=================================================================== RCS file: /cvs/mandoc/html.c,v retrieving revision 1.117 retrieving revision 1.231 diff -u -p -r1.117 -r1.231 --- mandoc/html.c 2010/12/17 08:17:40 1.117 +++ mandoc/html.c 2018/05/29 02:10:10 1.231 @@ -1,420 +1,594 @@ -/* $Id: html.c,v 1.117 2010/12/17 08:17:40 kristaps Exp $ */ +/* $Id: html.c,v 1.231 2018/05/29 02:10:10 schwarze Exp $ */ /* - * Copyright (c) 2008, 2009, 2010 Kristaps Dzonsons + * Copyright (c) 2008-2011, 2014 Kristaps Dzonsons + * Copyright (c) 2011-2015, 2017, 2018 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. */ -#ifdef HAVE_CONFIG_H #include "config.h" -#endif #include #include #include #include +#include #include #include #include #include #include +#include "mandoc_aux.h" +#include "mandoc_ohash.h" #include "mandoc.h" +#include "roff.h" #include "out.h" -#include "chars.h" #include "html.h" +#include "manconf.h" #include "main.h" struct htmldata { const char *name; int flags; -#define HTML_CLRLINE (1 << 0) -#define HTML_NOSTACK (1 << 1) -#define HTML_AUTOCLOSE (1 << 2) /* Tag has auto-closure. */ +#define HTML_NOSTACK (1 << 0) +#define HTML_AUTOCLOSE (1 << 1) +#define HTML_NLBEFORE (1 << 2) +#define HTML_NLBEGIN (1 << 3) +#define HTML_NLEND (1 << 4) +#define HTML_NLAFTER (1 << 5) +#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 << 6) +#define HTML_NOINDENT (1 << 7) }; 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 | HTML_AUTOCLOSE}, /* TAG_META */ - {"title", HTML_CLRLINE}, /* TAG_TITLE */ - {"div", HTML_CLRLINE}, /* TAG_DIV */ - {"h1", 0}, /* TAG_H1 */ - {"h2", 0}, /* TAG_H2 */ - {"span", 0}, /* TAG_SPAN */ - {"link", HTML_CLRLINE | HTML_NOSTACK | HTML_AUTOCLOSE}, /* TAG_LINK */ - {"br", HTML_CLRLINE | HTML_NOSTACK | HTML_AUTOCLOSE}, /* TAG_BR */ - {"a", 0}, /* TAG_A */ - {"table", HTML_CLRLINE}, /* TAG_TABLE */ - {"tbody", HTML_CLRLINE}, /* TAG_TBODY */ - {"col", HTML_CLRLINE | HTML_NOSTACK | HTML_AUTOCLOSE}, /* 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 */ - {"dl", HTML_CLRLINE}, /* TAG_DL */ - {"dt", HTML_CLRLINE}, /* TAG_DT */ - {"dd", HTML_CLRLINE}, /* TAG_DD */ - {"blockquote", HTML_CLRLINE}, /* TAG_BLOCKQUOTE */ - {"p", HTML_CLRLINE | HTML_NOSTACK | HTML_AUTOCLOSE}, /* TAG_P */ + {"html", HTML_NLALL}, + {"head", HTML_NLALL | HTML_INDENT}, + {"body", HTML_NLALL}, + {"meta", HTML_NOSTACK | HTML_AUTOCLOSE | HTML_NLALL}, + {"title", HTML_NLAROUND}, + {"div", HTML_NLAROUND}, + {"div", 0}, + {"h1", HTML_NLAROUND}, + {"h2", HTML_NLAROUND}, + {"span", 0}, + {"link", HTML_NOSTACK | HTML_AUTOCLOSE | HTML_NLALL}, + {"br", HTML_NOSTACK | HTML_AUTOCLOSE | HTML_NLALL}, + {"a", 0}, + {"table", HTML_NLALL | HTML_INDENT}, + {"colgroup", HTML_NLALL | HTML_INDENT}, + {"col", HTML_NOSTACK | HTML_AUTOCLOSE | HTML_NLALL}, + {"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}, + {"pre", HTML_NLALL | HTML_NOINDENT}, + {"var", 0}, + {"cite", 0}, + {"b", 0}, + {"i", 0}, + {"code", 0}, + {"small", 0}, + {"style", HTML_NLALL | HTML_INDENT}, + {"math", 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 htmlfonts[HTMLFONT_MAX] = { - "roman", - "bold", - "italic" +static const char *const roffscales[SCALE_MAX] = { + "cm", /* SCALE_CM */ + "in", /* SCALE_IN */ + "pc", /* SCALE_PC */ + "pt", /* SCALE_PT */ + "em", /* SCALE_EM */ + "em", /* SCALE_MM */ + "ex", /* SCALE_EN */ + "ex", /* SCALE_BU */ + "em", /* SCALE_VS */ + "ex", /* SCALE_FS */ }; -static const char *const