=================================================================== RCS file: /cvs/mandoc/html.c,v retrieving revision 1.256 retrieving revision 1.266 diff -u -p -r1.256 -r1.266 --- mandoc/html.c 2019/08/02 17:06:04 1.256 +++ mandoc/html.c 2020/04/07 22:56:02 1.266 @@ -1,7 +1,7 @@ -/* $Id: html.c,v 1.256 2019/08/02 17:06:04 schwarze Exp $ */ +/* $Id: html.c,v 1.266 2020/04/07 22:56:02 schwarze Exp $ */ /* + * Copyright (c) 2011-2015, 2017-2020 Ingo Schwarze * Copyright (c) 2008-2011, 2014 Kristaps Dzonsons - * Copyright (c) 2011-2015, 2017-2019 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 @@ -14,6 +14,9 @@ * 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 "config.h" @@ -42,34 +45,30 @@ struct htmldata { const char *name; int flags; -#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_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 << 6) -#define HTML_NOINDENT (1 << 7) +#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_NLALL}, {"head", HTML_NLALL | HTML_INDENT}, - {"body", HTML_NLALL}, - {"meta", HTML_NOSTACK | HTML_AUTOCLOSE | HTML_NLALL}, + {"meta", HTML_NOSTACK | HTML_NLALL}, + {"link", HTML_NOSTACK | HTML_NLALL}, + {"style", HTML_NLALL | HTML_INDENT}, {"title", HTML_NLAROUND}, + {"body", HTML_NLALL}, {"div", HTML_NLAROUND}, - {"div", 0}, {"section", HTML_NLALL}, - {"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}, {"tr", HTML_NLALL | HTML_INDENT}, {"td", HTML_NLAROUND}, @@ -79,16 +78,21 @@ static const struct htmldata htmltags[TAG_MAX] = { {"dl", HTML_NLALL | HTML_INDENT}, {"dt", HTML_NLAROUND}, {"dd", HTML_NLAROUND | HTML_INDENT}, - {"p", 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}, + {"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}, @@ -271,21 +275,18 @@ print_metaf(struct html *h) void html_close_paragraph(struct html *h) { - struct tag *t; + struct tag *this, *next; + int flags; - for (t = h->tag; t != NULL && t->closed == 0; t = t->next) { - switch(t->tag) { - case TAG_P: - case TAG_PRE: - print_tagq(h, t); + 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; - case TAG_A: - print_tagq(h, t); - continue; - default: - continue; - } - break; + this = next; } } @@ -323,6 +324,18 @@ html_fillmode(struct html *h, enum roff_tok want) return had; } +/* + * 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. + * If the "unique" argument is 0, the caller is responsible for + * free(3)ing the returned string after using it. + * If the "unique" argument is non-zero, the "id_unique" ohash table + * is used for de-duplication and owns the returned string, so the + * caller must not free(3) it. In this case, it will be freed + * automatically by html_reset() or html_free(). + */ char * html_make_id(const struct roff_node *n, int unique) { @@ -331,15 +344,31 @@ html_make_id(const struct roff_node *n, int unique) unsigned int slot; int suffix; - for (nch = n->child; nch != NULL; nch = nch->next) - if (nch->type != ROFFT_TEXT) - return NULL; + if (n->string != NULL) + buf = mandoc_strdup(n->string); + 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; + } + } - 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 @@ -584,6 +613,25 @@ print_otag(struct html *h, enum htmltag tag, const cha 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) { @@ -701,7 +749,7 @@ print_otag(struct html *h, enum htmltag tag, const cha /* Accommodate for "well-formed" singleton escaping. */ - if (HTML_AUTOCLOSE & htmltags[tag].flags) + if (htmltags[tag].flags & HTML_NOSTACK) print_byte(h, '/'); print_byte(h, '>'); @@ -719,6 +767,48 @@ print_otag(struct html *h, enum htmltag tag, const cha return t; } +/* + * 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) +{ + struct roff_node *nch; + struct tag *ret, *t; + char *id, *href; + + 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, 0) : 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); + } + } + if (id == NULL) + free(href); + return ret; +} + static void print_ctag(struct html *h, struct tag *tag) { @@ -788,6 +878,16 @@ print_gen_comment(struct html *h, struct roff_node *n) void print_text(struct html *h, const char *word) { + /* + * Always wrap text in a paragraph unless already contained in + * some flow container; never put it directly into a section. + */ + + if (h->tag->tag == TAG_SECTION) + print_otag(h, TAG_P, "c", "Pp"); + + /* Output whitespace before this text? */ + if (h->col && (h->flags & HTML_NOSPACE) == 0) { if ( ! (HTML_KEEP & h->flags)) { if (HTML_PREKEEP & h->flags) @@ -797,6 +897,11 @@ print_text(struct html *h, const char *word) print_word(h, " "); } + /* + * Print the text, optionally surrounded by HTML whitespace, + * optionally manually switching fonts before and after. + */ + assert(h->metaf == NULL); print_metaf(h); print_indent(h); @@ -937,15 +1042,12 @@ print_indent(struct html *h) { size_t i; - if (h->col) + if (h->col || h->noindent) return; - if (h->noindent == 0) { - h->col = h->indent * 2; - for (i = 0; i < h->col; i++) - putchar(' '); - } - h->flags &= ~HTML_NOSPACE; + h->col = h->indent * 2; + for (i = 0; i < h->col; i++) + putchar(' '); } /*