandoc/term.c,v retrieving revision 1.71 retrieving revision 1.129 diff -u -p -r1.71 -r1.129 --- mandoc/term.c 2009/03/31 13:50:19 1.71 +++ mandoc/term.c 2010/03/23 12:42:22 1.129 @@ -1,48 +1,47 @@ -/* $Id: term.c,v 1.71 2009/03/31 13:50:19 kristaps Exp $ */ +/* $Id: term.c,v 1.129 2010/03/23 12:42:22 kristaps Exp $ */ /* - * Copyright (c) 2008, 2009 Kristaps Dzonsons + * Copyright (c) 2008, 2009 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. + * 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 WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED - * WARRANTIES OF MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE - * AUTHOR 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. + * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES + * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF + * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR 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 + #include -#include +#include #include #include #include +#include +#include "chars.h" +#include "out.h" #include "term.h" #include "man.h" #include "mdoc.h" +#include "main.h" -extern int man_run(struct termp *, - const struct man *); -extern int mdoc_run(struct termp *, - const struct mdoc *); - static struct termp *term_alloc(enum termenc); static void term_free(struct termp *); -static void term_pword(struct termp *, const char *, int); -static void term_pescape(struct termp *, - const char *, int *, int); -static void term_nescape(struct termp *, - const char *, size_t); -static void term_chara(struct termp *, char); -static void term_stringa(struct termp *, - const char *, size_t); -static int term_isopendelim(const char *, int); -static int term_isclosedelim(const char *, int); +static void spec(struct termp *, const char *, size_t); +static void res(struct termp *, const char *, size_t); +static void buffera(struct termp *, const char *, size_t); +static void bufferc(struct termp *, char); +static void adjbuf(struct termp *p, size_t); +static void encode(struct termp *, const char *, size_t); void * @@ -53,26 +52,6 @@ ascii_alloc(void) } -int -terminal_run(void *arg, const struct man *man, - const struct mdoc *mdoc) -{ - struct termp *p; - - p = (struct termp *)arg; - - if (NULL == p->symtab) - p->symtab = term_ascii2htab(); - - if (man) - return(man_run(p, man)); - if (mdoc) - return(mdoc_run(p, mdoc)); - - return(1); -} - - void terminal_free(void *arg) { @@ -87,8 +66,8 @@ term_free(struct termp *p) if (p->buf) free(p->buf); - if (TERMENC_ASCII == p->enc && p->symtab) - term_asciifree(p->symtab); + if (p->symtab) + chars_free(p->symtab); free(p); } @@ -99,71 +78,16 @@ term_alloc(enum termenc enc) { struct termp *p; - if (NULL == (p = malloc(sizeof(struct termp)))) - err(1, "malloc"); - bzero(p, sizeof(struct termp)); - p->maxrmargin = 78; + p = calloc(1, sizeof(struct termp)); + if (NULL == p) { + perror(NULL); + exit(EXIT_FAILURE); + } p->enc = enc; return(p); } -static int -term_isclosedelim(const char *p, int len) -{ - - if (1 != len) - return(0); - - switch (*p) { - case('.'): - /* FALLTHROUGH */ - case(','): - /* FALLTHROUGH */ - case(';'): - /* FALLTHROUGH */ - case(':'): - /* FALLTHROUGH */ - case('?'): - /* FALLTHROUGH */ - case('!'): - /* FALLTHROUGH */ - case(')'): - /* FALLTHROUGH */ - case(']'): - /* FALLTHROUGH */ - case('}'): - return(1); - default: - break; - } - - return(0); -} - - -static int -term_isopendelim(const char *p, int len) -{ - - if (1 != len) - return(0); - - switch (*p) { - case('('): - /* FALLTHROUGH */ - case('['): - /* FALLTHROUGH */ - case('{'): - return(1); - default: - break; - } - - return(0); -} - - /* * 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 @@ -173,8 +97,7 @@ term_isopendelim(const char *p, int len) * Specifically, a line is whatever's in p->buf of length p->col, which * is zeroed after this function returns. * - * The variables TERMP_NOLPAD, TERMP_LITERAL and TERMP_NOBREAK are of - * critical importance here. Their behaviour