htmlattrs[ATTR_MAX] = { - "http-equiv", - "content", - "name", - "rel", - "href", - "type", - "media", - "class", - "style", - "width", - "valign", - "target", - "id", - "summary", -}; +/* Avoid duplicate HTML id= attributes. */ +static struct ohash id_unique; -static void print_spec(struct html *, enum roffdeco, - const char *, size_t); -static void print_res(struct html *, const char *, size_t); -static void print_ctag(struct html *, enum htmltag); -static void print_doctype(struct html *); -static void print_xmltype(struct html *); -static int print_encode(struct html *, const char *, int); -static void print_metaf(struct html *, enum roffdeco); -static void print_attr(struct html *, - const char *, const char *); -static void *ml_alloc(char *, enum htmltype); +static void a2width(const char *, struct roffsu *); +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 *); +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 *, enum mandoc_esc); -static void * -ml_alloc(char *outopts, enum htmltype type) -{ - struct html *h; - const char *toks[4]; - char *v; - toks[0] = "style"; - toks[1] = "man"; - toks[2] = "includes"; - toks[3] = NULL; - - h = calloc(1, sizeof(struct html)); - if (NULL == h) { - perror(NULL); - exit((int)MANDOCLEVEL_SYSERR); - } - - h->type = type; - h->tags.head = NULL; - h->symtab = chars_init(CHARS_HTML); - - while (outopts && *outopts) - switch (getsubopt(&outopts, UNCONST(toks), &v)) { - case (0): - h->style = v; - break; - case (1): - h->base_man = v; - break; - case (2): - h->base_includes = v; - break; - default: - break; - } - - return(h); -} - void * -html_alloc(char *outopts) +html_alloc(const struct manoutput *outopts) { + struct html *h; - return(ml_alloc(outopts, HTML_HTML_4_01_STRICT)); -} + h = mandoc_calloc(1, sizeof(struct html)); + h->tag = NULL; + h->style = outopts->style; + h->base_man = outopts->man; + h->base_includes = outopts->includes; + if (outopts->fragment) + h->oflags |= HTML_FRAGMENT; -void * -xhtml_alloc(char *outopts) -{ + mandoc_ohash_init(&id_unique, 4, 0); - return(ml_alloc(outopts, HTML_XHTML_1_0_STRICT)); + return h; } - void html_free(void *p) { struct tag *tag; struct html *h; + char *cp; + unsigned int slot; h = (struct html *)p; - - while ((tag = h->tags.head) != NULL) { - h->tags.head = tag->next; + while ((tag = h->tag) != NULL) { + h->tag = tag->next; free(tag); } - - if (h->symtab) - chars_free(h->symtab); - free(h); + + cp = ohash_first(&id_unique, &slot); + while (cp != NULL) { + free(cp); + cp = ohash_next(&id_unique, &slot); + } + ohash_delete(&id_unique); } - void print_gen_head(struct html *h) { - struct htmlpair tag[4]; + struct tag *t; - tag[0].key = ATTR_HTTPEQUIV; - tag[0].val = "Content-Type"; - tag[1].key = ATTR_CONTENT; - tag[1].val = "text/html; charset=utf-8"; - print_otag(h, TAG_META, 2, tag); + 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; + } - tag[0].key = ATTR_NAME; - tag[0].val = "resource-type"; - tag[1].key = ATTR_CONTENT; - tag[1].val = "document"; - print_otag(h, TAG_META, 2, tag); + /* + * Print a minimal embedded style sheet. + */ - if (h->style) { - tag[0].key = ATTR_REL; - tag[0].val = "stylesheet"; - tag[1].key = ATTR_HREF; - tag[1].val = h->style; - tag[2].key = ATTR_TYPE; - tag[2].val = "text/css"; - tag[3].key = ATTR_MEDIA; - tag[3].val = "all"; - print_otag(h, TAG_LINK, 4, tag); - } + 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, "div.Pp { margin: 1ex 0ex; }"); + print_endline(h); + print_text(h, "div.Nd, div.Bf, div.Op { display: inline; }"); + print_endline(h); + print_text(h, "span.Pa, span.Ad { font-style: italic; }"); + print_endline(h); + print_text(h, "span.Ms { font-weight: bold; }"); + print_endline(h); + print_text(h, "dl.Bl-diag "); + print_byte(h, '>'); + print_text(h, " dt { font-weight: bold; }"); + print_endline(h); + print_text(h, "code.Nm, code.Fl, code.Cm, code.Ic, " + "code.In, code.Fd, code.Fn,"); + print_endline(h); + print_text(h, "code.Cd { font-weight: bold; " + "font-family: inherit; }"); + print_tagq(h, t); } - static void -print_spec(struct html *h, enum roffdeco d, const char *p, size_t len) +print_metaf(struct html *h, enum mandoc_esc deco) { - int cp; - const char *rhs; - size_t sz; + enum htmlfont font; - if ((cp = chars_spec2cp(h->symtab, p, len)) > 0) { - printf("&#%d;", cp); - return; - } else if (-1 == cp && DECO_SSPECIAL == d) { - fwrite(p, 1, len, stdout); - return; - } else if (-1 == cp) - return; + switch (deco) { + case ESCAPE_FONTPREV: + font = h->metal; + break; + case ESCAPE_FONTITALIC: + font = HTMLFONT_ITALIC; + break; + case ESCAPE_FONTBOLD: + font = HTMLFONT_BOLD; + break; + case ESCAPE_FONTBI: + font = HTMLFONT_BI; + break; + case ESCAPE_FONT: + case ESCAPE_FONTROMAN: + font = HTMLFONT_NONE; + break; + default: + abort(); + } - if (NULL != (rhs = chars_spec2str(h->symtab, p, len, &sz))) - fwrite(rhs, 1, sz, stdout); -} + if (h->metaf) { + print_tagq(h, h->metaf); + h->metaf = NULL; + } + h->metal = h->metac; + h->metac = font; -static void -print_res(struct html *h, const char *p, size_t len) + switch (font) { + case HTMLFONT_ITALIC: + h->metaf = print_otag(h, TAG_I, ""); + break; + case HTMLFONT_BOLD: + h->metaf = print_otag(h, TAG_B, ""); + break; + case HTMLFONT_BI: + h->metaf = print_otag(h, TAG_B, ""); + print_otag(h, TAG_I, ""); + break; + default: + break; + } +} + +char * +html_make_id(const struct roff_node *n, int unique) { - int cp; - const char *rhs; - size_t sz; + const struct roff_node *nch; + char *buf, *bufs, *cp; + unsigned int slot; + int suffix; - if ((cp = chars_res2cp(h->symtab, p, len)) > 0) { - printf("&#%d;", cp); - return; - } else if (-1 == cp) - return; + for (nch = n->child; nch != NULL; nch = nch->next) + if (nch->type != ROFFT_TEXT) + return NULL; - if (NULL != (rhs = chars_res2str(h->symtab, p, len, &sz))) - fwrite(rhs, 1, sz, stdout); -} + buf = NULL; + deroff(&buf, n); + if (buf == NULL) + return NULL; + /* + * 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 + */ -struct tag * -print_ofont(struct html *h, enum htmlfont font) -{ - struct htmlpair tag; + for (cp = buf; *cp != '\0'; cp++) + if (isalnum((unsigned char)*cp) == 0 && + strchr("!$&'()*+,-./:;=?@_~", *cp) == NULL) + *cp = '_'; - h->metal = h->metac; - h->metac = font; + if (unique == 0) + return buf; - /* FIXME: DECO_ROMAN should just close out preexisting. */ + /* Avoid duplicate HTML id= attributes. */ - if (h->metaf && h->tags.head == h->metaf) - print_tagq(h, h->metaf); - - PAIR_CLASS_INIT(&tag, htmlfonts[font]); - h->metaf = print_otag(h, TAG_SPAN, 1, &tag); - return(h->metaf); + 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; } +int +html_strlen(const char *cp) +{ + size_t rsz; + int skip, sz; -static void -print_metaf(struct html *h, enum roffdeco deco) + /* + * Account for escaped sequences within string length + * calculations. This follows the logic in term_strlen() as we + * must calculate the width of produced strings. + * Assume that characters