=================================================================== RCS file: /cvs/mandoc/html.c,v retrieving revision 1.31 retrieving revision 1.263 diff -u -p -r1.31 -r1.263 --- mandoc/html.c 2009/09/17 07:41:28 1.31 +++ mandoc/html.c 2020/02/27 22:28:13 1.263 @@ -1,813 +1,989 @@ -/* $Id: html.c,v 1.31 2009/09/17 07:41:28 kristaps Exp $ */ +/* $Id: html.c,v 1.263 2020/02/27 22:28:13 schwarze Exp $ */ /* - * Copyright (c) 2008, 2009 Kristaps Dzonsons + * Copyright (c) 2008-2011, 2014 Kristaps Dzonsons + * Copyright (c) 2011-2015, 2017-2020 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. * - * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES + * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHORS DISCLAIM ALL WARRANTIES * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF - * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR + * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHORS 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 +#include "config.h" +#include +#include + #include -#include +#include +#include +#include #include +#include #include +#include +#include -#include "mdoc.h" -#include "man.h" +#include "mandoc_aux.h" +#include "mandoc_ohash.h" +#include "mandoc.h" +#include "roff.h" +#include "out.h" +#include "html.h" +#include "manconf.h" +#include "main.h" -#define DOCTYPE "-//W3C//DTD HTML 4.01//EN" -#define DTD "http://www.w3.org/TR/html4/strict.dtd" - -enum htmltag { - TAG_HTML, - TAG_HEAD, - TAG_BODY, - TAG_META, - TAG_TITLE, - TAG_DIV, - TAG_H1, - TAG_H2, - TAG_P, - TAG_SPAN, - TAG_LINK, - TAG_BR, - TAG_A, - TAG_MAX -}; - -enum htmlattr { - ATTR_HTTPEQUIV, - ATTR_CONTENT, - ATTR_NAME, - ATTR_REL, - ATTR_HREF, - ATTR_TYPE, - ATTR_MEDIA, - ATTR_CLASS, - ATTR_MAX -}; - struct htmldata { - char *name; + const char *name; int flags; -#define HTML_CLRLINE (1 << 0) -#define HTML_NOSTACK (1 << 1) +#define HTML_INPHRASE (1 << 0) /* Can appear in phrasing context. */ +#define HTML_TOPHRASE (1 << 1) /* Establishes phrasing context. */ +#define HTML_NOSTACK (1 << 2) /* Does not have an end tag. */ +#define HTML_NLBEFORE (1 << 3) /* Output line break before opening. */ +#define HTML_NLBEGIN (1 << 4) /* Output line break after opening. */ +#define HTML_NLEND (1 << 5) /* Output line break before closing. */ +#define HTML_NLAFTER (1 << 6) /* Output line break after closing. */ +#define HTML_NLAROUND (HTML_NLBEFORE | HTML_NLAFTER) +#define HTML_NLINSIDE (HTML_NLBEGIN | HTML_NLEND) +#define HTML_NLALL (HTML_NLAROUND | HTML_NLINSIDE) +#define HTML_INDENT (1 << 7) /* Indent content by two spaces. */ +#define HTML_NOINDENT (1 << 8) /* Exception: never indent content. */ }; static const struct htmldata htmltags[TAG_MAX] = { - {"html", HTML_CLRLINE}, /* TAG_HTML */ - {"head", HTML_CLRLINE}, /* TAG_HEAD */ - {"body", HTML_CLRLINE}, /* TAG_BODY */ - {"meta", HTML_CLRLINE | HTML_NOSTACK}, /* TAG_META */ - {"title", HTML_CLRLINE | HTML_NOSTACK}, /* TAG_TITLE */ - {"div", HTML_CLRLINE}, /* TAG_DIV */ - {"h1", 0}, /* TAG_H1 */ - {"h2", 0}, /* TAG_H2 */ - {"p", HTML_CLRLINE}, /* TAG_P */ - {"span", 0}, /* TAG_SPAN */ - {"link", HTML_CLRLINE | HTML_NOSTACK}, /* TAG_LINK */ - {"br", HTML_CLRLINE | HTML_NOSTACK}, /* TAG_LINK */ - {"a", 0}, /* TAG_A */ + {"html", HTML_NLALL}, + {"head", HTML_NLALL | HTML_INDENT}, + {"meta", HTML_NOSTACK | HTML_NLALL}, + {"link", HTML_NOSTACK | HTML_NLALL}, + {"style", HTML_NLALL | HTML_INDENT}, + {"title", HTML_NLAROUND}, + {"body", HTML_NLALL}, + {"div", HTML_NLAROUND}, + {"section", HTML_NLALL}, + {"table", HTML_NLALL | HTML_INDENT}, + {"tr", HTML_NLALL | HTML_INDENT}, + {"td", HTML_NLAROUND}, + {"li", HTML_NLAROUND | HTML_INDENT}, + {"ul", HTML_NLALL | HTML_INDENT}, + {"ol", HTML_NLALL | HTML_INDENT}, + {"dl", HTML_NLALL | HTML_INDENT}, + {"dt", HTML_NLAROUND}, + {"dd", HTML_NLAROUND | HTML_INDENT}, + {"h1", HTML_TOPHRASE | HTML_NLAROUND}, + {"h2", HTML_TOPHRASE | HTML_NLAROUND}, + {"p", HTML_TOPHRASE | HTML_NLAROUND | HTML_INDENT}, + {"pre", HTML_TOPHRASE | HTML_NLALL | HTML_NOINDENT}, + {"a", HTML_INPHRASE | HTML_TOPHRASE}, + {"b", HTML_INPHRASE | HTML_TOPHRASE}, + {"cite", HTML_INPHRASE | HTML_TOPHRASE}, + {"code", HTML_INPHRASE | HTML_TOPHRASE}, + {"i", HTML_INPHRASE | HTML_TOPHRASE}, + {"small", HTML_INPHRASE | HTML_TOPHRASE}, + {"span", HTML_INPHRASE | HTML_TOPHRASE}, + {"var", HTML_INPHRASE | HTML_TOPHRASE}, + {"br", HTML_INPHRASE | HTML_NOSTACK | HTML_NLALL}, + {"mark", HTML_INPHRASE }, + {"math", HTML_INPHRASE | HTML_NLALL | HTML_INDENT}, + {"mrow", 0}, + {"mi", 0}, + {"mn", 0}, + {"mo", 0}, + {"msup", 0}, + {"msub", 0}, + {"msubsup", 0}, + {"mfrac", 0}, + {"msqrt", 0}, + {"mfenced", 0}, + {"mtable", 0}, + {"mtr", 0}, + {"mtd", 0}, + {"munderover", 0}, + {"munder", 0}, + {"mover", 0}, }; -static const char *const htmlattrs[ATTR_MAX] = { - "http-equiv", - "content", - "name", - "rel", - "href", - "type", - "media", - "class" -}; +/* Avoid duplicate HTML id= attributes. */ +static struct ohash id_unique; -struct htmlpair { - enum htmlattr key; - char *val; -}; +static void html_reset_internal(struct html *); +static void print_byte(struct html *, char); +static void print_endword(struct html *); +static void print_indent(struct html *); +static void print_word(struct html *, const char *); -struct tag { - enum htmltag tag; - SLIST_ENTRY(tag) entry; -}; +static void print_ctag(struct html *, struct tag *); +static int print_escape(struct html *, char); +static int print_encode(struct html *, const char *, const char *, int); +static void print_href(struct html *, const char *, const char *, int); +static void print_metaf(struct html *); -SLIST_HEAD(tagq, tag); -struct html { - int flags; -#define HTML_NOSPACE (1 << 0) -#define HTML_NEWLINE (1 << 1) - struct tagq stack; -}; +void * +html_alloc(const struct manoutput *outopts) +{ + struct html *h; -#define MDOC_ARGS const struct mdoc_meta *m, \ - const struct mdoc_node *n, \ - struct html *h -#define MAN_ARGS const struct man_meta *m, \ - const struct man_node *n, \ - struct html *h -struct htmlmdoc { - int (*pre)(MDOC_ARGS); - void (*post)(MDOC_ARGS); -}; + h = mandoc_calloc(1, sizeof(struct html)); -static void print_gen_doctype(struct html *); -static void print_gen_head(struct html *); -static void print_mdoc(MDOC_ARGS); -static void print_mdoc_head(MDOC_ARGS); -static void print_mdoc_title(MDOC_ARGS); -static void print_mdoc_node(MDOC_ARGS); -static void print_man(MAN_ARGS); -static void print_man_head(MAN_ARGS); -static void print_man_body(MAN_ARGS); -static struct tag *print_otag(struct html *, enum htmltag, - int, const struct htmlpair *); -static void print_tagq(struct html *, const struct tag *); -static void print_stagq(struct html *, const struct tag *); -static void print_ctag(struct html *, enum htmltag); -static void print_encode(const char *); -static void print_text(struct html *, const char *); -static int mdoc_root_pre(MDOC_ARGS); + h->tag = NULL; + h->style = outopts->style; + if ((h->base_man1 = outopts->man) == NULL) + h->base_man2 = NULL; + else if ((h->base_man2 = strchr(h->base_man1, ';')) != NULL) + *h->base_man2++ = '\0'; + h->base_includes = outopts->includes; + if (outopts->fragment) + h->oflags |= HTML_FRAGMENT; + if (outopts->toc) + h->oflags |= HTML_TOC; -static int mdoc_ar_pre(MDOC_ARGS); -static int mdoc_fl_pre(MDOC_ARGS); -static int mdoc_nd_pre(MDOC_ARGS); -static int mdoc_nm_pre(MDOC_ARGS); -static int mdoc_ns_pre(MDOC_ARGS); -static int mdoc_op_pre(MDOC_ARGS); -static void mdoc_op_post(MDOC_ARGS); -static int mdoc_pp_pre(MDOC_ARGS); -static int mdoc_sh_pre(MDOC_ARGS); -static int mdoc_ss_pre(MDOC_ARGS); -static int mdoc_xr_pre(MDOC_ARGS); + mandoc_ohash_init(&id_unique, 4, 0); -static const struct htmlmdoc mdocs[MDOC_MAX] = { - {NULL, NULL}, /* Ap */ - {NULL, NULL}, /* Dd */ - {NULL, NULL}, /* Dt */ - {NULL, NULL}, /* Os */ - {mdoc_sh_pre, NULL }, /* Sh */ - {mdoc_ss_pre, NULL }, /* Ss */ - {mdoc_pp_pre, NULL}, /* Pp */ - {NULL, NULL}, /* D1 */ - {NULL, NULL}, /* Dl */ - {NULL, NULL}, /* Bd */ - {NULL, NULL}, /* Ed */ - {NULL, NULL}, /* Bl */ - {NULL, NULL}, /* El */ - {NULL, NULL}, /* It */ - {NULL, NULL}, /* Ad */ - {NULL, NULL}, /* An */ - {mdoc_ar_pre, NULL}, /* 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 */ - {mdoc_fl_pre, NULL}, /* Fl */ - {NULL, NULL}, /* Fn */ - {NULL, NULL}, /* Ft */ - {NULL, NULL}, /* Ic */ - {NULL, NULL}, /* In */ - {NULL, NULL}, /* Li */ - {mdoc_nd_pre, NULL}, /* Nd */ - {mdoc_nm_pre, NULL}, /* Nm */ - {mdoc_op_pre, mdoc_op_post}, /* Op */ - {NULL, NULL}, /* Ot */ - {NULL, NULL}, /* Pa */ - {NULL, NULL}, /* Rv */ - {NULL, NULL}, /* St */ - {NULL, NULL}, /* Va */ - {NULL, NULL}, /* Vt */ - {mdoc_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 */ - {mdoc_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 */ - {NULL, NULL}, /* Lb */ - {NULL, NULL}, /* Lp */ - {NULL, NULL}, /* Lk */ - {NULL, NULL}, /* Mt */ - {NULL, NULL}, /* Brq */ - {NULL, NULL}, /* Bro */ - {NULL, NULL}, /* Brc */ - {NULL, NULL}, /* %C */ - {NULL, NULL}, /* Es */ - {NULL, NULL}, /* En */ - {NULL, NULL}, /* Dx */ - {NULL, NULL}, /* %Q */ - {NULL, NULL}, /* br */ - {NULL, NULL}, /* sp */ -}; + return h; +} +static void +html_reset_internal(struct html *h) +{ + struct tag *tag; + char *cp; + unsigned int slot; + while ((tag = h->tag) != NULL) { + h->tag = tag->next; + free(tag); + } + cp = ohash_first(&id_unique, &slot); + while (cp != NULL) { + free(cp); + cp = ohash_next(&id_unique, &slot); + } + ohash_delete(&id_unique); +} + void -html_mdoc(void *arg, const struct mdoc *m) +html_reset(void *p) { - struct html *h; - struct tag *t; + html_reset_internal(p); + mandoc_ohash_init(&id_unique, 4, 0); +} - h = (struct html *)arg; - - print_gen_doctype(h); - t = print_otag(h, TAG_HTML, 0, NULL); - print_mdoc(mdoc_meta(m), mdoc_node(m), h); - print_tagq(h, t); - - printf("\n"); +void +html_free(void *p) +{ + html_reset_internal(p); + free(p); } - void -html_man(void *arg, const struct man *m) +print_gen_head(struct html *h) { - struct html *h; struct tag *t; - h = (struct html *)arg; + print_otag(h, TAG_META, "?", "charset", "utf-8"); + if (h->style != NULL) { + print_otag(h, TAG_LINK, "?h??", "rel", "stylesheet", + h->style, "type", "text/css", "media", "all"); + return; + } - print_gen_doctype(h); - t = print_otag(h, TAG_HTML, 0, NULL); - print_man(man_meta(m), man_node(m), h); - print_tagq(h, t); + /* + * Print a minimal embedded style sheet. + */ - printf("\n"); + t = print_otag(h, TAG_STYLE, ""); + print_text(h, "table.head, table.foot { width: 100%; }"); + print_endline(h); + print_text(h, "td.head-rtitle, td.foot-os { text-align: right; }"); + print_endline(h); + print_text(h, "td.head-vol { text-align: center; }"); + print_endline(h); + print_text(h, ".Nd, .Bf, .Op { display: inline; }"); + print_endline(h); + print_text(h, ".Pa, .Ad { font-style: italic; }"); + print_endline(h); + print_text(h, ".Ms { font-weight: bold; }"); + print_endline(h); + print_text(h, ".Bl-diag "); + print_byte(h, '>'); + print_text(h, " dt { font-weight: bold; }"); + print_endline(h); + print_text(h, "code.Nm, .Fl, .Cm, .Ic, code.In, .Fd, .Fn, .Cd " + "{ font-weight: bold; font-family: inherit; }"); + print_tagq(h, t); } - -void * -html_alloc(void) +int +html_setfont(struct html *h, enum mandoc_esc font) { - struct html *h; + switch (font) { + case ESCAPE_FONTPREV: + font = h->metal; + break; + case ESCAPE_FONTITALIC: + case ESCAPE_FONTBOLD: + case ESCAPE_FONTBI: + case ESCAPE_FONTCW: + case ESCAPE_FONTROMAN: + break; + case ESCAPE_FONT: + font = ESCAPE_FONTROMAN; + break; + default: + return 0; + } + h->metal = h->metac; + h->metac = font; + return 1; +} - if (NULL == (h = calloc(1, sizeof(struct html)))) - return(NULL); - - SLIST_INIT(&h->stack); - return(h); +static void +print_metaf(struct html *h) +{ + if (h->metaf) { + print_tagq(h, h->metaf); + h->metaf = NULL; + } + switch (h->metac) { + case ESCAPE_FONTITALIC: + h->metaf = print_otag(h, TAG_I, ""); + break; + case ESCAPE_FONTBOLD: + h->metaf = print_otag(h, TAG_B, ""); + break; + case ESCAPE_FONTBI: + h->metaf = print_otag(h, TAG_B, ""); + print_otag(h, TAG_I, ""); + break; + case ESCAPE_FONTCW: + h->metaf = print_otag(h, TAG_SPAN, "c", "Li"); + break; + default: + break; + } } - void -html_free(void *p) +html_close_paragraph(struct html *h) { - struct tag *tag; - struct html *h; + struct tag *this, *next; + int flags; - h = (struct html *)p; - - while ( ! SLIST_EMPTY(&h->stack)) { - tag = SLIST_FIRST(&h->stack); - SLIST_REMOVE_HEAD(&h->stack, entry); - free(tag); + this = h->tag; + for (;;) { + next = this->next; + flags = htmltags[this->tag].flags; + if (flags & (HTML_INPHRASE | HTML_TOPHRASE)) + print_ctag(h, this); + if ((flags & HTML_INPHRASE) == 0) + break; + this = next; } - free(h); } - -static void -print_mdoc(MDOC_ARGS) +/* + * ROFF_nf switches to no-fill mode, ROFF_fi to fill mode. + * TOKEN_NONE does not switch. The old mode is returned. + */ +enum roff_tok +html_fillmode(struct html *h, enum roff_tok want) { struct tag *t; + enum roff_tok had; - t = print_otag(h, TAG_HEAD, 0, NULL); - print_mdoc_head(m, n, h); - print_tagq(h, t); + for (t = h->tag; t != NULL; t = t->next) + if (t->tag == TAG_PRE) + break; - t = print_otag(h, TAG_BODY, 0, NULL); - print_mdoc_title(m, n, h); - print_mdoc_node(m, n, h); - print_tagq(h, t); + had = t == NULL ? ROFF_fi : ROFF_nf; + + if (want != had) { + switch (want) { + case ROFF_fi: + print_tagq(h, t); + break; + case ROFF_nf: + html_close_paragraph(h); + print_otag(h, TAG_PRE, ""); + break; + case TOKEN_NONE: + break; + default: + abort(); + } + } + return had; } - -static void -print_gen_head(struct html *h) +char * +html_make_id(const struct roff_node *n, int unique) { - struct htmlpair meta0[2]; - struct htmlpair meta1[2]; - struct htmlpair link[4]; + const struct roff_node *nch; + char *buf, *bufs, *cp; + unsigned int slot; + int suffix; - meta0[0].key = ATTR_HTTPEQUIV; - meta0[0].val = "Content-Type"; - meta0[1].key = ATTR_CONTENT; - meta0[1].val = "text/html; charest-utf-8"; + for (nch = n->child; nch != NULL; nch = nch->next) + if (nch->type != ROFFT_TEXT) + return NULL; - meta1[0].key = ATTR_NAME; - meta1[0].val = "resource-type"; - meta1[1].key = ATTR_CONTENT; - meta1[1].val = "document"; + buf = NULL; + deroff(&buf, n); + if (buf == NULL) + return NULL; - link[0].key = ATTR_REL; - link[0].val = "stylesheet"; - link[1].key = ATTR_HREF; - link[1].val = "style.css"; /* XXX */ - link[2].key = ATTR_TYPE; - link[2].val = "text/css"; - link[3].key = ATTR_MEDIA; - link[3].val = "all"; + /* + * In ID attributes, only use ASCII characters that are + * permitted in URL-fragment strings according to the + * explicit list at: + * https://url.spec.whatwg.org/#url-fragment-string + */ - print_otag(h, TAG_META, 2, meta0); - print_otag(h, TAG_META, 2, meta1); - print_otag(h, TAG_LINK, 4, link); -} + for (cp = buf; *cp != '\0'; cp++) + if (isalnum((unsigned char)*cp) == 0 && + strchr("!$&'()*+,-./:;=?@_~", *cp) == NULL) + *cp = '_'; + if (unique == 0) + return buf; -/* ARGSUSED */ -static void -print_mdoc_head(MDOC_ARGS) -{ + /* Avoid duplicate HTML id= attributes. */ - print_gen_head(h); - print_otag(h, TAG_TITLE, 0, NULL); - print_encode(m->title); + bufs = NULL; + suffix = 1; + slot = ohash_qlookup(&id_unique, buf); + cp = ohash_find(&id_unique, slot); + if (cp != NULL) { + while (cp != NULL) { + free(bufs); + if (++suffix > 127) { + free(buf); + return NULL; + } + mandoc_asprintf(&bufs, "%s_%d", buf, suffix); + slot = ohash_qlookup(&id_unique, bufs); + cp = ohash_find(&id_unique, slot); + } + free(buf); + buf = bufs; + } + ohash_insert(&id_unique, slot, buf); + return buf; } - -/* ARGSUSED */ -static void -print_mdoc_title(MDOC_ARGS) +static int +print_escape(struct html *h, char c) { - /* TODO */ -} - - -static void -print_mdoc_node(MDOC_ARGS) -{ - int child; - struct tag *t; - - child = 1; - t = SLIST_FIRST(&h->stack); - - switch (n->type) { - case (MDOC_ROOT): - child = mdoc_root_pre(m, n, h); + switch (c) { + case '<': + print_word(h, "<"); break; - case (MDOC_TEXT): - print_text(h, n->string); + case '>': + print_word(h, ">"); break; - default: - if (mdocs[n->tok].pre) - child = (*mdocs[n->tok].pre)(m, n, h); + case '&': + print_word(h, "&"); break; - } - - if (child && n->child) - print_mdoc_node(m, n->child, h); - - print_stagq(h, t); - - switch (n->type) { - case (MDOC_ROOT): + case '"': + print_word(h, """); break; - case (MDOC_TEXT): + case ASCII_NBRSP: + print_word(h, " "); break; - default: - if (mdocs[n->tok].post) - (*mdocs[n->tok].post)(m, n, h); + case ASCII_HYPH: + print_byte(h, '-'); break; + case ASCII_BREAK: + break; + default: + return 0; } - - if (n->next) - print_mdoc_node(m, n->next, h); + return 1; } - -static void -print_man(MAN_ARGS) +static int +print_encode(struct html *h, const char *p, const char *pend, int norecurse) { - struct tag *t; + char numbuf[16]; + const char *seq; + size_t sz; + int c, len, breakline, nospace; + enum mandoc_esc esc; + static const char rejs[10] = { ' ', '\\', '<', '>', '&', '"', + ASCII_NBRSP, ASCII_HYPH, ASCII_BREAK, '\0' }; - t = print_otag(h, TAG_HEAD, 0, NULL); - print_man_head(m, n, h); - print_tagq(h, t); + if (pend == NULL) + pend = strchr(p, '\0'); - t = print_otag(h, TAG_BODY, 0, NULL); - print_man_body(m, n, h); - print_tagq(h, t); -} + breakline = 0; + nospace = 0; + while (p < pend) { + if (HTML_SKIPCHAR & h->flags && '\\' != *p) { + h->flags &= ~HTML_SKIPCHAR; + p++; + continue; + } -/* ARGSUSED */ -static void -print_man_head(MAN_ARGS) -{ + for (sz = strcspn(p, rejs); sz-- && p < pend; p++) + print_byte(h, *p); - print_gen_head(h); - print_otag(h, TAG_TITLE, 0, NULL); - print_encode(m->title); -} + if (breakline && + (p >= pend || *p == ' ' || *p == ASCII_NBRSP)) { + print_otag(h, TAG_BR, ""); + breakline = 0; + while (p < pend && (*p == ' ' || *p == ASCII_NBRSP)) + p++; + continue; + } + if (p >= pend) + break; -/* ARGSUSED */ -static void -print_man_body(MAN_ARGS) -{ + if (*p == ' ') { + print_endword(h); + p++; + continue; + } - /* TODO */ -} + if (print_escape(h, *p++)) + continue; + esc = mandoc_escape(&p, &seq, &len); + switch (esc) { + case ESCAPE_FONT: + case ESCAPE_FONTPREV: + case ESCAPE_FONTBOLD: + case ESCAPE_FONTITALIC: + case ESCAPE_FONTBI: + case ESCAPE_FONTCW: + case ESCAPE_FONTROMAN: + if (0 == norecurse) { + h->flags |= HTML_NOSPACE; + if (html_setfont(h, esc)) + print_metaf(h); + h->flags &= ~HTML_NOSPACE; + } + continue; + case ESCAPE_SKIPCHAR: + h->flags |= HTML_SKIPCHAR; + continue; + case ESCAPE_ERROR: + continue; + default: + break; + } + if (h->flags & HTML_SKIPCHAR) { + h->flags &= ~HTML_SKIPCHAR; + continue; + } + + switch (esc) { + case ESCAPE_UNICODE: + /* Skip past "u" header. */ + c = mchars_num2uc(seq + 1, len - 1); + break; + case ESCAPE_NUMBERED: + c = mchars_num2char(seq, len); + if (c < 0) + continue; + break; + case ESCAPE_SPECIAL: + c = mchars_spec2cp(seq, len); + if (c <= 0) + continue; + break; + case ESCAPE_UNDEF: + c = *seq; + break; + case ESCAPE_DEVICE: + print_word(h, "html"); + continue; + case ESCAPE_BREAK: + breakline = 1; + continue; + case ESCAPE_NOSPACE: + if ('\0' == *p) + nospace = 1; + continue; + case ESCAPE_OVERSTRIKE: + if (len == 0) + continue; + c = seq[len - 1]; + break; + default: + continue; + } + if ((c < 0x20 && c != 0x09) || + (c > 0x7E && c < 0xA0)) + c = 0xFFFD; + if (c > 0x7E) { + (void)snprintf(numbuf, sizeof(numbuf), "&#x%.4X;", c); + print_word(h, numbuf); + } else if (print_escape(h, c) == 0) + print_byte(h, c); + } + + return nospace; +} + static void -print_encode(const char *p) +print_href(struct html *h, const char *name, const char *sec, int man) { + struct stat sb; + const char *p, *pp; + char *filename; - printf("%s", p); /* XXX */ + if (man) { + pp = h->base_man1; + if (h->base_man2 != NULL) { + mandoc_asprintf(&filename, "%s.%s", name, sec); + if (stat(filename, &sb) == -1) + pp = h->base_man2; + free(filename); + } + } else + pp = h->base_includes; + + while ((p = strchr(pp, '%')) != NULL) { + print_encode(h, pp, p, 1); + if (man && p[1] == 'S') { + if (sec == NULL) + print_byte(h, '1'); + else + print_encode(h, sec, NULL, 1); + } else if ((man && p[1] == 'N') || + (man == 0 && p[1] == 'I')) + print_encode(h, name, NULL, 1); + else + print_encode(h, p, p + 2, 1); + pp = p + 2; + } + if (*pp != '\0') + print_encode(h, pp, NULL, 1); } - -static struct tag * -print_otag(struct html *h, enum htmltag tag, - int sz, const struct htmlpair *p) +struct tag * +print_otag(struct html *h, enum htmltag tag, const char *fmt, ...) { - int i; + va_list ap; struct tag *t; + const char *attr; + char *arg1, *arg2; + int style_written, tflags; - if ( ! (HTML_NOSTACK & htmltags[tag].flags)) { - if (NULL == (t = malloc(sizeof(struct tag)))) - err(EXIT_FAILURE, "malloc"); + tflags = htmltags[tag].flags; + + /* Flow content is not allowed in phrasing context. */ + + if ((tflags & HTML_INPHRASE) == 0) { + for (t = h->tag; t != NULL; t = t->next) { + if (t->closed) + continue; + assert((htmltags[t->tag].flags & HTML_TOPHRASE) == 0); + break; + } + + /* + * Always wrap phrasing elements in a paragraph + * unless already contained in some flow container; + * never put them directly into a section. + */ + + } else if (tflags & HTML_TOPHRASE && h->tag->tag == TAG_SECTION) + print_otag(h, TAG_P, "c", "Pp"); + + /* Push this tag onto the stack of open scopes. */ + + if ((tflags & HTML_NOSTACK) == 0) { + t = mandoc_malloc(sizeof(struct tag)); t->tag = tag; - SLIST_INSERT_HEAD(&h->stack, t, entry); + t->next = h->tag; + t->refcnt = 0; + t->closed = 0; + h->tag = t; } else t = NULL; - if ( ! (HTML_NOSPACE & h->flags)) - if ( ! (HTML_CLRLINE & htmltags[tag].flags)) - printf(" "); - - printf("<%s", htmltags[tag].name); - for (i = 0; i < sz; i++) { - printf(" %s=\"", htmlattrs[p[i].key]); - assert(p->val); - print_encode(p[i].val); - printf("\""); + if (tflags & HTML_NLBEFORE) + print_endline(h); + if (h->col == 0) + print_indent(h); + else if ((h->flags & HTML_NOSPACE) == 0) { + if (h->flags & HTML_KEEP) + print_word(h, " "); + else { + if (h->flags & HTML_PREKEEP) + h->flags |= HTML_KEEP; + print_endword(h); + } } - printf(">"); - h->flags |= HTML_NOSPACE; - if (HTML_CLRLINE & htmltags[tag].flags) - h->flags |= HTML_NEWLINE; + if ( ! (h->flags & HTML_NONOSPACE)) + h->flags &= ~HTML_NOSPACE; else - h->flags &= ~HTML_NEWLINE; + h->flags |= HTML_NOSPACE; - return(t); -} + /* Print out the tag name and attributes. */ + print_byte(h, '<'); + print_word(h, htmltags[tag].name); -/* ARGSUSED */ -static void -print_ctag(struct html *h, enum htmltag tag) -{ - - printf("", htmltags[tag].name); - if (HTML_CLRLINE & htmltags[tag].flags) - h->flags |= HTML_NOSPACE; - if (HTML_CLRLINE & htmltags[tag].flags) - h->flags |= HTML_NEWLINE; - else - h->flags &= ~HTML_NEWLINE; -} + va_start(ap, fmt); + while (*fmt != '\0' && *fmt != 's') { -/* ARGSUSED */ -static void -print_gen_doctype(struct html *h) -{ - - printf("\n", DOCTYPE, DTD); -} + /* Parse attributes and arguments. */ + arg1 = va_arg(ap, char *); + arg2 = NULL; + switch (*fmt++) { + case 'c': + attr = "class"; + break; + case 'h': + attr = "href"; + break; + case 'i': + attr = "id"; + break; + case '?': + attr = arg1; + arg1 = va_arg(ap, char *); + break; + default: + abort(); + } + if (*fmt == 'M') + arg2 = va_arg(ap, char *); + if (arg1 == NULL) + continue; -static void -print_text(struct html *h, const char *p) -{ + /* Print the attributes. */ - if (*p && 0 == *(p + 1)) - switch (*p) { - case('.'): - /* FALLTHROUGH */ - case(','): - /* FALLTHROUGH */ - case(';'): - /* FALLTHROUGH */ - case(':'): - /* FALLTHROUGH */ - case('?'): - /* FALLTHROUGH */ - case('!'): - /* FALLTHROUGH */ - case(')'): - /* FALLTHROUGH */ - case(']'): - /* FALLTHROUGH */ - case('}'): - h->flags |= HTML_NOSPACE; + print_byte(h, ' '); + print_word(h, attr); + print_byte(h, '='); + print_byte(h, '"'); + switch (*fmt) { + case 'I': + print_href(h, arg1, NULL, 0); + fmt++; break; + case 'M': + print_href(h, arg1, arg2, 1); + fmt++; + break; + case 'R': + print_byte(h, '#'); + print_encode(h, arg1, NULL, 1); + fmt++; + break; default: + print_encode(h, arg1, NULL, 1); break; } + print_byte(h, '"'); + } - if ( ! (h->flags & HTML_NOSPACE)) - printf(" "); + style_written = 0; + while (*fmt++ == 's') { + arg1 = va_arg(ap, char *); + arg2 = va_arg(ap, char *); + if (arg2 == NULL) + continue; + print_byte(h, ' '); + if (style_written == 0) { + print_word(h, "style=\""); + style_written = 1; + } + print_word(h, arg1); + print_byte(h, ':'); + print_byte(h, ' '); + print_word(h, arg2); + print_byte(h, ';'); + } + if (style_written) + print_byte(h, '"'); - h->flags &= ~HTML_NOSPACE; - h->flags &= ~HTML_NEWLINE; + va_end(ap); - if (p) - print_encode(p); + /* Accommodate for "well-formed" singleton escaping. */ - if (*p && 0 == *(p + 1)) - switch (*p) { - case('('): - /* FALLTHROUGH */ - case('['): - /* FALLTHROUGH */ - case('{'): - h->flags |= HTML_NOSPACE; - break; - default: - break; - } -} + if (htmltags[tag].flags & HTML_NOSTACK) + print_byte(h, '/'); + print_byte(h, '>'); -static void -print_tagq(struct html *h, const struct tag *until) -{ - struct tag *tag; + if (tflags & HTML_NLBEGIN) + print_endline(h); + else + h->flags |= HTML_NOSPACE; - while ( ! SLIST_EMPTY(&h->stack)) { - tag = SLIST_FIRST(&h->stack); - print_ctag(h, tag->tag); - SLIST_REMOVE_HEAD(&h->stack, entry); - free(tag); - if (until && tag == until) - return; - } + if (tflags & HTML_INDENT) + h->indent++; + if (tflags & HTML_NOINDENT) + h->noindent++; + + return t; } - static void -print_stagq(struct html *h, const struct tag *suntil) +print_ctag(struct html *h, struct tag *tag) { - struct tag *tag; + int tflags; - while ( ! SLIST_EMPTY(&h->stack)) { - tag = SLIST_FIRST(&h->stack); - if (suntil && tag == suntil) - return; - print_ctag(h, tag->tag); - SLIST_REMOVE_HEAD(&h->stack, entry); + if (tag->closed == 0) { + tag->closed = 1; + if (tag == h->metaf) + h->metaf = NULL; + if (tag == h->tblt) + h->tblt = NULL; + + tflags = htmltags[tag->tag].flags; + if (tflags & HTML_INDENT) + h->indent--; + if (tflags & HTML_NOINDENT) + h->noindent--; + if (tflags & HTML_NLEND) + print_endline(h); + print_indent(h); + print_byte(h, '<'); + print_byte(h, '/'); + print_word(h, htmltags[tag->tag].name); + print_byte(h, '>'); + if (tflags & HTML_NLAFTER) + print_endline(h); + } + if (tag->refcnt == 0) { + h->tag = tag->next; free(tag); } } - -/* ARGSUSED */ -static int -mdoc_root_pre(MDOC_ARGS) +void +print_gen_decls(struct html *h) { - struct htmlpair tag; - - tag.key = ATTR_CLASS; - tag.val = "body"; - - print_otag(h, TAG_DIV, 1, &tag); - return(1); + print_word(h, ""); + print_endline(h); } - -/* ARGSUSED */ -static int -mdoc_ss_pre(MDOC_ARGS) +void +print_gen_comment(struct html *h, struct roff_node *n) { + int wantblank; - if (MDOC_BODY == n->type) - print_otag(h, TAG_P, 0, NULL); - if (MDOC_HEAD == n->type) - print_otag(h, TAG_H2, 0, NULL); - return(1); + print_word(h, "") == NULL && + (wantblank || *n->string != '\0')) { + print_endline(h); + print_indent(h); + print_word(h, n->string); + wantblank = *n->string != '\0'; + } + n = n->next; + } + if (wantblank) + print_endline(h); + print_word(h, " -->"); + print_endline(h); + h->indent = 0; } - -/* ARGSUSED */ -static int -mdoc_fl_pre(MDOC_ARGS) +void +print_text(struct html *h, const char *word) { - struct htmlpair tag; + /* + * Always wrap text in a paragraph unless already contained in + * some flow container; never put it directly into a section. + */ - tag.key = ATTR_CLASS; - tag.val = "flag"; + if (h->tag->tag == TAG_SECTION) + print_otag(h, TAG_P, "c", "Pp"); - print_otag(h, TAG_SPAN, 1, &tag); - print_text(h, "\\-"); - h->flags |= HTML_NOSPACE; - return(1); -} + /* Output whitespace before this text? */ + if (h->col && (h->flags & HTML_NOSPACE) == 0) { + if ( ! (HTML_KEEP & h->flags)) { + if (HTML_PREKEEP & h->flags) + h->flags |= HTML_KEEP; + print_endword(h); + } else + print_word(h, " "); + } -/* ARGSUSED */ -static int -mdoc_pp_pre(MDOC_ARGS) -{ + /* + * Print the text, optionally surrounded by HTML whitespace, + * optionally manually switching fonts before and after. + */ - print_otag(h, TAG_BR, 0, NULL); - print_otag(h, TAG_BR, 0, NULL); - return(0); -} + assert(h->metaf == NULL); + print_metaf(h); + print_indent(h); + if ( ! print_encode(h, word, NULL, 0)) { + if ( ! (h->flags & HTML_NONOSPACE)) + h->flags &= ~HTML_NOSPACE; + h->flags &= ~HTML_NONEWLINE; + } else + h->flags |= HTML_NOSPACE | HTML_NONEWLINE; + if (h->metaf != NULL) { + print_tagq(h, h->metaf); + h->metaf = NULL; + } -/* ARGSUSED */ -static int -mdoc_nd_pre(MDOC_ARGS) + h->flags &= ~HTML_IGNDELIM; +} + +void +print_tagq(struct html *h, const struct tag *until) { + struct tag *this, *next; - if (MDOC_BODY == n->type) - print_text(h, "\\(en"); - return(1); + for (this = h->tag; this != NULL; this = next) { + next = this == until ? NULL : this->next; + print_ctag(h, this); + } } - -/* ARGSUSED */ -static int -mdoc_op_pre(MDOC_ARGS) +/* + * Close