follows: + * 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 @@ -184,37 +107,58 @@ term_isopendelim(const char *p, int len) * columns. In short: don't print a newline and instead pad to the * right margin. Used in conjunction with TERMP_NOLPAD. * - * - TERMP_NONOBREAK: don't newline when TERMP_NOBREAK is specified. + * - 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. Extremely long - * lines will cause the system to emit a warning (TODO: hyphenate, if - * possible). + * writing from the right-margin, which will lead to the above scenario + * upon exit. Otherwise, the line will break at the right margin. */ void term_flushln(struct termp *p) { - int i, j; - size_t vsz, vis, maxvis, mmax, bp; + 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 bp; /* visual right border position */ + int j; /* temporary loop index */ + size_t maxvis, mmax; /* * 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. */ assert(p->offset < p->rmargin); - maxvis = p->rmargin - p->offset; - mmax = p->maxrmargin - p->offset; + + maxvis = (int)(p->rmargin - p->offset) - p->overstep < 0 ? + /* LINTED */ + 0 : p->rmargin - p->offset - p->overstep; + mmax = (int)(p->maxrmargin - p->offset) - p->overstep < 0 ? + /* LINTED */ + 0 : p->maxrmargin - p->offset - p->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; /* @@ -238,82 +182,106 @@ term_flushln(struct termp *p) /* LINTED */ for (j = i, vsz = 0; j < (int)p->col; j++) { - if (' ' == p->buf[j]) + if (j && ' ' == p->buf[j]) break; else if (8 == p->buf[j]) - j += 1; + vsz--; else vsz++; } /* - * Do line-breaking. If we're greater than our - * break-point and already in-line, break to the next - * line and start writing. If we're at the line start, - * then write out the word (TODO: hyphenate) and break - * in a subsequent loop invocation. + * 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); - if ( ! (TERMP_NOBREAK & p->flags)) { - if (vis && vis + vsz > bp) { - putchar('\n'); + /* + * 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 (vis && vis + vbl + vsz > bp) { + putchar('\n'); + if (TERMP_NOBREAK & p->flags) { + for (j = 0; j < (int)p->rmargin; j++) + putchar(' '); + vis = p->rmargin - p->offset; + } else { for (j = 0; j < (int)p->offset; j++) putchar(' '); vis = 0; - } - } else if (vis && vis + vsz > bp) { - putchar('\n'); - for (j = 0; j < (int)p->rmargin; j++) + } + /* Remove the p->overstep width. */ + bp += (int)/* LINTED */ + p->overstep; + p->overstep = 0; + } else { + for (j = 0; j < (int)vbl; j++) putchar(' '); - vis = p->rmargin - p->offset; + vis += vbl; } - /* - * Write out the word and a trailing space. Omit the - * space if we're the last word in the line or beyond - * our breakpoint. + /* + * Finally, write out the word. */ - for ( ; i < (int)p->col; i++) { if (' ' == p->buf[i]) break; - putchar(p->buf[i]); + + /* The unit sep. is a non-breaking space. */ + if (31 == p->buf[i]) + putchar(' '); + else + putchar(p->buf[i]); } vis += vsz; - if (i < (int)p->col && vis <= bp) { - putchar(' '); - vis++; - } } - /* - * If we've overstepped our maximum visible no-break space, then - * cause a newline and offset at the right margin. - */ + p->col = 0; + p->overstep = 0; - if ((TERMP_NOBREAK & p->flags) && vis >= maxvis) { - if ( ! (TERMP_NONOBREAK & p->flags)) { - putchar('\n'); - for (i = 0; i < (int)p->rmargin; i++) - putchar(' '); - } - p->col = 0; + if ( ! (TERMP_NOBREAK & p->flags)) { + putchar('\n'); return; } - /* - * If we're not to right-marginalise it (newline), then instead - * pad to the right margin and stay off. - */ + if (TERMP_HANG & p->flags) { + /* We need one blank after the tag. */ + p->overstep = /* LINTED */ + vis - maxvis + 1; - if (p->flags & TERMP_NOBREAK) { - if ( ! (TERMP_NONOBREAK & p->flags)) - for ( ; vis < maxvis; vis++) - putchar(' '); - } else - putchar('\n'); + /* + * 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. + */ + if (p->overstep >= -1) { + assert((int)maxvis + p->overstep >= 0); + /* LINTED */ + maxvis += p->overstep; + } else + p->overstep = 0; - p->col = 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(' '); + } } @@ -351,144 +319,92 @@ term_vspace(struct termp *p) } -/* - * Break apart a word into "pwords" (partial-words, usually from - * breaking up a phrase into individual words) and, eventually, put them - * into the output buffer. If we're a literal word, then don't break up - * the word and put it verbatim into the output buffer. - */ +static void +spec(struct termp *p, const char *word, size_t len) +{ + const char *rhs; + size_t sz; + + rhs = chars_a2ascii(p->symtab, word, len, &sz); + if (rhs) + encode(p, rhs, sz); +} + + +static void +res(struct termp *p, const char *word, size_t len) +{ + const char *rhs; + size_t sz; + + rhs = chars_a2res(p->symtab, word, len, &sz); + if (rhs) + encode(p, rhs, sz); +} + + void -term_word(struct termp *p, const char *word) +term_fontlast(struct termp *p) { - int i, j, len; + enum termfont f; - len = (int)strlen(word); + f = p->fontl; + p->fontl = p->fontq[p->fonti]; + p->fontq[p->fonti] = f; +} - if (p->flags & TERMP_LITERAL) { - term_pword(p, word, len); - return; - } - /* LINTED */ - for (j = i = 0; i < len; i++) { - if (' ' != word[i]) { - j++; - continue; - } - - /* Escaped spaces don't delimit... */ - if (i && ' ' == word[i] && '\\' == word[i - 1]) { - j++; - continue; - } +void +term_fontrepl(struct termp *p, enum termfont f) +{ - if (0 == j) - continue; - assert(i >= j); - term_pword(p, &word[i - j], j); - j = 0; - } - if (j > 0) { - assert(i >= j); - term_pword(p, &word[i - j], j); - } + p->fontl = p->fontq[p->fonti]; + p->fontq[p->fonti] = f; } -/* - * Determine the symbol indicated by an escape sequences, that is, one - * starting with a backslash. Once done, we pass this value into the - * output buffer by way of the symbol table. - */ -static void -term_nescape(struct termp *p, const char *word, size_t len) +void +term_fontpush(struct termp *p, enum termfont f) { - const char *rhs; - size_t sz; - if (NULL == (rhs = term_a2ascii(p->symtab, word, len, &sz))) - return; - term_stringa(p, rhs, sz); + assert(p->fonti + 1 < 10); + 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 -term_pescape(struct termp *p, const char *word, int *i, int len) +const void * +term_fontq(struct termp *p) { - int j; - if (++(*i) >= len) - return; + return(&p->fontq[p->fonti]); +} - if ('(' == word[*i]) { - (*i)++; - if (*i + 1 >= len) - return; - term_nescape(p, &word[*i], 2); - (*i)++; - return; +enum termfont +term_fonttop(struct termp *p) +{ - } else if ('*' == word[*i]) { - (*i)++; - if (*i >= len) - return; + return(p->fontq[p->fonti]); +} - switch (word[*i]) { - case ('('): - (*i)++; - if (*i + 1 >= len) - return; - term_nescape(p, &word[*i], 2); - (*i)++; - return; - case ('['): - break; - default: - term_nescape(p, &word[*i], 1); - return; - } - - } else if ('f' == word[*i]) { - (*i)++; - if (*i >= len) - return; - switch (word[*i]) { - case ('B'): - p->flags |= TERMP_BOLD; - break; - case ('I'): - p->flags |= TERMP_UNDER; - break; - case ('P'): - /* FALLTHROUGH */ - case ('R'): - p->flags &= ~TERMP_STYLE; - break; - default: - break; - } - return; +void +term_fontpopq(struct termp *p, const void *key) +{ - } else if ('[' != word[*i]) { - term_nescape(p, &word[*i], 1); - return; - } + while (p->fonti >= 0 && key != &p->fontq[p->fonti]) + p->fonti--; + assert(p->fonti >= 0); +} - (*i)++; - for (j = 0; word[*i] && ']' != word[*i]; (*i)++, j++) - /* Loop... */ ; - if (0 == word[*i]) - return; +void +term_fontpop(struct termp *p) +{ - term_nescape(p, &word[*i - j], (size_t)j); + assert(p->fonti); + p->fonti--; } @@ -497,99 +413,249 @@ term_pescape(struct termp *p, const char *word, int *i * phrase that cannot be broken down (such as a literal string). This * handles word styling. */ -static void -term_pword(struct termp *p, const char *word, int len) +void +term_word(struct termp *p, const char *word) { - int i; + const char *sv, *seq; + int sz; + size_t ssz; + enum