are always width of "1". This is + * hacky, but it gets the job done for approximation of widths. + */ + + sz = 0; + skip = 0; + while (1) { + rsz = strcspn(cp, "\\"); + if (rsz) { + cp += rsz; + if (skip) { + skip = 0; + rsz--; + } + sz += rsz; + } + if ('\0' == *cp) + break; + cp++; + switch (mandoc_escape(&cp, NULL, NULL)) { + case ESCAPE_ERROR: + return sz; + case ESCAPE_UNICODE: + case ESCAPE_NUMBERED: + case ESCAPE_SPECIAL: + case ESCAPE_OVERSTRIKE: + if (skip) + skip = 0; + else + sz++; + break; + case ESCAPE_SKIPCHAR: + skip = 1; + break; + default: + break; + } + } + return sz; +} + +static int +print_escape(struct html *h, char c) { - enum htmlfont font; - switch (deco) { - case (DECO_PREVIOUS): - font = h->metal; + switch (c) { + case '<': + print_word(h, "<"); break; - case (DECO_ITALIC): - font = HTMLFONT_ITALIC; + case '>': + print_word(h, ">"); break; - case (DECO_BOLD): - font = HTMLFONT_BOLD; + case '&': + print_word(h, "&"); break; - case (DECO_ROMAN): - font = HTMLFONT_NONE; + case '"': + print_word(h, """); break; + case ASCII_NBRSP: + print_word(h, " "); + break; + case ASCII_HYPH: + print_byte(h, '-'); + break; + case ASCII_BREAK: + break; default: - abort(); - /* NOTREACHED */ + return 0; } - - (void)print_ofont(h, font); + return 1; } - static int -print_encode(struct html *h, const char *p, int norecurse) +print_encode(struct html *h, const char *p, const char *pend, int norecurse) { - size_t sz; - int len, nospace; + char numbuf[16]; + struct tag *t; const char *seq; - enum roffdeco deco; - static const char rejs[6] = { '\\', '<', '>', '&', ASCII_HYPH, '\0' }; + size_t sz; + int c, len, breakline, nospace; + enum mandoc_esc esc; + static const char rejs[10] = { ' ', '\\', '<', '>', '&', '"', + ASCII_NBRSP, ASCII_HYPH, ASCII_BREAK, '\0' }; + if (pend == NULL) + pend = strchr(p, '\0'); + + breakline = 0; nospace = 0; - for (; *p; p++) { - sz = strcspn(p, rejs); + while (p < pend) { + if (HTML_SKIPCHAR & h->flags && '\\' != *p) { + h->flags &= ~HTML_SKIPCHAR; + p++; + continue; + } - fwrite(p, 1, sz, stdout); - p += /* LINTED */ - sz; + for (sz = strcspn(p, rejs); sz-- && p < pend; p++) + print_byte(h, *p); - if ('<' == *p) { - printf("<"); + if (breakline && + (p >= pend || *p == ' ' || *p == ASCII_NBRSP)) { + t = print_otag(h, TAG_DIV, ""); + print_text(h, "\\~"); + print_tagq(h, t); + breakline = 0; + while (p < pend && (*p == ' ' || *p == ASCII_NBRSP)) + p++; continue; - } else if ('>' == *p) { - printf(">"); + } + + if (p >= pend) + break; + + if (*p == ' ') { + print_endword(h); + p++; continue; - } else if ('&' == *p) { - printf("&"); + } + + if (print_escape(h, *p++)) continue; - } else if (ASCII_HYPH == *p) { - /* - * Note: "soft hyphens" aren't graphically - * displayed when not breaking the text; we want - * them to be displayed. - */ - /*printf("­");*/ - putchar('-'); + + esc = mandoc_escape(&p, &seq, &len); + if (ESCAPE_ERROR == esc) + break; + + switch (esc) { + case ESCAPE_FONT: + case ESCAPE_FONTPREV: + case ESCAPE_FONTBOLD: + case ESCAPE_FONTITALIC: + case ESCAPE_FONTBI: + case ESCAPE_FONTROMAN: + if (0 == norecurse) + print_metaf(h, esc); continue; - } else if ('\0' == *p) + case ESCAPE_SKIPCHAR: + h->flags |= HTML_SKIPCHAR; + continue; + default: break; + } - seq = ++p; - len = a2roffdeco(&deco, &seq, &sz); + if (h->flags & HTML_SKIPCHAR) { + h->flags &= ~HTML_SKIPCHAR; + continue; + } - switch (deco) { - case (DECO_RESERVED): - print_res(h, seq, sz); + switch (esc) { + case ESCAPE_UNICODE: + /* Skip past "u" header. */ + c = mchars_num2uc(seq + 1, len - 1); break; - case (DECO_SSPECIAL): - /* FALLTHROUGH */ - case (DECO_SPECIAL): - print_spec(h, deco, seq, sz); + case ESCAPE_NUMBERED: + c = mchars_num2char(seq, len); + if (c < 0) + continue; break; - case (DECO_PREVIOUS): - /* FALLTHROUGH */ - case (DECO_BOLD): - /* FALLTHROUGH */ - case (DECO_ITALIC): - /* FALLTHROUGH */ - case (DECO_ROMAN): - if (norecurse) - break; - print_metaf(h, deco); + case ESCAPE_SPECIAL: + c = mchars_spec2cp(seq, len); + if (c <= 0) + continue; break; - default: + 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 += len - 1; - - if (DECO_NOSPACE == deco && '\0' == *(p + 1)) - nospace = 1; + 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); + return nospace; } - static void -print_attr(struct html *h, const char *key, const char *val) +print_href(struct html *h, const char *name, const char *sec, int man) { - printf(" %s=\"", key); - (void)print_encode(h, val, 1); - putchar('\"'); + const char *p, *pp; + + pp = man ? h->base_man : 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); } - struct tag * -print_otag(struct html *h, enum htmltag tag, - int sz, const struct htmlpair *p) +print_otag(struct html *h, enum htmltag tag, const char *fmt, ...) { - int i; + va_list ap; + struct roffsu mysu, *su; + char numbuf[16]; struct tag *t; + const char *attr; + char *arg1, *arg2; + double v; + int i, have_style, tflags; - /* Push this tags onto the stack of open scopes. */ + tflags = htmltags[tag].flags; - if ( ! (HTML_NOSTACK & htmltags[tag].flags)) { - t = malloc(sizeof(struct tag)); - if (NULL == t) { - perror(NULL); - exit((int)MANDOCLEVEL_SYSERR); - } + /* Push this tag onto the stack of open scopes. */ + + if ((tflags & HTML_NOSTACK) == 0) { + t = mandoc_malloc(sizeof(struct tag)); t->tag = tag; - t->next = h->tags.head; - h->tags.head = t; + t->next = h->tag; + h->tag = t; } else t = NULL; - if ( ! (HTML_NOSPACE & h->flags)) - if ( ! (HTML_CLRLINE & htmltags[tag].flags)) { - /* Manage keeps! */ - if ( ! (HTML_KEEP & h->flags)) { - if (HTML_PREKEEP & h->flags) - h->flags |= HTML_KEEP; - putchar(' '); - } else - 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); } + } if ( ! (h->flags & HTML_NONOSPACE)) h->flags &= ~HTML_NOSPACE; @@ -423,371 +597,464 @@ print_otag(struct html *h, enum htmltag tag, /* Print out the tag name and attributes. */ - printf("<%s", htmltags[tag].name); - for (i = 0; i < sz; i++) - print_attr(h, htmlattrs[p[i].key], p[i].val); + print_byte(h, '<'); + print_word(h, htmltags[tag].name); - /* Add non-overridable attributes. */ + va_start(ap, fmt); - if (TAG_HTML == tag && HTML_XHTML_1_0_STRICT == h->type) { - print_attr(h, "xmlns", "http://www.w3.org/1999/xhtml"); - print_attr(h, "xml:lang", "en"); - print_attr(h, "lang", "en"); + have_style = 0; + while (*fmt != '\0') { + if (*fmt == 's') { + have_style = 1; + fmt++; + break; + } + + /* Parse a non-style attribute and its arguments. */ + + arg1 = va_arg(ap, char *); + 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(); + } + arg2 = NULL; + if (*fmt == 'M') + arg2 = va_arg(ap, char *); + if (arg1 == NULL) + continue; + + /* Print the non-style attributes. */ + + 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; + case 'T': + print_encode(h, arg1, NULL, 1); + print_word(h, "\" title=\""); + print_encode(h, arg1, NULL, 1); + fmt++; + break; + default: + print_encode(h, arg1, NULL, 