=================================================================== RCS file: /cvs/mandoc/html.c,v retrieving revision 1.35 retrieving revision 1.271 diff -u -p -r1.35 -r1.271 --- mandoc/html.c 2009/09/20 11:05:22 1.35 +++ mandoc/html.c 2020/10/16 17:22:43 1.271 @@ -1,1639 +1,1086 @@ -/* $Id: html.c,v 1.35 2009/09/20 11:05:22 kristaps Exp $ */ +/* $Id: html.c,v 1.271 2020/10/16 17:22:43 schwarze Exp $ */ /* - * Copyright (c) 2008, 2009 Kristaps Dzonsons + * Copyright (c) 2011-2015, 2017-2020 Ingo Schwarze + * Copyright (c) 2008-2011, 2014 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. * - * 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. + * + * Common functions for mandoc(1) HTML formatters. + * For use by individual formatters and by the main program. */ -#include +#include "config.h" +#include +#include + #include #include -#include +#include +#include #include +#include #include #include +#include -#include "chars.h" -#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" - -#define INDENT 5 -#define HALFINDENT 3 -#define PX_MULT 8 - -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_TABLE, - TAG_COL, - TAG_TR, - TAG_TD, - TAG_LI, - TAG_UL, - TAG_OL, - TAG_MAX -}; - -enum htmlattr { - ATTR_HTTPEQUIV, - ATTR_CONTENT, - ATTR_NAME, - ATTR_REL, - ATTR_HREF, - ATTR_TYPE, - ATTR_MEDIA, - ATTR_CLASS, - ATTR_STYLE, - ATTR_WIDTH, - ATTR_VALIGN, - 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}, /* 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 */ - {"table", HTML_CLRLINE}, /* TAG_TABLE */ - {"col", HTML_CLRLINE | HTML_NOSTACK}, /* TAG_COL */ - {"tr", HTML_CLRLINE}, /* TAG_TR */ - {"td", HTML_CLRLINE}, /* TAG_TD */ - {"li", HTML_CLRLINE}, /* TAG_LI */ - {"ul", HTML_CLRLINE}, /* TAG_UL */ - {"ol", HTML_CLRLINE}, /* TAG_OL */ + {"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_NLAROUND | 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", - "style", - "width", - "valign", -}; +/* Avoid duplicate HTML id= attributes. */ -struct htmlpair { - enum htmlattr key; - char *val; +struct id_entry { + int ord; /* Ordinal number of the latest occurrence. */ + char id[]; /* The id= attribute without any ordinal suffix. */ }; +static struct ohash id_unique; -struct tag { - enum htmltag tag; - SLIST_ENTRY(tag) entry; -}; +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 *); -SLIST_HEAD(tagq, tag); +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 *); -struct html { - int flags; -#define HTML_NOSPACE (1 << 0) -#define HTML_NEWLINE (1 << 1) - struct tagq stack; - void *symtab; -}; -#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); -}; - -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(struct html *, const char *); -static void print_escape(struct html *, const char **); -static void print_text(struct html *, const char *); -static void print_res(struct html *, const char *, int); -static void print_spec(struct html *, const char *, int); - -static int a2width(const char *); -static int a2offs(const char *); - -static int mdoc_list_pre(MDOC_ARGS, int); -static int mdoc_listitem_pre(MDOC_ARGS); -static int mdoc_root_pre(MDOC_ARGS); -static int mdoc_tbl_pre(MDOC_ARGS, int); -static int mdoc_tbl_block_pre(MDOC_ARGS, int, int, int); -static int mdoc_tbl_body_pre(MDOC_ARGS, int, int); -static int mdoc_tbl_head_pre(MDOC_ARGS, int, int); - -static int mdoc_ar_pre(MDOC_ARGS); -static int mdoc_bl_pre(MDOC_ARGS); -static int mdoc_d1_pre(MDOC_ARGS); -static void mdoc_dq_post(MDOC_ARGS); -static int mdoc_dq_pre(MDOC_ARGS); -static int mdoc_fl_pre(MDOC_ARGS); -static int mdoc_em_pre(MDOC_ARGS); -static int mdoc_ex_pre(MDOC_ARGS); -static int mdoc_it_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 void mdoc_op_post(MDOC_ARGS); -static int mdoc_op_pre(MDOC_ARGS); -static int mdoc_pp_pre(MDOC_ARGS); -static void mdoc_pq_post(MDOC_ARGS); -static int mdoc_pq_pre(MDOC_ARGS); -static int mdoc_sh_pre(MDOC_ARGS); -static void mdoc_sq_post(MDOC_ARGS); -static int mdoc_sq_pre(MDOC_ARGS); -static int mdoc_ss_pre(MDOC_ARGS); -static int mdoc_sx_pre(MDOC_ARGS); -static int mdoc_xr_pre(MDOC_ARGS); -static int mdoc_xx_pre(MDOC_ARGS); - -#ifdef __linux__ -extern size_t strlcpy(char *, const char *, size_t); -extern size_t strlcat(char *, const char *, size_t); -#endif - -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 */ - {mdoc_d1_pre, NULL}, /* D1 */ - {mdoc_d1_pre, NULL}, /* Dl */ - {NULL, NULL}, /* Bd */ - {NULL, NULL}, /* Ed */ - {mdoc_bl_pre, NULL}, /* Bl */ - {NULL, NULL}, /* El */ - {mdoc_it_pre, 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 */ - {mdoc_ex_pre, 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 */ - {mdoc_xx_pre, NULL}, /* Bsx */ - {NULL, NULL}, /* Bx */ - {NULL, NULL}, /* Db */ - {NULL, NULL}, /* Dc */ - {NULL, NULL}, /* Do */ - {mdoc_dq_pre, mdoc_dq_post}, /* Dq */ - {NULL, NULL}, /* Ec */ - {NULL, NULL}, /* Ef */ - {mdoc_em_pre, NULL}, /* Em */ - {NULL, NULL}, /* Eo */ - {mdoc_xx_pre, NULL}, /* Fx */ - {NULL, NULL}, /* Ms */ - {NULL, NULL}, /* No */ - {mdoc_ns_pre, NULL}, /* Ns */ - {mdoc_xx_pre, NULL}, /* Nx */ - {mdoc_xx_pre, NULL}, /* Ox */ - {NULL, NULL}, /* Pc */ - {NULL, NULL}, /* Pf */ - {mdoc_pq_pre, mdoc_pq_post}, /* Po */ - {mdoc_pq_pre, mdoc_pq_post}, /* Pq */ - {NULL, NULL}, /* Qc */ - {NULL, NULL}, /* Ql */ - {NULL, NULL}, /* Qo */ - {NULL, NULL}, /* Qq */ - {NULL, NULL}, /* Re */ - {NULL, NULL}, /* Rs */ - {NULL, NULL}, /* Sc */ - {mdoc_sq_pre, mdoc_sq_post}, /* So */ - {mdoc_sq_pre, mdoc_sq_post}, /* Sq */ - {NULL, NULL}, /* Sm */ - {mdoc_sx_pre, NULL}, /* Sx */ - {NULL, NULL}, /* Sy */ - {NULL, NULL}, /* Tn */ - {mdoc_xx_pre, 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 */ - {mdoc_xx_pre, NULL}, /* Dx */ - {NULL, NULL}, /* %Q */ - {NULL, NULL}, /* br */ - {NULL, NULL}, /* sp */ -}; - -static char buf[BUFSIZ]; /* XXX */ - -#define bufcat(x) (void)strlcat(buf, (x), BUFSIZ) -#define bufinit() buf[0] = 0 -#define buffmt(...) (void)snprintf(buf, BUFSIZ - 1, __VA_ARGS__) - -void -html_mdoc(void *arg, const struct mdoc *m) +void * +html_alloc(const struct manoutput *outopts) { - struct html *h; - struct tag *t; - - 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_man(void *arg, const struct man *m) -{ struct html *h; - struct tag *t; - h = (struct html *)arg; + h = mandoc_calloc(1, sizeof(struct html)); - 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); + 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; - printf("\n"); -} + mandoc_ohash_init(&id_unique, 4, offsetof(struct id_entry, id)); - -void * -html_alloc(void) -{ - struct html *h; - - if (NULL == (h = calloc(1, sizeof(struct html)))) - return(NULL); - - SLIST_INIT(&h->stack); - if (NULL == (h->symtab = chars_init(CHARS_HTML))) { - free(h); - return(NULL); - } - return(h); + return h; } - -void -html_free(void *p) +static void +html_reset_internal(struct html *h) { struct tag *tag; - struct html *h; + struct id_entry *entry; + unsigned int slot; - h = (struct html *)p; - - while ( ! SLIST_EMPTY(&h->stack)) { - tag = SLIST_FIRST(&h->stack); - SLIST_REMOVE_HEAD(&h->stack, entry); + while ((tag = h->tag) != NULL) { + h->tag = tag->next; free(tag); } - free(h); + entry = ohash_first(&id_unique, &slot); + while (entry != NULL) { + free(entry); + entry = ohash_next(&id_unique, &slot); + } + ohash_delete(&id_unique); } - -static void -print_mdoc(MDOC_ARGS) +void +html_reset(void *p) { - struct tag *t; - - t = print_otag(h, TAG_HEAD, 0, NULL); - print_mdoc_head(m, n, h); - print_tagq(h, t); - - t = print_otag(h, TAG_BODY, 0, NULL); - print_mdoc_title(m, n, h); - print_mdoc_node(m, n, h); - print_tagq(h, t); + html_reset_internal(p); + mandoc_ohash_init(&id_unique, 4, offsetof(struct id_entry, id)); } - -static void -print_gen_head(struct html *h) +void +html_free(void *p) { - struct htmlpair meta0[2]; - struct htmlpair meta1[2]; - struct htmlpair link[4]; - - meta0[0].key = ATTR_HTTPEQUIV; - meta0[0].val = "Content-Type"; - meta0[1].key = ATTR_CONTENT; - meta0[1].val = "text/html; charset=utf-8"; - - meta1[0].key = ATTR_NAME; - meta1[0].val = "resource-type"; - meta1[1].key = ATTR_CONTENT; - meta1[1].val = "document"; - - 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"; - - print_otag(h, TAG_META, 2, meta0); - print_otag(h, TAG_META, 2, meta1); - print_otag(h, TAG_LINK, 4, link); + html_reset_internal(p); + free(p); } - -/* ARGSUSED */ -static void -print_mdoc_head(MDOC_ARGS) +void +print_gen_head(struct html *h) { + struct tag *t; - print_gen_head(h); - print_otag(h, TAG_TITLE, 0, NULL); - print_encode(h, m->title); -} + 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 a minimal embedded style sheet. + */ -/* ARGSUSED */ -static void -print_mdoc_title(MDOC_ARGS) -{ - - /* TODO */ + 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); } - -static