=================================================================== RCS file: /cvs/mandoc/term.c,v retrieving revision 1.4 retrieving revision 1.227 diff -u -p -r1.4 -r1.227 --- mandoc/term.c 2009/02/21 15:34:46 1.4 +++ mandoc/term.c 2014/08/10 23:54:41 1.227 @@ -1,1012 +1,806 @@ -/* $Id: term.c,v 1.4 2009/02/21 15:34:46 kristaps Exp $ */ +/* $Id: term.c,v 1.227 2014/08/10 23:54:41 schwarze Exp $ */ /* - * Copyright (c) 2008 Kristaps Dzonsons + * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons + * 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 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. */ +#include "config.h" + +#include + #include #include -#include -#include -#include #include +#include #include -#include -#include -#include "private.h" +#include "mandoc.h" +#include "mandoc_aux.h" +#include "out.h" +#include "term.h" +#include "main.h" -enum termesc { - ESC_CLEAR, - ESC_BOLD, - ESC_UNDERLINE -}; +static size_t cond_width(const struct termp *, int, 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); -struct termp { - size_t rmargin; - size_t maxrmargin; - size_t maxcols; - size_t offset; - size_t col; - int flags; -#define TERMP_BOLD (1 << 0) /* Embolden words. */ -#define TERMP_UNDERLINE (1 << 1) /* Underline words. */ -#define TERMP_NOSPACE (1 << 2) /* No space before words. */ -#define TERMP_NOLPAD (1 << 3) /* No left-padding. */ -#define TERMP_NOBREAK (1 << 4) /* No break after line */ - char *buf; -}; -struct termact { - int (*pre)(struct termp *, - const struct mdoc_meta *, - const struct mdoc_node *); - int (*post)(struct termp *, - const struct mdoc_meta *, - const struct mdoc_node *); -}; +void +term_free(struct termp *p) +{ -static void termprint_r(struct termp *, - const struct mdoc_meta *, - const struct mdoc_node *); -static void termprint_header(struct termp *, - const struct mdoc_meta *); -static void termprint_footer(struct termp *, - const struct mdoc_meta *); + if (p->buf) + free(p->buf); + if (p->symtab) + mchars_free(p->symtab); -static int arg_hasattr(int, size_t, - const struct mdoc_arg *); -static int arg_getattr(int, size_t, - const struct mdoc_arg *); + free(p); +} -static void newln(struct termp *); -static void vspace(struct termp *); -static void pword(struct termp *, const char *, size_t); -static void word(struct termp *, const char *); - -#define decl_prepost(name, suffix) \ -static int name##_##suffix(struct termp *, \ - const struct mdoc_meta *, \ - const struct mdoc_node *) - -#define decl_pre(name) decl_prepost(name, pre) -#define decl_post(name) decl_prepost(name, post) - -decl_pre(termp_fl); -decl_pre(termp_it); -decl_pre(termp_nd); -decl_pre(termp_ns); -decl_pre(termp_op); -decl_pre(termp_pp); -decl_pre(termp_sh); -decl_pre(termp_xr); - -decl_post(termp_bl); -decl_post(termp_it); -decl_post(termp_op); -decl_post(termp_sh); - -decl_pre(termp_bold); -decl_pre(termp_under); - -decl_post(termp_bold); -decl_post(termp_under); - -const struct termact termacts[MDOC_MAX] = { - { NULL, NULL }, /* \" */ - { NULL, NULL }, /* Dd */ - { NULL, NULL }, /* Dt */ - { NULL, NULL }, /* Os */ - { termp_sh_pre, termp_sh_post }, /* Sh */ - { NULL, NULL }, /* Ss */ - { termp_pp_pre, NULL }, /* Pp */ - { NULL, NULL }, /* D1 */ - { NULL, NULL }, /* Dl */ - { NULL, NULL }, /* Bd */ - { NULL, NULL }, /* Ed */ - { NULL, termp_bl_post }, /* Bl */ - { NULL, NULL }, /* El */ - { termp_it_pre, termp_it_post }, /* It */ - { NULL, NULL }, /* Ad */ - { NULL, NULL }, /* An */ - { termp_under_pre, termp_under_post }, /* Ar */ - { NULL, NULL }, /* Cd */ - { NULL, NULL }, /* Cm */ - { NULL, NULL }, /* Dv */ - { NULL, NULL }, /* Er */ - { NULL, NULL }, /* Ev */ - { NULL, NULL }, /* Ex */ - { NULL, NULL }, /* Fa */ - { NULL, NULL }, /* Fd */ - { termp_fl_pre, termp_bold_post }, /* Fl */ - { NULL, NULL }, /* Fn */ - { NULL, NULL }, /* Ft */ - { NULL, NULL }, /* Ic */ - { NULL, NULL }, /* In */ - { NULL, NULL }, /* Li */ - { termp_nd_pre, NULL }, /* Nd */ - { termp_bold_pre, termp_bold_post }, /* Nm */ - { termp_op_pre, termp_op_post }, /* Op */ - { NULL, NULL }, /* Ot */ - { NULL, NULL }, /* Pa */ - { NULL, NULL }, /* Rv */ - { NULL, NULL }, /* St */ - { NULL, NULL }, /* Va */ - { NULL, NULL }, /* Vt */ - { termp_xr_pre, NULL }, /* Xr */ - { NULL, NULL }, /* %A */ - { NULL, NULL }, /* %B */ - { NULL, NULL }, /* %D */ - { NULL, NULL }, /* %I */ - { NULL, NULL }, /* %J */ - { NULL, NULL }, /* %N */ - { NULL, NULL }, /* %O */ - { NULL, NULL }, /* %P */ - { NULL, NULL }, /* %R */ - { NULL, NULL }, /* %T */ - { NULL, NULL }, /* %V */ - { NULL, NULL }, /* Ac */ - { NULL, NULL }, /* Ao */ - { NULL, NULL }, /* Aq */ - { NULL, NULL }, /* At */ - { NULL, NULL }, /* Bc */ - { NULL, NULL }, /* Bf */ - { NULL, NULL }, /* Bo */ - { NULL, NULL }, /* Bq */ - { NULL, NULL }, /* Bsx */ - { NULL, NULL }, /* Bx */ - { NULL, NULL }, /* Db */ - { NULL, NULL }, /* Dc */ - { NULL, NULL }, /* Do */ - { NULL, NULL }, /* Dq */ - { NULL, NULL }, /* Ec */ - { NULL, NULL }, /* Ef */ - { NULL, NULL }, /* Em */ - { NULL, NULL }, /* Eo */ - { NULL, NULL }, /* Fx */ - { NULL, NULL }, /* Ms */ - { NULL, NULL }, /* No */ - { termp_ns_pre, NULL }, /* Ns */ - { NULL, NULL }, /* Nx */ - { NULL, NULL }, /* Ox */ - { NULL, NULL }, /* Pc */ - { NULL, NULL }, /* Pf */ - { NULL, NULL }, /* Po */ - { NULL, NULL }, /* Pq */ - { NULL, NULL }, /* Qc */ - { NULL, NULL }, /* Ql */ - { NULL, NULL }, /* Qo */ - { NULL, NULL }, /* Qq */ - { NULL, NULL }, /* Re */ - { NULL, NULL }, /* Rs */ - { NULL, NULL }, /* Sc */ - { NULL, NULL }, /* So */ - { NULL, NULL }, /* Sq */ - { NULL, NULL }, /* Sm */ - { NULL, NULL }, /* Sx */ - { NULL, NULL }, /* Sy */ - { NULL, NULL }, /* Tn */ - { NULL, NULL }, /* Ux */ - { NULL, NULL }, /* Xc */ - { NULL, NULL }, /* Xo */ - { NULL, NULL }, /* Fo */ - { NULL, NULL }, /* Fc */ - { NULL, NULL }, /* Oo */ - { NULL, NULL }, /* Oc */ - { NULL, NULL }, /* Bk */ - { NULL, NULL }, /* Ek */ - { NULL, NULL }, /* Bt */ - { NULL, NULL }, /* Hf */ - { NULL, NULL }, /* Fr */ - { NULL, NULL }, /* Ud */ -}; - - -static int -arg_hasattr(int arg, size_t argc, const struct mdoc_arg *argv) +void +term_begin(struct termp *p, term_margin head, + term_margin foot, const void *arg) { - return(-1 != arg_getattr(arg, argc, argv)); + p->headf = head; + p->footf = foot; + p->argf = arg; + (*p->begin)(p); } - -static int -arg_getattr(int arg, size_t argc, const struct mdoc_arg *argv) +void +term_end(struct termp *p) { - int i; - for (i = 0; i < (int)argc; i++) - if (argv[i].arg == arg) - return(i); - return(-1); + (*p->end)(p); } - -static void -flushln(struct termp *p) +/* + * 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: 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) { - size_t i, j, vsz, vis, maxvis; + 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 */ + 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 */ /* * 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. + * + * 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; + 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 = p->rmargin - p->offset; - vis = 0; + bp = TERMP_NOBREAK & p->flags ? mmax : maxvis; /* - * If in the standard case (left-justified), then begin with our - * indentation, otherwise (columns, etc.) just start spitting - * out