1); + break; + } + print_byte(h, '"'); } - /* Accomodate for XML "well-formed" singleton escaping. */ + /* Print out styles. */ - if (HTML_AUTOCLOSE & htmltags[tag].flags) - switch (h->type) { - case (HTML_XHTML_1_0_STRICT): - putchar('/'); + while (*fmt != '\0') { + arg1 = NULL; + su = NULL; + + /* First letter: input argument type. */ + + switch (*fmt++) { + case 'h': + i = va_arg(ap, int); + su = &mysu; + SCALE_HS_INIT(su, i); break; + case 's': + arg1 = va_arg(ap, char *); + break; + case 'u': + su = va_arg(ap, struct roffsu *); + break; + case 'w': + if ((arg2 = va_arg(ap, char *)) != NULL) { + su = &mysu; + a2width(arg2, su); + } + if (*fmt == '+') { + if (su != NULL) { + /* Make even bold text fit. */ + su->scale *= 1.2; + /* Add padding. */ + su->scale += 3.0; + } + fmt++; + } + break; default: + abort(); + } + + /* Second letter: style name. */ + + switch (*fmt++) { + case 'h': + attr = "height"; break; + case 'i': + attr = "text-indent"; + break; + case 'l': + attr = "margin-left"; + break; + case 'w': + attr = "width"; + break; + case 'W': + attr = "min-width"; + break; + case '?': + attr = arg1; + arg1 = va_arg(ap, char *); + break; + default: + abort(); } + if (su == NULL && arg1 == NULL) + continue; - putchar('>'); + if (have_style == 1) + print_word(h, " style=\""); + else + print_byte(h, ' '); + print_word(h, attr); + print_byte(h, ':'); + print_byte(h, ' '); + if (su != NULL) { + v = su->scale; + if (su->unit == SCALE_MM && (v /= 100.0) == 0.0) + v = 1.0; + else if (su->unit == SCALE_BU) + v /= 24.0; + (void)snprintf(numbuf, sizeof(numbuf), "%.2f", v); + print_word(h, numbuf); + print_word(h, roffscales[su->unit]); + } else + print_word(h, arg1); + print_byte(h, ';'); + have_style = 2; + } + if (have_style == 2) + print_byte(h, '"'); - h->flags |= HTML_NOSPACE; + va_end(ap); - if ((HTML_AUTOCLOSE | HTML_CLRLINE) & htmltags[tag].flags) - putchar('\n'); + /* Accommodate for "well-formed" singleton escaping. */ - return(t); -} + if (HTML_AUTOCLOSE & htmltags[tag].flags) + print_byte(h, '/'); + print_byte(h, '>'); -static void -print_ctag(struct html *h, enum htmltag tag) -{ - - printf("", htmltags[tag].name); - if (HTML_CLRLINE & htmltags[tag].flags) { + if (tflags & HTML_NLBEGIN) + print_endline(h); + else h->flags |= HTML_NOSPACE; - putchar('\n'); - } -} + if (tflags & HTML_INDENT) + h->indent++; + if (tflags & HTML_NOINDENT) + h->noindent++; -void -print_gen_decls(struct html *h) -{ - - print_xmltype(h); - print_doctype(h); + return t; } - static void -print_xmltype(struct html *h) +print_ctag(struct html *h, struct tag *tag) { + int tflags; - if (HTML_XHTML_1_0_STRICT == h->type) - printf(""); -} + /* + * Remember to close out and nullify the current + * meta-font and table, if applicable. + */ + if (tag == h->metaf) + h->metaf = NULL; + if (tag == h->tblt) + h->tblt = NULL; + tflags = htmltags[tag->tag].flags; -static void -print_doctype(struct html *h) -{ - const char *doctype; - const char *dtd; - const char *name; + 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); - switch (h->type) { - case (HTML_HTML_4_01_STRICT): - name = "HTML"; - doctype = "-//W3C//DTD HTML 4.01//EN"; - dtd = "http://www.w3.org/TR/html4/strict.dtd"; - break; - default: - name = "html"; - doctype = "-//W3C//DTD XHTML 1.0 Strict//EN"; - dtd = "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"; - break; - } + h->tag = tag->next; + free(tag); +} - printf("\n", - name, doctype, dtd); +void +print_gen_decls(struct