void -print_mdoc_node(MDOC_ARGS) +int +html_setfont(struct html *h, enum mandoc_esc font) { - int child; - struct tag *t; - - child = 1; - t = SLIST_FIRST(&h->stack); - - bufinit(); - - switch (n->type) { - case (MDOC_ROOT): - child = mdoc_root_pre(m, n, h); + switch (font) { + case ESCAPE_FONTPREV: + font = h->metal; break; - case (MDOC_TEXT): - print_text(h, n->string); + case ESCAPE_FONTITALIC: + case ESCAPE_FONTBOLD: + case ESCAPE_FONTBI: + case ESCAPE_FONTCW: + case ESCAPE_FONTROMAN: break; - default: - if (mdocs[n->tok].pre) - child = (*mdocs[n->tok].pre)(m, n, h); + case ESCAPE_FONT: + font = ESCAPE_FONTROMAN; break; + default: + return 0; } + h->metal = h->metac; + h->metac = font; + return 1; +} - if (child && n->child) - print_mdoc_node(m, n->child, h); - - print_stagq(h, t); - - bufinit(); - - switch (n->type) { - case (MDOC_ROOT): +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 (MDOC_TEXT): + 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: - if (mdocs[n->tok].post) - (*mdocs[n->tok].post)(m, n, h); break; } - - if (n->next) - print_mdoc_node(m, n->next, h); } +void +html_close_paragraph(struct html *h) +{ + struct tag *this, *next; + int flags; -static void -print_man(MAN_ARGS) + 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; + } +} + +/* + * 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_man_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_man_body(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; } - -/* ARGSUSED */ -static void -print_man_head(MAN_ARGS) +/* + * Allocate a string to be used for the "id=" attribute of an HTML + * element and/or as a segment identifier for a URI in an element. + * The function may fail and return NULL if the node lacks text data + * to create the attribute from. + * The caller is responsible for free(3)ing the returned string. + * + * If the "unique" argument is non-zero, the "id_unique" ohash table + * is used for de-duplication. If the "unique" argument is 1, + * it is the first time the function is called for this tag and + * location, so if an ordinal suffix is needed, it is incremented. + * If the "unique" argument is 2, it is the second time the function + * is called for this tag and location, so the ordinal suffix + * remains unchanged. + */ +char * +html_make_id(const struct roff_node *n, int unique) { + const struct roff_node *nch; + struct id_entry *entry; + char *buf, *cp; + size_t len; + unsigned int slot; - print_gen_head(h); - print_otag(h, TAG_TITLE, 0, NULL); - print_encode(h, m->title); -} + if (n->tag != NULL) + buf = mandoc_strdup(n->tag); + else { + switch (n->tok) { + case MDOC_Sh: + case MDOC_Ss: + case MDOC_Sx: + case MAN_SH: + case MAN_SS: + for (nch = n->child; nch != NULL; nch = nch->next) + if (nch->type != ROFFT_TEXT) + return NULL; + buf = NULL; + deroff(&buf, n); + if (buf == NULL) + return NULL; + break; + default: + if (n->child == NULL || n->child->type != ROFFT_TEXT) + return NULL; + buf = mandoc_strdup(n->child->string); + break; + } + } + /* + * 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 + * In addition, reserve '~' for ordinal suffixes. + */ -/* ARGSUSED */ -static void -print_man_body(MAN_ARGS) -{ + for (cp = buf; *cp != '\0'; cp++) + if (isalnum((unsigned char)*cp) == 0 && + strchr("!$&'()*+,-./:;=?@_", *cp) == NULL) + *cp = '_'; - /* TODO */ -} + if (unique == 0) + return buf; + /* Avoid duplicate HTML id= attributes. */ -static void -print_spec(struct html *h, const char *p, int len) -{ - const char *rhs; - int i; - size_t sz; + slot = ohash_qlookup(&id_unique, buf); + if ((entry = ohash_find(&id_unique, slot)) == NULL) { + len = strlen(buf) + 1; + entry = mandoc_malloc(sizeof(*entry) + len); + entry->ord = 1; + memcpy(entry->id, buf, len); + ohash_insert(&id_unique, slot, entry); + } else if (unique == 1) + entry->ord++; - rhs = chars_a2ascii(h->symtab, p, (size_t)len, &sz); - - if (NULL == rhs) - return; - for (i = 0; i < (int)sz; i++) - putchar(rhs[i]); + if (entry->ord > 1) { + cp = buf; + mandoc_asprintf(&buf, "%s~%d", cp, entry->ord); + free(cp); + } + return buf; } - -static void -print_res(struct html *h, const char *p, int len) +static int +print_escape(struct html *h, char c) { - const char *rhs; - int i; - size_t sz; - rhs = chars_a2res(h->symtab, p, (size_t)len, &sz); - - if (NULL == rhs) - return; - for (i = 0; i < (int)sz; i++) - putchar(rhs[i]); + switch (c) { + case '<': + print_word(h, "<"); + break; + case '>': + print_word(h, ">"); + break; + case '&': + print_word(h, "&"); + break; + case '"': + print_word(h, """); + break; + case ASCII_NBRSP: + print_word(h, " "); + break; + case ASCII_HYPH: + print_byte(h, '-'); + break; + case ASCII_BREAK: + break; + default: + return 0; + } + return 1; } - -static void -print_escape(struct html *h, const char **p) +static int +print_encode(struct html *h, const char *p, const char *pend, int norecurse) { - int j, type; - const char *wp; + 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' }; - wp = *p; - type = 1; + if (pend == NULL) + pend = strchr(p, '\0'); - if (0 == *(++wp)) { - *p = wp; - return; - } + breakline = 0; + nospace = 0; - if ('(' == *wp) { - wp++; - if (0 == *wp || 0 == *(wp + 1)) { - *p = 0 == *wp ? wp : wp + 1; - return; + while (p < pend) { + if (HTML_SKIPCHAR & h->flags && '\\' != *p) { + h->flags &= ~HTML_SKIPCHAR; + p++; + continue; } - print_spec(h, wp, 2); - *p = ++wp; - return; + for (sz = strcspn(p, rejs); sz-- && p < pend; p++) + print_byte(h, *p); - } else if ('*' == *wp) { - if (0 == *(++wp)) { - *p = wp; - return; + if (breakline && + (p >= pend || *p == ' ' || *p == ASCII_NBRSP)) { + print_otag(h, TAG_BR, ""); + breakline = 0; + while (p < pend && (*p == ' ' || *p == ASCII_NBRSP)) + p++; + continue; } - switch (*wp) { - case ('('): - wp++; - if (0 == *wp || 0 == *(wp + 1)) { - *p = 0 == *wp ? wp : wp + 1; - return; - } - - print_res(h, wp, 2); - *p = ++wp; - return; - case ('['): - type = 0; + if (p >= pend) break; + + if (*p == ' ') { + print_endword(h); + p++; + continue; + } + + 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: - print_res(h, wp, 1); - *p = wp; - return; + break; } - - } else if ('f' == *wp) { - if (0 == *(++wp)) { - *p = wp; - return; + + if (h->flags & HTML_SKIPCHAR) { + h->flags &= ~HTML_SKIPCHAR; + continue; } - switch (*wp) { - case ('B'): - /* TODO */ + switch (esc) { + case ESCAPE_UNICODE: + /* Skip past "u" header. */ + c = mchars_num2uc(seq + 1, len - 1); break; - case ('I'): - /* TODO */ + case ESCAPE_NUMBERED: + c = mchars_num2char(seq, len); + if (c < 0) + continue; break; - case ('P'): - /* FALLTHROUGH */ - case ('R'): - /* TODO */ + case ESCAPE_SPECIAL: + c = mchars_spec2cp(seq, len); + if (c <= 0) + continue; break; - default: + 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; } - - *p = wp; - return; - - } else if ('[' != *wp) { - print_spec(h, wp, 1); - *p = wp; - return; + 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); } - wp++; - for (j = 0; *wp && ']' != *wp; wp++, j++) - /* Loop... */ ; - - if (0 == *wp) { - *p = wp; - return; - } - - if (type) - print_spec(h, wp - j, j); - else - print_res(h, wp - j, j); - - *p = wp; + return nospace; } - static void -print_encode(struct html *h, 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; - for (; *p; p++) { - if ('\\' == *p) { - print_escape(h, &p); - continue; + 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); } - switch (*p) { - case ('<'): - printf("<"); - break; - case ('>'): - printf(">"); - break; - case ('&'): - printf("&"); - break; - default: - putchar(*p); - break; - } + } 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(h, 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; - - return(t); -} - - -/* 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; -} + /* Print out the tag name and attributes. */ -/* ARGSUSED */ -static void -print_gen_doctype(struct html *h) -{ - - printf("\n", DOCTYPE, DTD); -} + print_byte(h, '<'); + print_word(h, htmltags[tag].name); + va_start(ap, fmt); -static void -print_text(struct html *h, const char *p) -{ + while (*fmt != '\0' && *fmt != 's') { - 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; + /* Parse attributes and arguments. */ + + arg1 = va_arg(ap, char *); + arg2 = NULL; + switch (*fmt++) { + case 'c': + attr = "class"; break; - default: + 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; - if ( ! (h->flags & HTML_NOSPACE)) - printf(" "); + /* Print the attributes. */ - h->flags &= ~HTML_NOSPACE; - h->flags &= ~HTML_NEWLINE; - - if (p) - print_encode(h, p); - - if (*p && 0 == *(p + 1)) - switch (*p) { - 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; } -} - - -static void -print_tagq(struct html *h, const struct tag *until) -{ - struct tag *tag; - - 