=================================================================== RCS file: /cvs/mandoc/term.c,v retrieving revision 1.4 retrieving revision 1.144 diff -u -p -r1.4 -r1.144 --- mandoc/term.c 2009/02/21 15:34:46 1.4 +++ mandoc/term.c 2010/06/08 09:20:08 1.144 @@ -1,1012 +1,941 @@ -/* $Id: term.c,v 1.4 2009/02/21 15:34:46 kristaps Exp $ */ +/* $Id: term.c,v 1.144 2010/06/08 09:20:08 kristaps Exp $ */ /* - * Copyright (c) 2008 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 +#include #include -#include -#include +#include -#include "private.h" +#include "mandoc.h" +#include "chars.h" +#include "out.h" +#include "term.h" +#include "man.h" +#include "mdoc.h" +#include "main.h" -enum termesc { - ESC_CLEAR, - ESC_BOLD, - ESC_UNDERLINE -}; +#define PS_CHAR_WIDTH 6 +#define PS_CHAR_HEIGHT 12 +#define PS_CHAR_TOPMARG (792 - 24) +#define PS_CHAR_TOP (PS_CHAR_TOPMARG - 36) +#define PS_CHAR_LEFT 36 +#define PS_CHAR_BOTMARG 24 +#define PS_CHAR_BOT (PS_CHAR_BOTMARG + 36) -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; -}; +static struct termp *alloc(char *, enum termenc, enum termtype); +static void term_free(struct termp *); +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); +static void advance(struct termp *, size_t); +static void endline(struct termp *); +static void letter(struct termp *, char); +static void pageopen(struct termp *); -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 *); -}; -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 *); +void * +ascii_alloc(char *outopts) +{ -static int arg_hasattr(int, size_t, - const struct mdoc_arg *); -static int arg_getattr(int, size_t, - const struct mdoc_arg *); + return(alloc(outopts, TERMENC_ASCII, TERMTYPE_CHAR)); +} -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 *) +void * +ps_alloc(void) +{ -#define decl_pre(name) decl_prepost(name, pre) -#define decl_post(name) decl_prepost(name, post) + return(alloc(NULL, TERMENC_ASCII, TERMTYPE_PS)); +} -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 +terminal_free(void *arg) { - return(-1 != arg_getattr(arg, argc, argv)); + term_free((struct termp *)arg); } -static int -arg_getattr(int arg, size_t argc, const struct mdoc_arg *argv) +static void +term_free(struct termp *p) { - int i; - for (i = 0; i < (int)argc; i++) - if (argv[i].arg == arg) - return(i); - return(-1); + if (p->buf) + free(p->buf); + if (p->symtab) + chars_free(p->symtab); + free(p); } +/* + * Push a single letter into our output engine. + */ static void -flushln(struct termp *p) +letter(struct termp *p, char c) { - size_t i, j, vsz, vis, maxvis; + + if (TERMTYPE_CHAR == p->type) { + /* + * If using the terminal device, just push the letter + * out into the screen. + */ + putchar(c); + return; + } - /* - * 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. - */ + if ( ! (PS_INLINE & p->psstate)) { + /* + * If we're not in a PostScript "word" context, then + * open one now at the current cursor. + */ + printf("%zu %zu moveto\n", p->pscol, p->psrow); + putchar('('); + p->psstate |= PS_INLINE; + } - assert(p->offset < p->rmargin); - maxvis = p->rmargin - p->offset; - vis = 0; - /* - * If in the standard case (left-justified), then begin with our - * indentation, otherwise (columns, etc.) just start spitting - * out text. + * We need to escape these characters as per the PostScript + * specification. We would also escape non-graphable characters + * (like tabs), but none of them would get to this point and + * it's superfluous to