text. + * Calculate the required amount of padding. */ + vbl = p->offset + p->overstep > p->viscol ? + p->offset + p->overstep - p->viscol : 0; - if ( ! (p->flags & TERMP_NOLPAD)) - /* LINTED */ - for (j = 0; j < p->offset; j++) - putchar(' '); + vis = vend = 0; + i = 0; - for (i = 0; i < p->col; i++) { + while (i < p->col) { /* + * Handle literal tab characters: collapse all + * subsequent tabs into a single huge set of spaces. + */ + ntab = 0; + while (i < p->col && '\t' == p->buf[i]) { + vend = (vis / p->tabwidth + 1) * p->tabwidth; + vbl += vend - vis; + vis = vend; + ntab++; + i++; + } + + /* * Count up visible word characters. Control sequences - * (starting with the CSI) aren't counted. + * (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). */ - assert( ! isspace(p->buf[i])); - /* LINTED */ - for (j = i, vsz = 0; j < p->col; j++) { - if (isspace(p->buf[j])) + for (j = i, jhy = 0; j < p->col; j++) { + if (' ' == p->buf[j] || '\t' == p->buf[j]) break; - else if (27 == p->buf[j]) { - assert(j + 4 <= p->col); - j += 3; - } 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] || + 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]); } - assert(vsz > 0); /* - * If a word is too long and we're within a line, put it - * on the next line. Puke if we're being asked to write - * something that will exceed the right margin (i.e., - * from a fresh line or when we're not allowed to break - * the line with TERMP_NOBREAK). + * Find out whether we would exceed the right margin. + * If so, break to the next line. */ + if (vend > bp && 0 == jhy && vis > 0) { + vend -= vis; + (*p->endline)(p); + p->viscol = 0; + if (TERMP_BRIND & p->flags) { + vbl = p->rmargin; + vend += p->rmargin - p->offset; + } else + vbl = p->offset; - if (vis && vis + vsz >= maxvis) { - /* FIXME */ - if (p->flags & TERMP_NOBREAK) - errx(1, "word breaks right margin"); - putchar('\n'); - for (j = 0; j < p->offset; j++) - putchar(' '); - vis = 0; - } else if (vis + vsz >= maxvis) { - /* FIXME */ - errx(1, "word breaks right margin"); - } + /* use pending tabs on the new line */ - /* - * Write out the word and a trailing space. Omit the - * space if we're the last word in the line. - */ + if (0 < ntab) + vbl += ntab * p->tabwidth; + /* + * 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; + } + + /* Write out the [remaining] word. */ for ( ; i < p->col; i++) { - if (isspace(p->buf[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 = (i - j) * (*p->width)(p, ' '); + vbl += dv; + vend += dv; + break; + } + if (ASCII_NBRSP == p->buf[i]) { + vbl += (*p->width)(p, ' '); + continue; + } + if (ASCII_BREAK == p->buf[i]) + 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; + } + + (*p->letter)(p, p->buf[i]); + if (8 == p->buf[i]) + p->viscol -= (*p->width)(p, p->buf[i-1]); + else + p->viscol += (*p->width)(p, p->buf[i]); } - vis += vsz; - if (i < p->col) { - putchar(' '); - vis++; - } + vis = vend; } /* - * If we're not to right-marginalise it (newline), then instead - * pad to the right margin and stay off. + * If there was trailing white space, it was not printed; + * so reset the cursor position accordingly. */ + if (vis) + vis -= vbl; - if (p->flags & TERMP_NOBREAK) { - for ( ; vis <= maxvis; vis++) - putchar(' '); - } else - putchar('\n'); - p->col = 0; -} + p->overstep = 0; + if ( ! (TERMP_NOBREAK & p->flags)) { + p->viscol = 0; + (*p->endline)(p); + return; + } -static void -newln(struct termp *p) -{ + if (TERMP_HANG & p->flags) { + p->overstep = (int)(vis - maxvis + + p->trailspace * (*p->width)(p, ' ')); - /* - * A newline only breaks an existing line; it won't assert - * vertical