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; + print_byte(h, '"'); } -} - -static void -print_stagq(struct html *h, const struct tag *suntil) -{ - struct tag *tag; - - 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); - free(tag); + 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, '"'); + va_end(ap); -/* FIXME: put in utility file for front-ends. */ -static int -a2offs(const char *p) -{ - int len, i; + /* Accommodate for "well-formed" singleton escaping. */ - if (0 == strcmp(p, "left")) - return(0); - if (0 == strcmp(p, "indent")) - return(INDENT + 1); - if (0 == strcmp(p, "indent-two")) - return((INDENT + 1) * 2); + if (htmltags[tag].flags & HTML_NOSTACK) + print_byte(h, '/'); - if (0 == (len = (int)strlen(p))) - return(0); + print_byte(h, '>'); - for (i = 0; i < len - 1; i++) - if ( ! isdigit((u_char)p[i])) - break; + if (tflags & HTML_NLBEGIN) + print_endline(h); + else + h->flags |= HTML_NOSPACE; - if (i == len - 1) - if ('n' == p[len - 1] || 'm' == p[len - 1]) - return(atoi(p)); + if (tflags & HTML_INDENT) + h->indent++; + if (tflags & HTML_NOINDENT) + h->noindent++; - return(len); + return t; } - -/* FIXME: put in utility file for front-ends. */ -static int -a2width(const char *p) +/* + * Print an element with an optional "id=" attribute. + * If the element has phrasing content and an "id=" attribute, + * also add a permalink: outside if it can be in phrasing context, + * inside otherwise. + */ +struct tag * +print_otag_id(struct html *h, enum htmltag elemtype, const char *cattr, + struct roff_node *n) { - int i, len; + struct roff_node *nch; + struct tag *ret, *t; + char *id, *href; - if (0 == (len = (int)strlen(p))) - return(0); - for (i = 0; i < len - 1; i++) - if ( ! isdigit((u_char)p[i])) - break; - - if (i == len - 1) - if ('n' == p[len - 1] || 'm' == p[len - 1]) - return(atoi(p) + 2); - - return(len + 2); -} - - - - -/* ARGSUSED */ -static int -mdoc_root_pre(MDOC_ARGS) -{ - struct htmlpair tag; - - tag.key = ATTR_CLASS; - tag.val = "body"; - - print_otag(h, TAG_DIV, 1, &tag); - return(1); -} - - -/* ARGSUSED */ -static int -mdoc_sh_pre(MDOC_ARGS) -{ - struct htmlpair tag[2]; - - if (MDOC_HEAD == n->type) { - tag[0].key = ATTR_CLASS; - tag[0].val = "sec-head"; - print_otag(h, TAG_DIV, 1, tag); - print_otag(h, TAG_SPAN, 1, tag); - return(1); - } else if (MDOC_BLOCK == n->type) { - tag[0].key = ATTR_CLASS; - tag[0].val = "sec-block"; - print_otag(h, TAG_DIV, 1, tag); - return(1); - } - - buffmt("margin-left: %dem;", INDENT); - - if (n->parent->next && n->child) - bufcat("margin-bottom: 1em;"); - - tag[0].key = ATTR_CLASS; - tag[0].val = "sec-body"; - tag[1].key = ATTR_STYLE; - tag[1].val = buf; - - print_otag(h, TAG_DIV, 2, tag); - return(1); -} - - -/* ARGSUSED */ -static int -mdoc_ss_pre(MDOC_ARGS) -{ - struct htmlpair tag[2]; - int i; - - i = 0; - - if (MDOC_BODY == n->type) { - tag[i].key = ATTR_CLASS; - tag[i++].val = "ssec-body"; - if (n->parent->next && n->child) { - bufcat("margin-bottom: 1em;"); - tag[i].key = ATTR_STYLE; - tag[i++].val = buf; + ret = NULL; + id = href = NULL; + if (n->flags & NODE_ID) + id = html_make_id(n, 1); + if (n->flags & NODE_HREF) + href = id == NULL ? html_make_id(n, 2) : id; + if (href != NULL && htmltags[elemtype].flags & HTML_INPHRASE) + ret = print_otag(h, TAG_A, "chR", "permalink", href); + t = print_otag(h, elemtype, "ci", cattr, id); + if (ret == NULL) { + ret = t; + if (href != NULL && (nch = n->child) != NULL) { + /* man(7) is safe, it tags phrasing content only. */ + if (n->tok > MDOC_MAX || + htmltags[elemtype].flags & HTML_TOPHRASE) + nch = NULL; + else /* For mdoc(7), beware of nested blocks. */ + while (nch != NULL && nch->type == ROFFT_TEXT) + nch = nch->next; + if (nch == NULL) + print_otag(h, TAG_A, "chR", "permalink", href); } - print_otag(h, TAG_DIV, i, tag); - return(1); - } else if (MDOC_BLOCK == n->type) { - tag[i].key = ATTR_CLASS; - tag[i++].val = "ssec-block"; - if (n->prev) { - bufcat("margin-top: 1em;"); - tag[i].key = ATTR_STYLE; - tag[i++].val = buf; - } - print_otag(h, TAG_DIV, i, tag); - return(1); } - - buffmt("margin-left: -%dem;", INDENT - HALFINDENT); - - tag[0].key = ATTR_CLASS; - tag[0].val = "ssec-head"; - tag[1].key = ATTR_STYLE; - tag[1].val = buf; - - print_otag(h, TAG_DIV, 2, tag); - print_otag(h, TAG_SPAN, 1, tag); - return(1); + free(id); + if (id == NULL) + free(href); + return ret; } - -/* ARGSUSED */ -static int -mdoc_fl_pre(MDOC_ARGS) -{ - struct htmlpair tag; - - tag.key = ATTR_CLASS; - tag.val = "flag"; - - print_otag(h, TAG_SPAN, 1, &tag); - print_text(h, "\\-"); - h->flags |= HTML_NOSPACE; - return(1); -} - - -/* ARGSUSED */ -static int -mdoc_pp_pre(MDOC_ARGS) -{ - struct htmlpair tag; - - bufcat("clear: both;"); - bufcat("height: 1em;"); - - tag.key = ATTR_STYLE; - tag.val = buf; - - print_otag(h, TAG_DIV, 1, &tag); - return(0); -} - - -/* ARGSUSED */ -static int -mdoc_nd_pre(MDOC_ARGS) -{ - struct htmlpair tag; - - if (MDOC_BODY != n->type) - return(1); - - /* XXX - this can contain block elements! */ - print_text(h, "\\(em"); - tag.key = ATTR_CLASS; - tag.val = "desc-body"; - print_otag(h, TAG_SPAN, 1, &tag); - return(1); -} - - -/* ARGSUSED */ -static int -mdoc_op_pre(MDOC_ARGS) -{ - struct htmlpair tag; - - if (MDOC_BODY != n->type) - return(1); - - /* XXX - this can contain block elements! */ - print_text(h, "\\(lB"); - tag.key = ATTR_CLASS; - tag.val = "opt"; - print_otag(h, TAG_SPAN, 1, &tag); - return(1); -} - - -/* ARGSUSED */ static void -mdoc_op_post(MDOC_ARGS) +print_ctag(struct html *h, struct tag *tag) { + int tflags; - if (MDOC_BODY != n->type) - return; - h->flags |= HTML_NOSPACE; - print_text(h, "\\(rB"); -} + if (tag->closed == 0) { + tag->closed = 1; + if (tag == h->metaf) + h->metaf = NULL; + if (tag == h->tblt) + h->tblt = NULL; - -static int -mdoc_nm_pre(MDOC_ARGS) -{ - struct htmlpair tag; - - if ( ! (HTML_NEWLINE & h->flags)) - if (SEC_SYNOPSIS == n->sec) { - tag.key = ATTR_STYLE; - tag.val = "clear: both;"; - print_otag(h, TAG_BR, 1, &tag); - } - - tag.key = ATTR_CLASS; - tag.val = "name"; - - print_otag(h, TAG_SPAN, 1, &tag); - if (NULL == n->child) - print_text(h, m->name); - - return(1); -} - - -/* ARGSUSED */ -static int -mdoc_xr_pre(MDOC_ARGS) -{ - struct htmlpair tag[2]; - - tag[0].key = ATTR_CLASS; - tag[0].val = "link-man"; - tag[1].key = ATTR_HREF; - tag[1].val = "#"; /* TODO */ - - print_otag(h, TAG_A, 2, 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); -} - - -/* ARGSUSED */ -static int -mdoc_ns_pre(MDOC_ARGS) -{ - - 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); -} - - -/* ARGSUSED */ -static int -mdoc_xx_pre(MDOC_ARGS) -{ - const char *pp; - struct htmlpair tag; - - switch (n->tok) { - case (MDOC_Bsx): - pp = "BSDI BSD/OS"; - break; - case (MDOC_Dx): - pp = "DragonFlyBSD"; - break; - case (MDOC_Fx): - pp = "FreeBSD"; - break; - case (MDOC_Nx): - pp = "NetBSD"; - break; - case (MDOC_Ox): - pp = "OpenBSD"; - break; - case (MDOC_Ux): - pp = "UNIX"; - break; - default: - return(1); + 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); } - - tag.key = ATTR_CLASS; - tag.val = "unix"; - - print_otag(h, TAG_SPAN, 1, &tag); - print_text(h, pp); - return(1); + if (tag->refcnt == 0) { + h->tag = tag->next; + free(tag); + } } - -/* ARGSUSED */ -static int -mdoc_tbl_block_pre(MDOC_ARGS, int w, int o, int c) +void +print_gen_decls(struct html *h) { - struct htmlpair tag; - - buffmt("margin-left: %dem; clear: both;", w + o); - - if ( ! c) - bufcat("padding-top: 1em;"); - - tag.key = ATTR_STYLE; - tag.val = buf; - print_otag(h, TAG_DIV, 1, &tag); - return(1); + print_word(h, ""); + print_endline(h); } - -/* ARGSUSED */ -static int -mdoc_tbl_body_pre(MDOC_ARGS, int t, int w) +void +print_gen_comment(struct html *h, struct roff_node *n) { - struct htmlpair tag; - int i; + int wantblank; - switch (t) { - case (MDOC_Tag): - i = 1; - tag.key = ATTR_STYLE; - tag.val = buf; - bufcat("clear: right;"); - bufcat("float: left;"); - bufcat("width: 100%%;"); - break; - default: - i = 0; - break; + 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; } - - print_otag(h, TAG_DIV, i, &tag); - return(1); + if (wantblank) + print_endline(h); + print_word(h, " -->"); + print_endline(h); + h->indent = 0; } - -/* ARGSUSED */ -static int -mdoc_tbl_head_pre(MDOC_ARGS, int t, int w) +void +print_text(struct html *h, const char *word) { - struct htmlpair tag; - int i; - - switch (t) { - case (MDOC_Hang): - /* FALLTHROUGH */ - case (MDOC_Tag): - i = 1; - tag.key = ATTR_STYLE; - tag.val = buf; - buffmt("margin-left: -%dem;", w); - bufcat("clear: left;"); - bufcat("float: left;"); - bufcat("padding-right: 1em;"); - break; - default: - i = 0; - break; - } - - print_otag(h, TAG_DIV, i, &tag); - return(1); + print_tagged_text(h, word, NULL); } - -static int -mdoc_tbl_pre(MDOC_ARGS, int type) +void +print_tagged_text(struct html *h, const char *word, struct roff_node *n) { - int i, w, o, c; - const struct mdoc_node *bl; + struct tag *t; + char *href; - bl = n->parent->parent; - if (MDOC_BLOCK != n->type) - bl = bl->parent; + /* + * Always wrap text in a paragraph unless already contained in + * some flow container; never put it directly into a section. + */ - /* FIXME: fmt_vspace() equivalent. */ + if (h->tag->tag == TAG_SECTION) + print_otag(h, TAG_P, "c", "Pp"); - assert(bl->args); + /* Output whitespace before this text? */ - w = o = c = 0; - - for (i = 0; i < (int)bl->args->argc; i++) - if (MDOC_Width == bl->args->argv[i].arg) { - assert(bl->args->argv[i].sz); - w = a2width(bl->args->argv[i].value[0]); - } else if (MDOC_Offset == bl->args->argv[i].arg) { - assert(bl->args->argv[i].sz); - o = a2offs(bl->args->argv[i].value[0]); - } else if (MDOC_Compact == bl->args->argv[i].arg) - c = 1; - - if (0 == w) - w = 10; - - switch (n->type) { - case (MDOC_BLOCK): - break; - case (MDOC_HEAD): - return(mdoc_tbl_head_pre(m, n, h, type, w)); - case (MDOC_BODY): - return(mdoc_tbl_body_pre(m, n, h, type, w)); - default: - abort(); - /* NOTREACHED */ + 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, " "); } - return(mdoc_tbl_block_pre(m, n, h, w, o, c)); -} + /* + * Optionally switch fonts, optionally write a permalink, then + * print the text, optionally surrounded by HTML whitespace. + */ + assert(h->metaf == NULL); + print_metaf(h); + print_indent(h); -/* ARGSUSED */ -static int -mdoc_listitem_pre(MDOC_ARGS) -{ - int i, w, o, c; - const struct mdoc_node *bl; - struct htmlpair tag; + if (n != NULL && (href = html_make_id(n, 2)) != NULL) { + t = print_otag(h, TAG_A, "chR", "permalink", href); + free(href); + } else + t = NULL; - /* FIXME: fmt_vspace() equivalent. */ + 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 (MDOC_BLOCK != n->type) - return(1); + if (h->metaf != NULL) { + print_tagq(h, h->metaf); + h->metaf = NULL; + } else if (t != NULL) + print_tagq(h, t); - bl = n->parent->parent; - assert(bl); - - w = o = c = 0; - - for (i = 0; i < (int)bl->args->argc; i++) - if (MDOC_Width == bl->args->argv[i].arg) { - assert(bl->args->argv[i].sz); - w = a2width(bl->args->argv[i].value[0]); - } else if (MDOC_Offset == bl->args->argv[i].arg) { - assert(bl->args->argv[i].sz); - o = a2offs(bl->args->argv[i].value[0]); - } else if (MDOC_Compact == bl->args->argv[i].arg) - c = 1; - - if (o && w) - buffmt("margin-left: %dem; padding-left: %dem;", o, w); - else if (w) - buffmt("padding-left: %dem;", w > 4 ? w - 4 : w); - else if (o) - buffmt("margin-left: %dem;", o); - - if ( ! c) - bufcat("padding-top: 1em;"); - - tag.key = ATTR_STYLE; - tag.val = buf; - - print_otag(h, TAG_LI, 1, &tag); - return(1); + h->flags &= ~HTML_IGNDELIM; } - -/* ARGSUSED */ -static int -mdoc_list_pre(MDOC_ARGS, int type) +void +print_tagq(struct html *h, const struct tag *until) { + struct tag *this, *next; - switch (type) { - case (MDOC_Enum): - print_otag(h, TAG_OL, 0, NULL); - break; - case (MDOC_Bullet): - print_otag(h, TAG_UL, 0, NULL); - break; - default: - break; + for (this = h->tag; this != NULL; this = next) { + next = this == until ? NULL : this->next; + print_ctag(h, this); } - - return(1); } - -static int -mdoc_bl_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) { - int i, len, type; + struct tag *this, *next; - if (MDOC_BLOCK != n->type) - return(1); - - assert(n->args); - len = (int)n->args->argc; - - for (i = 0; i < len; i++) - switch ((type = n->args->argv[i].arg)) { - case (MDOC_Enum): - /* FALLTHROUGH */ - case (MDOC_Bullet): - return(mdoc_list_pre(m, n, h, type)); - case (MDOC_Tag): - /* FALLTHROUGH */ - case (MDOC_Hang): - /* FALLTHROUGH */ - case (MDOC_Dash): - /* FALLTHROUGH */ - case (MDOC_Hyphen): - /* FALLTHROUGH */ - case (MDOC_Inset): - /* FALLTHROUGH */ - case (MDOC_Diag): - /* FALLTHROUGH */ - case (MDOC_Item): - /* FALLTHROUGH */ - case (MDOC_Column): - /* FALLTHROUGH */ - case (MDOC_Ohang): - return(1); - default: + 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; - } - - abort(); - /* NOTREACHED */ + print_ctag(h, this); + } } -static int -mdoc_it_pre(MDOC_ARGS) -{ - int i, len, type; - const struct mdoc_node *bl; +/*********************************************************************** + * Low level output functions. + * They implement line