abort() on them. */ - if ( ! (p->flags & TERMP_NOLPAD)) - /* LINTED */ - for (j = 0; j < p->offset; j++) - putchar(' '); + switch (c) { + case ('('): + /* FALLTHROUGH */ + case (')'): + /* FALLTHROUGH */ + case ('\\'): + putchar('\\'); + break; + default: + break; + } - for (i = 0; i < p->col; i++) { - /* - * Count up visible word characters. Control sequences - * (starting with the CSI) aren't counted. - */ - assert( ! isspace(p->buf[i])); + /* Write the character and adjust where we are on the page. */ + putchar(c); + p->pscol += PS_CHAR_WIDTH; +} - /* LINTED */ - for (j = i, vsz = 0; j < p->col; j++) { - if (isspace(p->buf[j])) - break; - else if (27 == p->buf[j]) { - assert(j + 4 <= p->col); - j += 3; - } else - vsz++; - } - 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). - */ +/* + * Begin a "terminal" context. Since terminal encompasses PostScript, + * the actual terminal, etc., there are a few things we can do here. + */ +void +term_begin(struct termp *p, term_margin head, + term_margin foot, const void *arg) +{ - 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"); - } + p->headf = head; + p->footf = foot; + p->argf = arg; - /* - * Write out the word and a trailing space. Omit the - * space if we're the last word in the line. - */ - - for ( ; i < p->col; i++) { - if (isspace(p->buf[i])) - break; - putchar(p->buf[i]); - } - vis += vsz; - if (i < p->col) { - putchar(' '); - vis++; - } + if (TERMTYPE_CHAR == p->type) { + /* Emit the header and be done. */ + (*p->headf)(p, p->argf); + return; } - + /* - * If we're not to right-marginalise it (newline), then instead - * pad to the right margin and stay off. + * Emit the standard PostScript prologue, set our initial page + * position, then run pageopen() on the initial page. */ - if (p->flags & TERMP_NOBREAK) { - for ( ; vis <= maxvis; vis++) - putchar(' '); - } else - putchar('\n'); + printf("%s\n", "%!PS"); + printf("%s\n", "/Courier"); + printf("%s\n", "10 selectfont"); - p->col = 0; + p->pspage = 1; + p->psstate = 0; + pageopen(p); } +/* + * Open a page. This is only used for -Tps at the moment. It opens a + * page context, printing the header and the footer. THE OUTPUT BUFFER + * MUST BE EMPTY. If it is not, output will ghost on the next line and + * we'll be all gross and out of state. + */ static void -newln(struct termp *p) +pageopen(struct termp *p) { + + assert(TERMTYPE_PS == p->type); + assert(0 == p->psstate); - /* - * A newline only breaks an existing line; it won't assert - * vertical space. - */ - p->flags |= TERMP_NOSPACE; - if (0 == p->col) - return; - flushln(p); -} + p->pscol = PS_CHAR_LEFT; + p->psrow = PS_CHAR_TOPMARG; + p->psstate |= PS_MARGINS; + (*p->headf)(p, p->argf); + endline(p); -static void -vspace(struct termp *p) -{ + p->psstate &= ~PS_MARGINS; + assert(0 == p->psstate); - /* - * Asserts a vertical space (a full, empty line-break between - * lines). - */ - newln(p); - putchar('\n'); -} + p->pscol = PS_CHAR_LEFT; + p->psrow = PS_CHAR_BOTMARG; + p->psstate |= PS_MARGINS; + (*p->footf)(p, p->argf); + endline(p); -static void -chara(struct termp *p, char c) -{ + p->psstate &= ~PS_MARGINS; + assert(0 == p->psstate); - /* TODO: dynamically expand the buffer. */ - if (p->col + 1 >= p->maxcols) - errx(1, "line overrun"); - p->buf[(p->col)++] = c; + p->pscol = PS_CHAR_LEFT; + p->psrow = PS_CHAR_TOP; + } -static void -escape(struct termp *p, enum termesc esc) +void +term_end(struct termp *p) { - 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 */ + if (TERMTYPE_CHAR == p->type) { + (*p->footf)(p, p->argf); + return; } - p->buf[(p->col)++] = 'm'; + + printf("%s\n", "%%END"); } static void -pword(struct termp *p, const char *word, size_t len) +endline(struct termp *p) { - size_t i; - assert(len > 0); + if (TERMTYPE_CHAR == p->type) { + putchar('\n'); + return; + } - if ( ! (p->flags & TERMP_NOSPACE)) - chara(p, ' '); + if (PS_INLINE & p->psstate) { + printf(") show\n"); + p->psstate &= ~PS_INLINE; + } - p->flags &= ~TERMP_NOSPACE; + if (PS_MARGINS & p->psstate) + return; - if (p->flags & TERMP_BOLD) - escape(p, ESC_BOLD); - if (p->flags & TERMP_UNDERLINE) - escape(p, ESC_UNDERLINE); + p->pscol = PS_CHAR_LEFT; + if (p->psrow >= PS_CHAR_HEIGHT + PS_CHAR_BOT) { + p->psrow -= PS_CHAR_HEIGHT; + return; + } - /* TODO: escape patterns. */ - - for (i = 0; i < len; i++) - chara(p, word[i]); - - if (p->flags & TERMP_BOLD || - p->flags & TERMP_UNDERLINE) - escape(p, ESC_CLEAR); + /* + * XXX: can't run pageopen() until we're certain a flushln() has + * occured, else the buf will reopen in an awkward state on the + * next line. + */ + printf("showpage\n"); + p->psrow = PS_CHAR_TOP; } +/* + * Advance the output engine by a certain amount of whitespace. + */ static void -word(struct termp *p, const char *word) +advance(struct termp *p, size_t len) { - size_t i, j, len; + size_t i; - if (mdoc_isdelim(word)) - p->flags |= TERMP_NOSPACE; + if (TERMTYPE_CHAR == p->type) { + /* Just print whitespace on the terminal. */ + for (i = 0; i < len; i++) + putchar(' '); + return; + } - 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 (PS_INLINE & p->psstate) { + /* Dump out any existing line scope. */ + printf(") show\n"); + p->psstate &= ~PS_INLINE; } - if (j > 0) { - assert(i >= j); - pword(p, &word[i - j], j); - } + + p->pscol += len ? len * PS_CHAR_WIDTH : 0; } -/* ARGSUSED */ -static int -termp_it_post(struct termp *p, const struct mdoc_meta *meta, - const struct mdoc_node *node) +static struct termp * +alloc(char *outopts, enum termenc enc, enum termtype type) { - const struct mdoc_node *n, *it; - const struct mdoc_block *bl; - int i; + struct termp *p; + const char *toks[2]; + char *v; size_t width; - switch (node->type) { - case (MDOC_BODY): - /* FALLTHROUGH */ - case (MDOC_HEAD): - break; - default: - return(1); + toks[0] = "width"; + toks[1] = NULL; + + p = calloc(1, sizeof(struct termp)); + if (NULL == p) { + perror(NULL); + exit(EXIT_FAILURE); } - it = node->parent; - assert(MDOC_BLOCK == it->type); - assert(MDOC_It == it->tok); + p->type = type; + p->tabwidth = 5; + p->enc = enc; - n = it->parent; - assert(MDOC_BODY == n->type); - assert(MDOC_Bl == n->tok); - n = n->parent; - bl = &n->data.block; + width = 80; - /* 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; + while (outopts && *outopts) + switch (getsubopt(&outopts, UNCONST(toks), &v)) { + case (0): + width = (size_t)atoi(v); + break; + default: + break; } - } - return(1); + /* Enforce some lower boundary. */ + if (width < 60) + width = 60; + p->defrmargin = width - 2; + return(p); } -/* ARGSUSED */ -static int -termp_it_pre(struct termp *p, const struct mdoc_meta *meta, - const struct mdoc_node *node) +/* + * 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. + * + * The following flags may be specified: + * + * - 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. + * + * - TERMP_NOBREAK: this is the most important and is used when making + * columns. In short: don't print a newline and