roffdeco deco; - if (term_isclosedelim(word, len)) - if ( ! (TERMP_IGNDELIM & p->flags)) - p->flags |= TERMP_NOSPACE; + sv = word; + 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)) - term_chara(p, ' '); + bufferc(p, ' '); if ( ! (p->flags & TERMP_NONOSPACE)) p->flags &= ~TERMP_NOSPACE; - /* - * If ANSI (word-length styling), then apply our style now, - * before the word. - */ + /* FIXME: use strcspn. */ - for (i = 0; i < len; i++) { - if ('\\' == word[i]) { - term_pescape(p, word, &i, len); + while (*word) { + if ('\\' != *word) { + encode(p, word, 1); + word++; continue; } - if (TERMP_STYLE & p->flags) { - if (TERMP_BOLD & p->flags) { - term_chara(p, word[i]); - term_chara(p, 8); - } - if (TERMP_UNDER & p->flags) { - term_chara(p, '_'); - term_chara(p, 8); - } + seq = ++word; + sz = a2roffdeco(&deco, &seq, &ssz); + + switch (deco) { + case (DECO_RESERVED): + res(p, seq, ssz); + break; + case (DECO_SPECIAL): + spec(p, seq, ssz); + break; + case (DECO_BOLD): + term_fontrepl(p, TERMFONT_BOLD); + break; + case (DECO_ITALIC): + term_fontrepl(p, TERMFONT_UNDER); + break; + case (DECO_ROMAN): + term_fontrepl(p, TERMFONT_NONE); + break; + case (DECO_PREVIOUS): + term_fontlast(p); + break; + default: + break; } - term_chara(p, word[i]); + word += sz; + if (DECO_NOSPACE == deco && '\0' == *word) + p->flags |= TERMP_NOSPACE; } - if (term_isopendelim(word, len)) - p->flags |= TERMP_NOSPACE; + if (sv[0] && 0 == sv[1]) + switch (sv[0]) { + case('('): + /* FALLTHROUGH */ + case('['): + /* FALLTHROUGH */ + case('{'): + p->flags |= TERMP_NOSPACE; + break; + default: + break; + } } -/* - * Like term_chara() but for arbitrary-length buffers. Resize the - * buffer by a factor of two (if the buffer is less than that) or the - * buffer's size. - */ static void -term_stringa(struct termp *p, const char *c, size_t sz) +adjbuf(struct termp *p, size_t sz) { - size_t s; - if (0 == sz) - return; + if (0 == p->maxcols) + p->maxcols = 1024; + while (sz >= p->maxcols) + p->maxcols <<= 2; - assert(c); - if (p->col + sz >= p->maxcols) { - if (0 == p->maxcols) - p->maxcols = 256; - s = sz > p->maxcols * 2 ? sz : p->maxcols * 2; - p->buf = realloc(p->buf, s); - if (NULL == p->buf) - err(1, "realloc"); - p->maxcols = s; + p->buf = realloc(p->buf, p->maxcols); + if (NULL == p->buf) { + perror(NULL); + exit(EXIT_FAILURE); } +} - (void)memcpy(&p->buf[(int)p->col], c, sz); + +static void +buffera(struct termp *p, const char *word, size_t sz) +{ + + if (p->col + sz >= p->maxcols) + adjbuf(p, p->col + sz); + + memcpy(&p->buf[(int)p->col], word, sz); p->col += sz; } -/* - * 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 -term_chara(struct termp *p, char c) +bufferc(struct termp *p, char c) { - 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"); - p->maxcols = s; + if (p->col + 1 >= p->maxcols) + adjbuf(p, p->col + 1); + + p->buf[(int)p->col++] = c; +} + + +static void +encode(struct termp *p, const char *word, size_t sz) +{ + enum termfont f; + int i; + + /* + * 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))) { + buffera(p, word, sz); + return; } - p->buf[(int)(p->col)++] = c; + + for (i = 0; i < (int)sz; i++) { + if ( ! isgraph((u_char)word[i])) { + bufferc(p, word[i]); + continue; + } + + if (TERMFONT_UNDER == f) + bufferc(p, '_'); + else + bufferc(p, word[i]); + + bufferc(p, 8); + bufferc(p, word[i]); + } } + + +size_t +term_vspan(const struct roffsu *su) +{ + double r; + + switch (su->unit) { + case (SCALE_CM): + r = su->scale * 2; + break; + case (SCALE_IN): + r = su->scale * 6; + break; + case (SCALE_PC): + r = su->scale; + break; + case (SCALE_PT): + r = su->scale / 8; + break; + case (SCALE_MM): + r = su->scale / 1000; + break; + case (SCALE_VS): + r = su->scale; + break; + default: + r = su->scale - 1; + break; + } + + if (r < 0.0) + r = 0.0; + return(/* LINTED */(size_t) + r); +} + + +size_t +term_hspan(const struct roffsu *su) +{ + double r; + + /* 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); +} +