space. - */ - p->flags |= TERMP_NOSPACE; - if (0 == p->col) + /* + * 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->trailspace && p->overstep < 0) + p->overstep = 0; return; - flushln(p); -} + } else if (TERMP_DANGLE & p->flags) + return; -static void -vspace(struct termp *p) + /* If the column was overrun, break the line. */ + 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. + */ +void +term_newln(struct termp *p) { - /* - * Asserts a vertical space (a full, empty line-break between - * lines). - */ - newln(p); - putchar('\n'); + p->flags |= TERMP_NOSPACE; + if (p->col || p->viscol) + term_flushln(p); } - -static void -chara(struct termp *p, char c) +/* + * 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. + * All data in the output buffer is flushed prior to the newline + * assertion. + */ +void +term_vspace(struct termp *p) { - /* TODO: dynamically expand the buffer. */ - if (p->col + 1 >= p->maxcols) - errx(1, "line overrun"); - p->buf[(p->col)++] = c; + term_newln(p); + p->viscol = 0; + if (0 < p->skipvsp) + p->skipvsp--; + else + (*p->endline)(p); } - -static void -escape(struct termp *p, enum termesc esc) +void +term_fontlast(struct termp *p) { + enum termfont f; - if (p->col + 4 >= p->maxcols) - errx(1, "line overrun"); - - p->buf[(p->col)++] = 27; - p->buf[(p->col)++] = '['; - switch (esc) { - case (ESC_CLEAR): - p->buf[(p->col)++] = '0'; - break; - case (ESC_BOLD): - p->buf[(p->col)++] = '1'; - break; - case (ESC_UNDERLINE): - p->buf[(p->col)++] = '4'; - break; - default: - abort(); - /* NOTREACHED */ - } - p->buf[(p->col)++] = 'm'; + f = p->fontl; + p->fontl = p->fontq[p->fonti]; + p->fontq[p->fonti] = f; } - -static void -pword(struct termp *p, const char *word, size_t len) +void +term_fontrepl(struct termp *p, enum termfont f) { - size_t i; - assert(len > 0); + p->fontl = p->fontq[p->fonti]; + p->fontq[p->fonti] = f; +} - if ( ! (p->flags & TERMP_NOSPACE)) - chara(p, ' '); +void +term_fontpush(struct termp *p, enum termfont f) +{ - p->flags &= ~TERMP_NOSPACE; + assert(p->fonti + 1 < 10); + p->fontl = p->fontq[p->fonti]; + p->fontq[++p->fonti] = f; +} - if (p->flags & TERMP_BOLD) - escape(p, ESC_BOLD); - if (p->flags & TERMP_UNDERLINE) - escape(p, ESC_UNDERLINE); +const void * +term_fontq(struct termp *p) +{ - /* TODO: escape patterns. */ + return(&p->fontq[p->fonti]); +} - for (i = 0; i < len; i++) - chara(p, word[i]); +enum termfont +term_fonttop(struct termp *p) +{ - if (p->flags & TERMP_BOLD || - p->flags & TERMP_UNDERLINE) - escape(p, ESC_CLEAR); + return(p->fontq[p->fonti]); } - -static void -word(struct termp *p, const char *word) +void +term_fontpopq(struct termp *p, const void *key) { - size_t i, j, len; - if (mdoc_isdelim(word)) - p->flags |= TERMP_NOSPACE; - - len = strlen(word); - assert(len > 0); - - /* LINTED */ - for (j = i = 0; i < len; i++) { - if ( ! isspace(word[i])) { - j++; - continue; - } - if (0 == j) - continue; - assert(i >= j); - pword(p, &word[i - j], j); - j = 0; - } - if (j > 0) { - assert(i >= j); - pword(p, &word[i - j], j); - } + while (p->fonti >= 0 && key < (void *)(p->fontq + p->fonti)) + p->fonti--; + assert(p->fonti >= 0); } - -/* ARGSUSED */ -static int -termp_it_post(struct termp *p, const struct mdoc_meta *meta, - const struct mdoc_node *node) +void +term_fontpop(struct termp *p) { - const struct mdoc_node *n, *it; - const struct mdoc_block *bl; - int i; - size_t width; - switch (node->type) { - case (MDOC_BODY): - /* FALLTHROUGH */ - case (MDOC_HEAD): - break; - default: - return(1); - } - - it = node->parent; - assert(MDOC_BLOCK == it->type); - assert(MDOC_It == it->tok); - - n = it->parent; - assert(MDOC_BODY == n->type); - assert(MDOC_Bl == n->tok); - n = n->parent; - bl = &n->data.block; - - /* If `-tag', adjust our margins accordingly. */ - - if (arg_hasattr(MDOC_Tag, bl->argc, bl->argv)) { - i = arg_getattr(MDOC_Width, bl->argc, bl->argv); - assert(i >= 0); - assert(1 == bl->argv[i].sz); - width = strlen(*bl->argv[i].value); /* XXX */ - - if (MDOC_HEAD == node->type) { - flushln(p); - /* FIXME: nested lists. */ - p->rmargin = p->maxrmargin; - p->flags &= ~TERMP_NOBREAK; - } else { - flushln(p); - p->offset -= width + 1; - p->flags &= ~TERMP_NOLPAD; - } - } - - return(1); + assert(p->fonti); + p->fonti--; } - -/* ARGSUSED */ -static int -termp_it_pre(struct termp *p, const struct mdoc_meta *meta, - const struct mdoc_node *node) +/* + * 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 + * handles word styling. + */ +void +term_word(struct termp *p, const char *word) { - const struct mdoc_node *n, *it; - const struct mdoc_block *bl; - int i; - size_t width; + const char nbrsp[2] = { ASCII_NBRSP, 0 }; + const char *seq, *cp; + char c; + int sz, uc; + size_t ssz; + enum mandoc_esc esc; - switch (node->type) { - case (MDOC_BODY): - /* FALLTHROUGH */ - case (MDOC_HEAD): - it = node->parent; - break; - case (MDOC_BLOCK): - it = node; - break; - default: - return(1); + if ( ! (TERMP_NOSPACE & p->flags)) { + if ( ! (TERMP_KEEP & p->flags)) { + bufferc(p, ' '); + if (TERMP_SENTENCE & p->flags) + bufferc(p, ' '); + } else + bufferc(p, ASCII_NBRSP); } + if (TERMP_PREKEEP & p->flags) + p->flags |= TERMP_KEEP; - assert(MDOC_BLOCK == it->type); - assert(MDOC_It == it->tok); + if ( ! (p->flags & TERMP_NONOSPACE)) + p->flags &= ~TERMP_NOSPACE; + else + p->flags |= TERMP_NOSPACE; - n = it->parent; - assert(MDOC_BODY == n->type); - assert(MDOC_Bl == n->tok); - n = n->parent; - bl = &n->data.block; + p->flags &= ~TERMP_SENTENCE; - /* If `-compact', don't assert vertical space. */ + while ('\0' != *word) { + if ('\\' != *word) { + if (TERMP_SKIPCHAR & p->flags) { + p->flags &= ~TERMP_SKIPCHAR; + word++; + continue; + } + 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; + } - if (MDOC_BLOCK == node->type) { - if (arg_hasattr(MDOC_Compact, bl->argc, bl->argv)) - newln(p); - else - vspace(p); - return(1); - } + word++; + esc = mandoc_escape(&word, &seq, &sz); + if (ESCAPE_ERROR == esc) + continue; - assert(MDOC_HEAD == node->type - || MDOC_BODY == node->type); + 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; + } - /* If `-tag', adjust our margins accordingly. */ - - if (arg_hasattr(MDOC_Tag, bl->argc, bl->argv)) { - i = arg_getattr(MDOC_Width, bl->argc, bl->argv); - assert(i >= 0); /* XXX */ - assert(1 == bl->argv[i].sz); - width = strlen(*bl->argv[i].value); /* XXX */ - - /* FIXME: nested lists. */ - - if (MDOC_HEAD == node->type) { - p->flags |= TERMP_NOBREAK; - p->flags |= TERMP_NOSPACE; - p->rmargin = p->offset + width; - } else { - p->flags |= TERMP_NOSPACE; - p->flags |= TERMP_NOLPAD; - p->offset += width + 1; + switch (esc) { + case ESCAPE_UNICODE: + encode1(p, '?'); + break; + case ESCAPE_NUMBERED: + c = mchars_num2char(seq, sz); + if ('\0' != c) + encode(p, &c, 1); + 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_FONTBI: + term_fontrepl(p, TERMFONT_BI); + break; + case ESCAPE_FONT: + /* FALLTHROUGH */ + case ESCAPE_FONTROMAN: + term_fontrepl(p, TERMFONT_NONE); + break; + case ESCAPE_FONTPREV: + term_fontlast(p); + break; + 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_SKIPCHAR; + break; + default: + break; } } - - return(1); + p->flags &= ~TERMP_NBRWORD; } - -/* ARGSUSED */ -static int -termp_bold_post(struct termp *p, const