instead pad to the + * right margin. Used in conjunction with TERMP_NOLPAD. + * + * - 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. + */ +void +term_flushln(struct termp *p) { - const struct mdoc_node *n, *it; - const struct mdoc_block *bl; - int i; - size_t width; + 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 vend; /* end of word visual position on output */ + size_t bp; /* visual right border position */ + int j; /* temporary loop index */ + int jhy; /* last hyphen before line overflow */ + size_t maxvis, mmax; - switch (node->type) { - case (MDOC_BODY): - /* FALLTHROUGH */ - case (MDOC_HEAD): - it = node->parent; - break; - case (MDOC_BLOCK): - it = node; - break; - default: - return(1); - } + /* + * 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. + */ - assert(MDOC_BLOCK == it->type); - assert(MDOC_It == it->tok); + assert(p->offset < p->rmargin); - n = it->parent; - assert(MDOC_BODY == n->type); - assert(MDOC_Bl == n->tok); - n = n->parent; - bl = &n->data.block; + 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; - /* If `-compact', don't assert vertical space. */ + bp = TERMP_NOBREAK & p->flags ? mmax : maxvis; - if (MDOC_BLOCK == node->type) { - if (arg_hasattr(MDOC_Compact, bl->argc, bl->argv)) - newln(p); - else - vspace(p); - return(1); - } + /* + * Indent the first line of a paragraph. + */ + vbl = p->flags & TERMP_NOLPAD ? 0 : p->offset; - assert(MDOC_HEAD == node->type - || MDOC_BODY == node->type); + /* + * FIXME: if bp is zero, we still output the first word before + * breaking the line. + */ - /* If `-tag', adjust our margins accordingly. */ + vis = vend = i = 0; + while (i < (int)p->col) { - 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 */ + /* + * Handle literal tab characters. + */ + for (j = i; j < (int)p->col; j++) { + if ('\t' != p->buf[j]) + break; + vend = (vis/p->tabwidth+1)*p->tabwidth; + vbl += vend - vis; + vis = vend; + } - /* FIXME: nested lists. */ + /* + * 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). + */ - 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; + /* LINTED */ + for (jhy = 0; j < (int)p->col; j++) { + if ((j && ' ' == p->buf[j]) || '\t' == p->buf[j]) + break; + if (8 != p->buf[j]) { + if (vend > vis && vend < bp && + ASCII_HYPH == p->buf[j]) + jhy = j; + vend++; + } else + vend--; } - } - return(1); -} + /* + * 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; + endline(p); + if (TERMP_NOBREAK & p->flags) { + p->viscol = p->rmargin; + advance(p, p->rmargin); + vend += p->rmargin - p->offset; + } else { + p->viscol = 0; + vbl = p->offset; + } + /* Remove the p->overstep width. */ -/* ARGSUSED */ -static int -termp_bold_post(struct termp *p, const struct mdoc_meta *meta, - const struct mdoc_node *node) -{ + bp += (int)/* LINTED */ + p->overstep; + p->overstep = 0; + } - p->flags &= ~TERMP_BOLD; - return(1); -} + /* + * Skip leading tabs, they were handled above. + */ + while (i < (int)p->col && '\t' == p->buf[i]) + i++; + /* Write out the [remaining] word. */ + for ( ; i < (int)p->col; i++) { + if (vend > bp && jhy > 0 && i > jhy) + break; + if ('\t' == p->buf[i]) + break; + if (' ' == p->buf[i]) { + while (' ' == p->buf[i]) { + vbl++; + i++; + } + break; + } + if (ASCII_NBRSP == p->buf[i]) { + vbl++; + continue; + } -/* ARGSUSED */ -static int -termp_under_pre(struct termp *p, const struct mdoc_meta *meta, - const struct mdoc_node *node) -{ + /* + * Now we definitely know there will be + * printable characters to output, + * so