html *h) +{ + print_word(h, ""); + print_endline(h); } - void -print_text(struct html *h, const char *word) +print_gen_comment(struct html *h, struct roff_node *n) { + int wantblank; - 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 ( ! (HTML_IGNDELIM & h->flags)) - h->flags |= HTML_NOSPACE; - break; - default: - 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; + } + if (wantblank) + print_endline(h); + print_word(h, " -->"); + print_endline(h); + h->indent = 0; +} - if ( ! (HTML_NOSPACE & h->flags)) { - /* Manage keeps! */ +void +print_text(struct html *h, const char *word) +{ + if (h->col && (h->flags & HTML_NOSPACE) == 0) { if ( ! (HTML_KEEP & h->flags)) { if (HTML_PREKEEP & h->flags) h->flags |= HTML_KEEP; - putchar(' '); + print_endword(h); } else - printf(" "); + print_word(h, " "); } + assert(NULL == h->metaf); + switch (h->metac) { + case HTMLFONT_ITALIC: + h->metaf = print_otag(h, TAG_I, ""); + break; + case HTMLFONT_BOLD: + h->metaf = print_otag(h, TAG_B, ""); + break; + case HTMLFONT_BI: + h->metaf = print_otag(h, TAG_B, ""); + print_otag(h, TAG_I, ""); + break; + default: + print_indent(h); + break; + } + assert(word); - if ( ! print_encode(h, word, 0)) + 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; - h->flags &= ~HTML_IGNDELIM; + if (h->metaf) { + print_tagq(h, h->metaf); + h->metaf = NULL; + } - /* - * Note that we don't process the pipe: the parser sees it as - * punctuation, but we don't in terms of typography. - */ - if (word[0] && '\0' == word[1]) - switch (word[0]) { - case('('): - /* FALLTHROUGH */ - case('['): - h->flags |= HTML_NOSPACE; - break; - default: - break; - } + h->flags &= ~HTML_IGNDELIM; } - void print_tagq(struct html *h, const struct tag *until) { struct tag *tag; - while ((tag = h->tags.head) != NULL) { - if (tag == h->metaf) - h->metaf = NULL; - print_ctag(h, tag->tag); - h->tags.head = tag->next; - free(tag); + while ((tag = h->tag) != NULL) { + print_ctag(h, tag); if (until && tag == until) return; } } - void print_stagq(struct html *h, const struct tag *suntil) { struct tag *tag; - while ((tag = h->tags.head) != NULL) { + while ((tag = h->tag) != NULL) { if (suntil && tag == suntil) return; - if (tag == h->metaf) - h->metaf = NULL; - print_ctag(h, tag->tag); - h->tags.head = tag->next; - free(tag); + print_ctag(h, tag); } } - void -bufinit(struct html *h) +print_paragraph(struct html *h) { + struct tag *t; - h->buf[0] = '\0'; - h->buflen = 0; + t = print_otag(h, TAG_DIV, "c", "Pp"); + print_tagq(h, t); } -void -bufcat_style(struct html *h, const char *key, const char *val) -{ +/*********************************************************************** + * Low level output functions. + * They implement line breaking using a short static buffer. + ***********************************************************************/ - bufcat(h, key); - bufncat(h, ":", 1); - bufcat(h, val); - bufncat(h, ";", 1); -} - - -void -bufcat(struct html *h, const char *p) +/* + * 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) { + if ((h->flags & HTML_BUFFER) == 0) { + putchar(c); + h->col++; + return; + } - bufncat(h, p, strlen(p)); -} + if (h->col + h->bufcol < sizeof(h->buf)) { + h->buf[h->bufcol++] = c; + return; + } - -void -buffmt(struct html *h, const char *fmt, ...) -{ - va_list ap; - - va_start(ap, fmt); - (void)vsnprintf(h->buf + (int)h->buflen, - BUFSIZ - h->buflen - 1, fmt, ap); - va_end(ap); - h->buflen = strlen(h->buf); + 