out all open elements up to but excluding suntil. + * Note that a paragraph just inside stays open together with it + * because paragraphs include subsequent phrasing content. + */ +void +print_stagq(struct html *h, const struct tag *suntil) { + struct tag *this, *next; - if (MDOC_BODY == n->type) { - print_text(h, "\\(lB"); - h->flags |= HTML_NOSPACE; + for (this = h->tag; this != NULL; this = next) { + next = this->next; + if (this == suntil || (next == suntil && + (this->tag == TAG_P || this->tag == TAG_PRE))) + break; + print_ctag(h, this); } - return(1); } -/* ARGSUSED */ +/*********************************************************************** + * Low level output functions. + * They implement line breaking using a short static buffer. + ***********************************************************************/ + +/* + * Buffer one HTML output byte. + * If the buffer is full, flush and deactivate it and start a new line. + * If the buffer is inactive, print directly. + */ static void -mdoc_op_post(MDOC_ARGS) +print_byte(struct html *h, char c) { + if ((h->flags & HTML_BUFFER) == 0) { + putchar(c); + h->col++; + return; + } - if (MDOC_BODY != n->type) + if (h->col + h->bufcol < sizeof(h->buf)) { + h->buf[h->bufcol++] = c; return; - h->flags |= HTML_NOSPACE; - print_text(h, "\\(rB"); + } + + putchar('\n'); + h->col = 0; + print_indent(h); + putchar(' '); + putchar(' '); + fwrite(h->buf, h->bufcol, 1, stdout); + putchar(c); + h->col = (h->indent + 1) * 2 + h->bufcol + 1; + h->bufcol = 0; + h->flags &= ~HTML_BUFFER; } - -static int -mdoc_nm_pre(MDOC_ARGS) +/* + * If something was printed on the current output line, end it. + * Not to be called right after print_indent(). + */ +void +print_endline(struct html *h) { - struct htmlpair class; + if (h->col == 0) + return; - if ( ! (HTML_NEWLINE & h->flags)) - if (SEC_SYNOPSIS == n->sec) - print_otag(h, TAG_BR, 0, NULL); - - class.key = ATTR_CLASS; - class.val = "name"; - - print_otag(h, TAG_SPAN, 1, &class); - if (NULL == n->child) - print_text(h, m->name); - - return(1); + if (h->bufcol) { + putchar(' '); + fwrite(h->buf, h->bufcol, 1, stdout); + h->bufcol = 0; + } + putchar('\n'); + h->col = 0; + h->flags |= HTML_NOSPACE; + h->flags &= ~HTML_BUFFER; } - -/* ARGSUSED */ -static int -mdoc_sh_pre(MDOC_ARGS) +/* + * Flush the HTML output buffer. + * If it is inactive, activate it. + */ +static void +print_endword(struct html *h) { + if (h->noindent) { + print_byte(h, ' '); + return; + } - if (MDOC_BODY == n->type) - print_otag(h, TAG_P, 0, NULL); - if (MDOC_HEAD == n->type) - print_otag(h, TAG_H1, 0, NULL); - return(1); + if ((h->flags & HTML_BUFFER) == 0) { + h->col++; + h->flags |= HTML_BUFFER; + } else if (h->bufcol) { + putchar(' '); + fwrite(h->buf, h->bufcol, 1, stdout); + h->col += h->bufcol + 1; + } + h->bufcol = 0; } - -/* ARGSUSED */ -static int -mdoc_xr_pre(MDOC_ARGS) +/* + * If at the beginning of a new output line, + * perform indentation and mark the line as containing output. + * Make sure to really produce some output right afterwards, + * but do not use print_otag() for producing it. + */ +static void +print_indent(struct html *h) { - struct htmlpair tag; + size_t i; - tag.key = ATTR_HREF; - tag.val = "#"; /* TODO */ + if (h->col || h->noindent) + return; - print_otag(h, TAG_A, 1, &tag); - - n = n->child; - print_text(h, n->string); - if (NULL == (n = n->next)) - return(0); - - h->flags |= HTML_NOSPACE; - print_text(h, "("); - h->flags |= HTML_NOSPACE; - print_text(h, n->string); - h->flags |= HTML_NOSPACE; - print_text(h, ")"); - - return(0); + h->col = h->indent * 2; + for (i = 0; i < h->col; i++) + putchar(' '); } - -/* ARGSUSED */ -static int -mdoc_ns_pre(MDOC_ARGS) +/* + * Print or buffer some characters + * depending on the current HTML output buffer state. + */ +static void +print_word(struct html *h, const char *cp) { - - h->flags |= HTML_NOSPACE; - return(1); -} - - -/* ARGSUSED */ -static int -mdoc_ar_pre(MDOC_ARGS) -{ - struct htmlpair tag; - - tag.key = ATTR_CLASS; - tag.val = "arg"; - - print_otag(h, TAG_SPAN, 1, &tag); - return(1); + while (*cp != '\0') + print_byte(h, *cp++); }