write preceding white space now. + */ + if (vbl) { + advance(p, vbl); + p->viscol += vbl; + vbl = 0; + } - p->flags |= TERMP_UNDERLINE; - return(1); -} + if (ASCII_HYPH == p->buf[i]) + letter(p, '-'); + else + letter(p, p->buf[i]); + p->viscol += 1; + } + vend += vbl; + vis = vend; + } -/* ARGSUSED */ -static int -termp_bold_pre(struct termp *p, const struct mdoc_meta *meta, - const struct mdoc_node *node) -{ + p->col = 0; + p->overstep = 0; - p->flags |= TERMP_BOLD; - return(1); + if ( ! (TERMP_NOBREAK & p->flags)) { + p->viscol = 0; + endline(p); + return; + } + + if (TERMP_HANG & p->flags) { + /* We need one blank after the tag. */ + p->overstep = /* LINTED */ + vis - maxvis + 1; + + /* + * 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; + + } else if (TERMP_DANGLE & p->flags) + return; + + /* Right-pad. */ + if (maxvis > vis + /* LINTED */ + ((TERMP_TWOSPACE & p->flags) ? 1 : 0)) { + p->viscol += maxvis - vis; + advance(p, maxvis - vis); + vis += (maxvis - vis); + } else { /* ...or newline break. */ + endline(p); + p->viscol = p->rmargin; + advance(p, p->rmargin); + } } -/* ARGSUSED */ -static int -termp_ns_pre(struct termp *p, const struct mdoc_meta *meta, - const struct mdoc_node *node) +/* + * 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) { p->flags |= TERMP_NOSPACE; - return(1); + if (0 == p->col && 0 == p->viscol) { + p->flags &= ~TERMP_NOLPAD; + return; + } + term_flushln(p); + p->flags &= ~TERMP_NOLPAD; } -/* ARGSUSED */ -static int -termp_pp_pre(struct termp *p, const struct mdoc_meta *meta, - const struct mdoc_node *node) +/* + * 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) { - vspace(p); - return(1); + term_newln(p); + p->viscol = 0; + endline(p); } -/* ARGSUSED */ -static int -termp_under_post(struct termp *p, const struct mdoc_meta *meta, - const struct mdoc_node *node) +static void +spec(struct termp *p, const char *word, size_t len) { + const char *rhs; + size_t sz; - p->flags &= ~TERMP_UNDERLINE; - return(1); + rhs = chars_a2ascii(p->symtab, word, len, &sz); + if (rhs) + encode(p, rhs, sz); } -/* ARGSUSED */ -static int -termp_nd_pre(struct termp *p, const struct mdoc_meta *meta, - const struct mdoc_node *node) +static void +res(struct termp *p, const char *word, size_t len) { + const char *rhs; + size_t sz; - word(p, "-"); - return(1); + rhs = chars_a2res(p->symtab, word, len, &sz); + if (rhs) + encode(p, rhs, sz); } -/* ARGSUSED */ -static int -termp_bl_post(struct termp *p, const struct mdoc_meta *meta, - const struct mdoc_node *node) +void +term_fontlast(struct termp *p) { + enum termfont f; - switch (node->type) { - case (MDOC_BLOCK): - newln(p); - break; - default: - break; - } - return(1); + f = p->fontl; + p->fontl = p->fontq[p->fonti]; + p->fontq[p->fonti] = f; } -/* ARGSUSED */ -static int -termp_op_post(struct termp *p, const struct mdoc_meta *meta, - const struct mdoc_node *node) +void +term_fontrepl(struct termp *p, enum termfont f) { - switch (node->type) { - case (MDOC_BODY): - p->flags |= TERMP_NOSPACE; - word(p, "\\]"); - break; - default: - break; - } - return(1); + p->fontl = p->fontq[p->fonti]; + p->fontq[p->fonti] = f; } -/* ARGSUSED */ -static int -termp_sh_post(struct termp *p, const struct mdoc_meta *meta, - const struct mdoc_node *node) +void +term_fontpush(struct termp *p, enum termfont f) { - switch (node->type) { - case (MDOC_HEAD): - p->flags &= ~TERMP_BOLD; - newln(p); - break; - case (MDOC_BODY): - newln(p); - p->offset -= 4; - break; - default: - break; - } - return(1); + assert(p->fonti + 1 < 10); + p->fontl = p->fontq[p->fonti]; + p->fontq[++p->fonti] = f; } -/* ARGSUSED */ -static int -termp_xr_pre(struct termp *p, const struct mdoc_meta *meta, - const struct mdoc_node *node) +const void * +term_fontq(struct termp *p) { - const struct mdoc_node *n; - n = node->child; - assert(n); + return(&p->fontq[p->fonti]); +} - assert(MDOC_TEXT == n->type); - word(p, n->data.text.string); - if (NULL == (n = n->next)) - return(0); +enum termfont +term_fonttop(struct termp *p) +{ - 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->fontq[p->fonti]); } -/* ARGSUSED */ -static int -termp_sh_pre(struct termp *p, const struct mdoc_meta *meta, - const struct mdoc_node *node) +void +term_fontpopq(struct termp *p, const void *key) { - switch (node->type) { - case (MDOC_HEAD): - vspace(p); - p->flags |= TERMP_BOLD; - break; - case (MDOC_BODY): - p->offset += 4; - break; - default: - break; - } - return(1); + while (p->fonti >= 0 && key != &p->fontq[p->fonti]) + p->fonti--; + assert(p->fonti >= 0); } -/* ARGSUSED */ -static int -termp_op_pre(struct termp *p, const struct mdoc_meta *meta, - const struct mdoc_node *node) +void +term_fontpop(struct termp *p) { - switch (node->type) { - case (MDOC_BODY): - word(p, "\\["); - p->flags |= TERMP_NOSPACE; - break; - default: - break; - } - return(1); + assert(p->fonti); + p->fonti--; } -/* ARGSUSED */ -static int -termp_fl_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 char *sv, *seq; + int sz; + size_t ssz; + enum roffdeco deco; - p->flags |= TERMP_BOLD; - word(p, "-"); - p->flags |= TERMP_NOSPACE; - return(1); -} + 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(']'): + if ( ! (TERMP_IGNDELIM & p->flags)) + p->flags |= TERMP_NOSPACE; + break; + default: + break; + } -static void -termprint_r(struct termp *p, const struct mdoc_meta *meta, - const struct mdoc_node *node) -{ - int dochild; + if ( ! (TERMP_NOSPACE & p->flags)) { + bufferc(p, ' '); + if (TERMP_SENTENCE & p->flags) + bufferc(p, ' '); + } - /* Pre-processing ----------------- */ + if ( ! (p->flags & TERMP_NONOSPACE)) + p->flags &= ~TERMP_NOSPACE; - dochild = 1; + p->flags &= ~TERMP_SENTENCE; - 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); + /* FIXME: use strcspn. */ - /* Children ---------------------- */ + while (*word) { + if ('\\' != *word) { + encode(p, word, 1); + word++; + continue; + } - if (dochild && NULL == node->child) { - /* No-child processing. */ - switch (node->type) { - case (MDOC_ELEM): - switch (node->tok) { - case (MDOC_Nm): - word(p, "progname"); /* TODO */ - break; - case (MDOC_Ar): - word(p, "..."); - break; - default: - break; - } + 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; } - } else if (dochild) - termprint_r(p, meta, node->child); - /* Post-processing --------------- */ + word += sz; + if (DECO_NOSPACE == deco && '\0' == *word) + p->flags |= TERMP_NOSPACE; + } - if (MDOC_TEXT != node->type) { - if (termacts[node->tok].post) - if ( ! (*termacts[node->tok].post)(p, meta, node)) - return; - } - - /* Siblings ---------------------- */ - - if (node->next) - termprint_r(p, meta, node->next); + /* + * Note that we don't process the pipe: the parser sees it as + * punctuation, but we don't in terms of typography. + */ + if (sv[0] && 0 == sv[1]) + switch (sv[0]) { + case('('): + /* FALLTHROUGH */ + case('['): + p->flags |= TERMP_NOSPACE; + break; + default: + break; + } } static void -termprint_footer(struct termp *p, const struct mdoc_meta *meta) +adjbuf(struct termp *p, size_t sz) { - 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"); + if (0 == p->maxcols) + p->maxcols = 1024; + while (sz >= p->maxcols) + p->maxcols <<= 2; - tm = localtime(&meta->date); - if (NULL == strftime(buf, p->rmargin, "%B %d, %Y", tm)) - err(1, "strftime"); + p->buf = realloc(p->buf, p->maxcols); + if (NULL == p->buf) { + perror(NULL); + exit(EXIT_FAILURE); + } +} - osz = strlcpy(os, meta->os, p->rmargin); - sz = strlen(buf); - ssz = sz + osz + 1; +static void +buffera(struct termp *p, const char *word, size_t sz) +{ - if (ssz > p->rmargin) { - ssz -= p->rmargin; - assert(ssz <= osz); - os[osz - ssz] = 0; - ssz = 1; - } else - ssz = p->rmargin - ssz + 1; + if (p->col + sz >= p->maxcols) + adjbuf(p, p->col + sz); - printf("\n"); - printf("%s", os); - for (i = 0; i < ssz; i++) - printf(" "); - - printf("%s\n", buf); - fflush(stdout); - - free(buf); - free(os); + memcpy(&p->buf[(int)p->col], word, sz); + p->col += sz; } static void -termprint_header(struct termp *p, const struct mdoc_meta *meta) +bufferc(struct termp *p, char c) { - 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 (p->col + 1 >= p->maxcols) + adjbuf(p, p->col + 1); - 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); - break; - case (MSEC_8): - pp = mdoc_vol2a(VOL_SMM); - break; - case (MSEC_2): - /* FALLTHROUGH */ - case (MSEC_3): - /* FALLTHROUGH */ - case (MSEC_4): - /* FALLTHROUGH */ - case (MSEC_5): - pp = mdoc_vol2a(VOL_PRM); - 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); + p->buf[(int)p->col++] = c; +} - 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); - } +static void +encode(struct termp *p, const char *word, size_t sz) +{ + enum termfont f; + int i; - ttsz = strlcpy(title, meta->title, p->rmargin); + /* + * Encode and buffer a string of characters. If the current + * font mode is unset, buffer directly, else encode then buffer + * character by character. + */ - if (NULL == (msec = mdoc_msec2a(meta->msec))) - msec = ""; + if (TERMTYPE_PS == p->type) { + buffera(p, word, sz); + return; + } else if (TERMFONT_NONE == (f = term_fonttop(p))) { + buffera(p, word, sz); + return; + } - ssz = (2 * (ttsz + 2 + strlen(msec))) + tsz + 2; + for (i = 0; i < (int)sz; i++) { + if ( ! isgraph((u_char)word[i])) { + bufferc(p, word[i]); + continue; + } - 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; + if (TERMFONT_UNDER == f) + bufferc(p, '_'); + else + bufferc(p, word[i]); - printf("%s(%s)", title, msec); + bufferc(p, 8); + bufferc(p, word[i]); + } +} - for (i = 0; i < ssz; i++) - printf(" "); - printf("%s", buf); +size_t +term_vspan(const struct roffsu *su) +{ + double r; - for (i = 0; i < ssz; i++) - printf(" "); + 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; + } - printf("%s(%s)\n", title, msec); - fflush(stdout); - - free(title); - free(buf); + if (r < 0.0) + r = 0.0; + return(/* LINTED */(size_t) + r); } -int -termprint(const struct mdoc_node *node, - const struct mdoc_meta *meta) +size_t +term_hspan(const struct roffsu *su) { - struct termp p; + double r; - if (ERR == setupterm(NULL, STDOUT_FILENO, NULL)) - return(0); + /* XXX: CM, IN, and PT are approximations. */ - p.maxrmargin = columns < 60 ? 60 : (size_t)columns; - p.rmargin = p.maxrmargin; - p.maxcols = 1024; - p.offset = p.col = 0; - p.flags = TERMP_NOSPACE; + 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 (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)/* LINTED */ + r); }