struct mdoc_meta *meta, - const struct mdoc_node *node) +static void +adjbuf(struct termp *p, size_t sz) { - p->flags &= ~TERMP_BOLD; - return(1); -} + if (0 == p->maxcols) + p->maxcols = 1024; + while (sz >= p->maxcols) + p->maxcols <<= 2; - -/* ARGSUSED */ -static int -termp_under_pre(struct termp *p, const struct mdoc_meta *meta, - const struct mdoc_node *node) -{ - - p->flags |= TERMP_UNDERLINE; - return(1); + p->buf = mandoc_reallocarray(p->buf, p->maxcols, sizeof(int)); } - -/* ARGSUSED */ -static int -termp_bold_pre(struct termp *p, const struct mdoc_meta *meta, - const struct mdoc_node *node) +static void +bufferc(struct termp *p, char c) { - p->flags |= TERMP_BOLD; - return(1); -} + if (p->col + 1 >= p->maxcols) + adjbuf(p, p->col + 1); - -/* ARGSUSED */ -static int -termp_ns_pre(struct termp *p, const struct mdoc_meta *meta, - const struct mdoc_node *node) -{ - - p->flags |= TERMP_NOSPACE; - return(1); + p->buf[p->col++] = c; } - -/* ARGSUSED */ -static int -termp_pp_pre(struct termp *p, const struct mdoc_meta *meta, - const struct mdoc_node *node) +/* + * See encode(). + * Do this for a single (probably unicode) value. + * Does not check for non-decorated glyphs. + */ +static void +encode1(struct termp *p, int c) { + enum termfont f; - vspace(p); - return(1); -} + if (TERMP_SKIPCHAR & p->flags) { + p->flags &= ~TERMP_SKIPCHAR; + return; + } + if (p->col + 6 >= p->maxcols) + adjbuf(p, p->col + 6); -/* ARGSUSED */ -static int -termp_under_post(struct termp *p, const struct mdoc_meta *meta, - const struct mdoc_node *node) -{ + f = term_fonttop(p); - p->flags &= ~TERMP_UNDERLINE; - return(1); + if (TERMFONT_UNDER == f || TERMFONT_BI == f) { + p->buf[p->col++] = '_'; + 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; } - -/* ARGSUSED */ -static int -termp_nd_pre(struct termp *p, const struct mdoc_meta *meta, - const struct mdoc_node *node) +static void +encode(struct termp *p, const char *word, size_t sz) { + size_t i; - word(p, "-"); - return(1); -} + if (TERMP_SKIPCHAR & p->flags) { + p->flags &= ~TERMP_SKIPCHAR; + return; + } + /* + * Encode and buffer a string of characters. If the current + * font mode is unset, buffer directly, else encode then buffer + * character by character. + */ -/* ARGSUSED */ -static int -termp_bl_post(struct termp *p, const struct mdoc_meta *meta, - const struct mdoc_node *node) -{ - - switch (node->type) { - case (MDOC_BLOCK): - newln(p); - break; - default: - break; + 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; } - return(1); -} + /* Pre-buffer, assuming worst-case. */ -/* ARGSUSED */ -static int -termp_op_post(struct termp *p, const struct mdoc_meta *meta, - const struct mdoc_node *node) -{ + if (p->col + 1 + (sz * 5) >= p->maxcols) + adjbuf(p, p->col + 1 + (sz * 5)); - switch (node->type) { - case (MDOC_BODY): - p->flags |= TERMP_NOSPACE; - word(p, "\\]"); - break; - default: - break; + 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]; } - return(1); } - -/* ARGSUSED */ -static int -termp_sh_post(struct termp *p, const struct mdoc_meta *meta, - const struct mdoc_node *node) +void +term_setwidth(struct termp *p, const char *wstr) { + struct roffsu su; + size_t width; + int iop; - switch (node->type) { - case (MDOC_HEAD): - p->flags &= ~TERMP_BOLD; - newln(p); - break; - case (MDOC_BODY): - newln(p); - p->offset -= 4; - break; - default: - break; + 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; } - return(1); + (*p->setwidth)(p, iop, width); } - -/* ARGSUSED */ -static int -termp_xr_pre(struct termp *p, const struct mdoc_meta *meta, - const struct mdoc_node *node) +size_t +term_len(const