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; } - +/* + * If something was printed on the current output line, end it. + * Not to be called right after print_indent(). + */ void -bufncat(struct html *h, const char *p, size_t sz) +print_endline(struct html *h) { + if (h->col == 0) + return; - if (h->buflen + sz > BUFSIZ - 1) - sz = BUFSIZ - 1 - h->buflen; - - (void)strncat(h->buf, p, sz); - h->buflen += sz; + 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; } - -void -buffmt_includes(struct html *h, const char *name) +/* + * Flush the HTML output buffer. + * If it is inactive, activate it. + */ +static void +print_endword(struct html *h) { - const char *p, *pp; - - pp = h->base_includes; - - while (NULL != (p = strchr(pp, '%'))) { - bufncat(h, pp, (size_t)(p - pp)); - switch (*(p + 1)) { - case('I'): - bufcat(h, name); - break; - default: - bufncat(h, p, 2); - break; - } - pp = p + 2; + if (h->noindent) { + print_byte(h, ' '); + return; } - if (pp) - bufcat(h, pp); -} - -void -buffmt_man(struct html *h, - const char *name, const char *sec) -{ - const char *p, *pp; - - pp = h->base_man; - - /* LINTED */ - while (NULL != (p = strchr(pp, '%'))) { - bufncat(h, pp, (size_t)(p - pp)); - switch (*(p + 1)) { - case('S'): - bufcat(h, sec ? sec : "1"); - break; - case('N'): - buffmt(h, name); - break; - default: - bufncat(h, p, 2); - break; - } - pp = p + 2; + 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; } - if (pp) - bufcat(h, pp); + h->bufcol = 0; } - -void -bufcat_su(struct html *h, const char *p, const struct roffsu *su) +/* + * 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) { - double v; - const char *u; + size_t i; - v = su->scale; + if (h->col) + return; - switch (su->unit) { - case (SCALE_CM): - u = "cm"; - break; - case (SCALE_IN): - u = "in"; - break; - case (SCALE_PC): - u = "pc"; - break; - case (SCALE_PT): - u = "pt"; - break; - case (SCALE_EM): - u = "em"; - break; - case (SCALE_MM): - if (0 == (v /= 100)) - v = 1; - u = "em"; - break; - case (SCALE_EN): - u = "ex"; - break; - case (SCALE_BU): - u = "ex"; - break; - case (SCALE_VS): - u = "em"; - break; - default: - u = "ex"; - break; + if (h->noindent == 0) { + h->col = h->indent * 2; + for (i = 0; i < h->col; i++) + putchar(' '); } + h->flags &= ~HTML_NOSPACE; +} - /* - * XXX: the CSS spec isn't clear as to which types accept - * integer or real numbers, so we just make them all decimals. - */ - buffmt(h, "%s: %.2f%s;", p, v, u); +/* + * Print or buffer some characters + * depending on the current HTML output buffer state. + */ +static void +print_word(struct html *h, const char *cp) +{ + while (*cp != '\0') + print_byte(h, *cp++); } - -void -html_idcat(char *dst, const char *src, int sz) +/* + * Calculate the scaling unit passed in a `-width' argument. This uses + * either a native scaling unit (e.g., 1i, 2m) or the string length of + * the value. + */ +static void +a2width(const char *p, struct roffsu *su) { - int ssz; + const char *end; - assert(sz); - - /* Cf. . */ - - for ( ; *dst != '\0' && sz; dst++, sz--) - /* Jump to end. */ ; - - assert(sz > 2); - - /* We can't start with a number (bah). */ - - *dst++ = 'x'; - *dst = '\0'; - sz--; - - for ( ; *src != '\0' && sz > 1; src++) { - ssz = snprintf(dst, (size_t)sz, "%.2x", *src); - sz -= ssz; - dst += ssz; - } + end = a2roffsu(p, su, SCALE_MAX); + if (end == NULL || *end != '\0') { + su->unit = SCALE_EN; + su->scale = html_strlen(p); + } else if (su->scale < 0.0) + su->scale = 0.0; }