breaking using a short static buffer. + ***********************************************************************/ - if (MDOC_BLOCK == n->type) - bl = n->parent->parent; - else - bl = n->parent->parent->parent; - - assert(bl->args); - len = (int)bl->args->argc; - - for (i = 0; i < len; i++) - switch ((type = bl->args->argv[i].arg)) { - case (MDOC_Tag): - /* FALLTHROUGH */ - case (MDOC_Hang): - return(mdoc_tbl_pre(m, n, h, type)); - case (MDOC_Enum): - /* FALLTHROUGH */ - case (MDOC_Bullet): - return(mdoc_listitem_pre(m, n, h)); - case (MDOC_Dash): - /* FALLTHROUGH */ - case (MDOC_Hyphen): - /* FALLTHROUGH */ - case (MDOC_Inset): - /* FALLTHROUGH */ - case (MDOC_Diag): - /* FALLTHROUGH */ - case (MDOC_Item): - /* FALLTHROUGH */ - case (MDOC_Column): - /* FALLTHROUGH */ - case (MDOC_Ohang): - return(0); - default: - break; - } - - abort(); - /* NOTREACHED */ -} - - -/* ARGSUSED */ -static int -mdoc_ex_pre(MDOC_ARGS) +/* + * 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 +print_byte(struct html *h, char c) { - const struct mdoc_node *nn; - struct tag *t; - struct htmlpair tag; + if ((h->flags & HTML_BUFFER) == 0) { + putchar(c); + h->col++; + return; + } - print_text(h, "The"); - - tag.key = ATTR_CLASS; - tag.val = "utility"; - - for (nn = n->child; nn; nn = nn->next) { - t = print_otag(h, TAG_SPAN, 1, &tag); - print_text(h, nn->string); - print_tagq(h, t); - - h->flags |= HTML_NOSPACE; - - if (nn->next && NULL == nn->next->next) - print_text(h, ", and"); - else if (nn->next) - print_text(h, ","); - else - h->flags &= ~HTML_NOSPACE; + if (h->col + h->bufcol < sizeof(h->buf)) { + h->buf[h->bufcol++] = c; + return; } - if (n->child->next) - print_text(h, "utilities exit"); - else - print_text(h, "utility exits"); - - print_text(h, "0 on success, and >0 if an error occurs."); - return(0); + 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; } - -/* ARGSUSED */ -static int -mdoc_dq_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) { + if (h->col == 0) + return; - if (MDOC_BODY != n->type) - return(1); - print_text(h, "\\(lq"); + if (h->bufcol) { + putchar(' '); + fwrite(h->buf, h->bufcol, 1, stdout); + h->bufcol = 0; + } + putchar('\n'); + h->col = 0; h->flags |= HTML_NOSPACE; - return(1); + h->flags &= ~HTML_BUFFER; } - -/* ARGSUSED */ +/* + * Flush the HTML output buffer. + * If it is inactive, activate it. + */ static void -mdoc_dq_post(MDOC_ARGS) +print_endword(struct html *h) { - - if (MDOC_BODY != n->type) + if (h->noindent) { + print_byte(h, ' '); return; - h->flags |= HTML_NOSPACE; - print_text(h, "\\(rq"); -} + } - -/* ARGSUSED */ -static int -mdoc_pq_pre(MDOC_ARGS) -{ - - if (MDOC_BODY != n->type) - return(1); - print_text(h, "\\&("); - h->flags |= HTML_NOSPACE; - 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 */ +/* + * 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 -mdoc_pq_post(MDOC_ARGS) +print_indent(struct html *h) { + size_t i; - if (MDOC_BODY != n->type) + if (h->col || h->noindent) return; - print_text(h, ")"); -} - -/* ARGSUSED */ -static int -mdoc_sq_pre(MDOC_ARGS) -{ - - if (MDOC_BODY != n->type) - return(1); - print_text(h, "\\(oq"); - h->flags |= HTML_NOSPACE; - return(1); + h->col = h->indent * 2; + for (i = 0; i < h->col; i++) + putchar(' '); } - -/* ARGSUSED */ +/* + * Print or buffer some characters + * depending on the current HTML output buffer state. + */ static void -mdoc_sq_post(MDOC_ARGS) +print_word(struct html *h, const char *cp) { - - if (MDOC_BODY != n->type) - return; - h->flags |= HTML_NOSPACE; - print_text(h, "\\(aq"); -} - - -/* ARGSUSED */ -static int -mdoc_em_pre(MDOC_ARGS) -{ - struct htmlpair tag; - - tag.key = ATTR_CLASS; - tag.val = "emph"; - - print_otag(h, TAG_SPAN, 1, &tag); - return(1); -} - - -/* ARGSUSED */ -static int -mdoc_d1_pre(MDOC_ARGS) -{ - struct htmlpair tag[2]; - - if (MDOC_BLOCK != n->type) - return(1); - - buffmt("margin-left: %dem;", INDENT); - bufcat("margin-bottom: 0px;"); - - tag[0].key = ATTR_CLASS; - tag[0].val = "lit-block"; - tag[1].key = ATTR_STYLE; - tag[1].val = buf; - - print_otag(h, TAG_DIV, 2, tag); - return(1); -} - - -/* ARGSUSED */ -static int -mdoc_sx_pre(MDOC_ARGS) -{ - struct htmlpair tag[2]; - - tag[0].key = ATTR_HREF; - tag[0].val = "#"; /* XXX */ - tag[1].key = ATTR_CLASS; - tag[1].val = "link-sec"; - - print_otag(h, TAG_A, 2, tag); - return(1); + while (*cp != '\0') + print_byte(h, *cp++); }