struct termp *p, size_t sz) { - const struct mdoc_node *n; - n = node->child; - assert(n); - - assert(MDOC_TEXT == n->type); - word(p, n->data.text.string); - - if (NULL == (n = n->next)) - return(0); - - assert(MDOC_TEXT == n->type); - p->flags |= TERMP_NOSPACE; - word(p, "\\("); - p->flags |= TERMP_NOSPACE; - word(p, n->data.text.string); - p->flags |= TERMP_NOSPACE; - word(p, "\\)"); - - return(0); + return((*p->width)(p, ' ') * sz); } - -/* ARGSUSED */ -static int -termp_sh_pre(struct termp *p, const struct mdoc_meta *meta, - const struct mdoc_node *node) +static size_t +cond_width(const struct termp *p, int c, int *skip) { - switch (node->type) { - case (MDOC_HEAD): - vspace(p); - p->flags |= TERMP_BOLD; - break; - case (MDOC_BODY): - p->offset += 4; - break; - default: - break; - } - return(1); + if (*skip) { + (*skip) = 0; + return(0); + } else + return((*p->width)(p, c)); } - -/* ARGSUSED */ -static int -termp_op_pre(struct termp *p, const struct mdoc_meta *meta, - const struct mdoc_node *node) +size_t +term_strlen(const struct termp *p, const char *cp) { + size_t sz, rsz, i; + int ssz, skip, c; + const char *seq, *rhs; + enum mandoc_esc esc; + static const char rej[] = { '\\', ASCII_NBRSP, ASCII_HYPH, + ASCII_BREAK, '\0' }; - switch (node->type) { - case (MDOC_BODY): - word(p, "\\["); - p->flags |= TERMP_NOSPACE; - break; - default: - break; - } - return(1); -} + /* + * 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; + skip = 0; + while ('\0' != *cp) { + rsz = strcspn(cp, rej); + for (i = 0; i < rsz; i++) + sz += cond_width(p, *cp++, &skip); -/* ARGSUSED */ -static int -termp_fl_pre(struct termp *p, const struct mdoc_meta *meta, - const struct mdoc_node *node) -{ + switch (*cp) { + case '\\': + cp++; + esc = mandoc_escape(&cp, &seq, &ssz); + if (ESCAPE_ERROR == esc) + continue; - p->flags |= TERMP_BOLD; - word(p, "-"); - p->flags |= TERMP_NOSPACE; - return(1); -} + 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; -static void -termprint_r(struct termp *p, const struct mdoc_meta *meta, - const struct mdoc_node *node) -{ - int dochild; + switch (esc) { + case ESCAPE_UNICODE: + sz += cond_width(p, '?', &skip); + break; + 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); - /* Pre-processing ----------------- */ + if (ssz != 1 || rhs) + break; - dochild = 1; - - if (MDOC_TEXT != node->type) { - if (termacts[node->tok].pre) - if ( ! (*termacts[node->tok].pre)(p, meta, node)) - dochild = 0; - } else /* MDOC_TEXT == node->type */ - word(p, node->data.text.string); - - /* Children ---------------------- */ - - if (dochild && NULL == node->child) { - /* No-child processing. */ - switch (node->type) { - case (MDOC_ELEM): - switch (node->tok) { - case (MDOC_Nm): - word(p, "progname"); /* TODO */ + rhs = seq; + rsz = ssz; break; - case (MDOC_Ar): - word(p, "..."); + case ESCAPE_SKIPCHAR: + skip = 1; break; default: break; } - break; - default: - break; - } - } else if (dochild) - termprint_r(p, meta, node->child); - /* Post-processing --------------- */ + if (NULL == rhs) + break; - if (MDOC_TEXT != node->type) { - if (termacts[node->tok].post) - if ( ! (*termacts[node->tok].post)(p, meta, node)) - return; - } + if (skip) { + skip = 0; + break; + } - /* Siblings ---------------------- */ - - if (node->next) - termprint_r(p, meta, node->next); -} - - -static void -termprint_footer(struct termp *p, const struct mdoc_meta *meta) -{ - struct tm *tm; - char *buf, *os; - size_t sz, osz, ssz, i; - - if (NULL == (buf = malloc(p->rmargin))) - err(1, "malloc"); - if (NULL == (os = malloc(p->rmargin))) - err(1, "malloc"); - - tm = localtime(&meta->date); - if (NULL == strftime(buf, p->rmargin, "%B %d, %Y", tm)) - err(1, "strftime"); - - osz = strlcpy(os, meta->os, p->rmargin); - - sz = strlen(buf); - ssz = sz + osz + 1; - - if (ssz > p->rmargin) { - ssz -= p->rmargin; - assert(ssz <= osz); - os[osz - ssz] = 0; - ssz = 1; - } else - ssz = p->rmargin - ssz + 1; - - printf("\n"); - printf("%s", os); - for (i = 0; i < ssz; i++) - printf(" "); - - printf("%s\n", buf); - fflush(stdout); - - free(buf); - free(os); -} - - -static void -termprint_header(struct termp *p, const struct mdoc_meta *meta) -{ - char *msec, *buf, *title, *pp; - size_t ssz, tsz, ttsz, i;; - - if (NULL == (buf = malloc(p->rmargin))) - err(1, "malloc"); - if (NULL == (title = malloc(p->rmargin))) - err(1, "malloc"); - - if (NULL == (pp = mdoc_vol2a(meta->vol))) - switch (meta->msec) { - case (MSEC_1): - /* FALLTHROUGH */ - case (MSEC_6): - /* FALLTHROUGH */ - case (MSEC_7): - pp = mdoc_vol2a(VOL_URM); + for (i = 0; i < rsz; i++) + sz += (*p->width)(p, *rhs++); break; - case (MSEC_8): - pp = mdoc_vol2a(VOL_SMM); + case ASCII_NBRSP: + sz += cond_width(p, ' ', &skip); + cp++; break; - case (MSEC_2): + case ASCII_HYPH: + sz += cond_width(p, '-', &skip); + cp++; /* FALLTHROUGH */ - case (MSEC_3): - /* FALLTHROUGH */ - case (MSEC_4): - /* FALLTHROUGH */ - case (MSEC_5): - pp = mdoc_vol2a(VOL_PRM); + case ASCII_BREAK: break; - case (MSEC_9): - pp = mdoc_vol2a(VOL_KM); - break; default: - /* FIXME: capitalise. */ - if (NULL == (pp = mdoc_msec2a(meta->msec))) - pp = mdoc_msec2a(MSEC_local); break; } - assert(pp); - - tsz = strlcpy(buf, pp, p->rmargin); - assert(tsz < p->rmargin); - - if ((pp = mdoc_arch2a(meta->arch))) { - tsz = strlcat(buf, " (", p->rmargin); - assert(tsz < p->rmargin); - tsz = strlcat(buf, pp, p->rmargin); - assert(tsz < p->rmargin); - tsz = strlcat(buf, ")", p->rmargin); - assert(tsz < p->rmargin); } - ttsz = strlcpy(title, meta->title, p->rmargin); - - if (NULL == (msec = mdoc_msec2a(meta->msec))) - msec = ""; - - ssz = (2 * (ttsz + 2 + strlen(msec))) + tsz + 2; - - if (ssz > p->rmargin) { - if ((ssz -= p->rmargin) % 2) - ssz++; - ssz /= 2; - - assert(ssz <= ttsz); - title[ttsz - ssz] = 0; - ssz = 1; - } else - ssz = ((p->rmargin - ssz) / 2) + 1; - - printf("%s(%s)", title, msec); - - for (i = 0; i < ssz; i++) - printf(" "); - - printf("%s", buf); - - for (i = 0; i < ssz; i++) - printf(" "); - - printf("%s(%s)\n", title, msec); - fflush(stdout); - - free(title); - free(buf); + return(sz); } - -int -termprint(const struct mdoc_node *node, - const struct mdoc_meta *meta) +size_t +term_vspan(const struct termp *p, const struct roffsu *su) { - struct termp p; + double r; - if (ERR == setupterm(NULL, STDOUT_FILENO, NULL)) - return(0); + switch (su->unit) { + case SCALE_CM: + r = su->scale * 2.0; + break; + case SCALE_IN: + r = su->scale * 6.0; + break; + case SCALE_PC: + r = su->scale; + break; + case SCALE_PT: + r = su->scale / 8.0; + break; + case SCALE_MM: + r = su->scale / 1000.0; + break; + case SCALE_VS: + r = su->scale; + break; + default: + r = su->scale - 1.0; + break; + } - p.maxrmargin = columns < 60 ? 60 : (size_t)columns; - p.rmargin = p.maxrmargin; - p.maxcols = 1024; - p.offset = p.col = 0; - p.flags = TERMP_NOSPACE; - - if (NULL == (p.buf = malloc(p.maxcols))) - err(1, "malloc"); - - termprint_header(&p, meta); - termprint_r(&p, meta, node); - termprint_footer(&p, meta); - - free(p.buf); - - return(1); + if (r < 0.0) + r = 0.0; + return((size_t)(r + 0.0005)); } +size_t +term_hspan(const struct termp *p, const struct roffsu *su) +{ + double v; + v = (*p->hspan)(p, su); + if (v < 0.0) + v = 0.0; + return((size_t)(v + 0.0005)); +}