=================================================================== RCS file: /cvs/mandoc/html.c,v retrieving revision 1.195 retrieving revision 1.212 diff -u -p -r1.195 -r1.212 --- mandoc/html.c 2017/01/17 15:32:43 1.195 +++ mandoc/html.c 2017/05/14 12:27:28 1.212 @@ -1,4 +1,4 @@ -/* $Id: html.c,v 1.195 2017/01/17 15:32:43 schwarze Exp $ */ +/* $Id: html.c,v 1.212 2017/05/14 12:27:28 schwarze Exp $ */ /* * Copyright (c) 2008-2011, 2014 Kristaps Dzonsons * Copyright (c) 2011-2015, 2017 Ingo Schwarze @@ -28,8 +28,9 @@ #include #include -#include "mandoc.h" #include "mandoc_aux.h" +#include "mandoc.h" +#include "roff.h" #include "out.h" #include "html.h" #include "manconf.h" @@ -38,58 +39,67 @@ 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 */ - {"pre", HTML_CLRLINE }, /* TAG_PRE */ - {"b", 0 }, /* TAG_B */ - {"i", 0 }, /* TAG_I */ - {"code", 0 }, /* TAG_CODE */ - {"small", 0 }, /* TAG_SMALL */ - {"style", HTML_CLRLINE}, /* TAG_STYLE */ - {"math", HTML_CLRLINE}, /* TAG_MATH */ - {"mrow", 0}, /* TAG_MROW */ - {"mi", 0}, /* TAG_MI */ - {"mo", 0}, /* TAG_MO */ - {"msup", 0}, /* TAG_MSUP */ - {"msub", 0}, /* TAG_MSUB */ - {"msubsup", 0}, /* TAG_MSUBSUP */ - {"mfrac", 0}, /* TAG_MFRAC */ - {"msqrt", 0}, /* TAG_MSQRT */ - {"mfenced", 0}, /* TAG_MFENCED */ - {"mtable", 0}, /* TAG_MTABLE */ - {"mtr", 0}, /* TAG_MTR */ - {"mtd", 0}, /* TAG_MTD */ - {"munderover", 0}, /* TAG_MUNDEROVER */ - {"munder", 0}, /* TAG_MUNDER*/ - {"mover", 0}, /* TAG_MOVER*/ + {"html", HTML_NLALL}, + {"head", HTML_NLALL | HTML_INDENT}, + {"body", HTML_NLALL}, + {"meta", HTML_NOSTACK | HTML_AUTOCLOSE | HTML_NLALL}, + {"title", HTML_NLAROUND}, + {"div", HTML_NLAROUND}, + {"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}, + {"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 roffscales[SCALE_MAX] = { @@ -106,8 +116,13 @@ static const char *const roffscales[SCALE_MAX] = { }; 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(char); +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); @@ -120,7 +135,7 @@ html_alloc(const struct manoutput *outopts) h = mandoc_calloc(1, sizeof(struct html)); - h->tags.head = NULL; + h->tag = NULL; h->style = outopts->style; h->base_man = outopts->man; h->base_includes = outopts->includes; @@ -138,8 +153,8 @@ html_free(void *p) 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); } @@ -156,13 +171,15 @@ print_gen_head(struct html *h) /* * Print a default style-sheet. */ + t = print_otag(h, TAG_STYLE, ""); - print_text(h, "table.head, table.foot { width: 100%; }\n" - "td.head-rtitle, td.foot-os { text-align: right; }\n" - "td.head-vol { text-align: center; }\n" - "table.foot td { width: 50%; }\n" - "table.head td { width: 33%; }\n" - "div.spacer { margin: 1em 0; }\n"); + 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_tagq(h, t); if (h->style) @@ -220,6 +237,28 @@ print_metaf(struct html *h, enum mandoc_esc deco) } } +char * +html_make_id(const struct roff_node *n) +{ + const struct roff_node *nch; + char *buf, *cp; + + for (nch = n->child; nch != NULL; nch = nch->next) + if (nch->type != ROFFT_TEXT) + return NULL; + + buf = NULL; + deroff(&buf, n); + + /* http://www.w3.org/TR/html5/dom.html#the-id-attribute */ + + for (cp = buf; *cp != '\0'; cp++) + if (*cp == ' ') + *cp = '_'; + + return buf; +} + int html_strlen(const char *cp) { @@ -272,27 +311,27 @@ html_strlen(const char *cp) } static int -print_escape(char c) +print_escape(struct html *h, char c) { switch (c) { case '<': - printf("<"); + print_word(h, "<"); break; case '>': - printf(">"); + print_word(h, ">"); break; case '&': - printf("&"); + print_word(h, "&"); break; case '"': - printf("""); + print_word(h, """); break; case ASCII_NBRSP: - printf(" "); + print_word(h, " "); break; case ASCII_HYPH: - putchar('-'); + print_byte(h, '-'); break; case ASCII_BREAK: break; @@ -305,6 +344,7 @@ print_escape(char c) static int print_encode(struct html *h, const char *p, const char *pend, int norecurse) { + char numbuf[16]; size_t sz; int c, len, nospace; const char *seq; @@ -324,17 +364,16 @@ print_encode(struct html *h, const char *p, const char continue; } - sz = strcspn(p, rejs); - if (p + sz > pend) - sz = pend - p; + for (sz = strcspn(p, rejs); sz-- && p < pend; p++) + if (*p == ' ') + print_endword(h); + else + print_byte(h, *p); - fwrite(p, 1, sz, stdout); - p += (int)sz; - if (p >= pend) break; - if (print_escape(*p++)) + if (print_escape(h, *p++)) continue; esc = mandoc_escape(&p, &seq, &len); @@ -393,10 +432,11 @@ print_encode(struct html *h, const char *p, const char if ((c < 0x20 && c != 0x09) || (c > 0x7E && c < 0xA0)) c = 0xFFFD; - if (c > 0x7E) - printf("&#%d;", c); - else if ( ! print_escape(c)) - putchar(c); + if (c > 0x7E) { + (void)snprintf(numbuf, sizeof(numbuf), "&#%d;", c); + print_word(h, numbuf); + } else if (print_escape(h, c) == 0) + print_byte(h, c); } return nospace; @@ -412,7 +452,7 @@ print_href(struct html *h, const char *name, const cha print_encode(h, pp, p, 1); if (man && p[1] == 'S') { if (sec == NULL) - putchar('1'); + print_byte(h, '1'); else print_encode(h, sec, NULL, 1); } else if ((man && p[1] == 'N') || @@ -431,32 +471,38 @@ print_otag(struct html *h, enum htmltag tag, const cha { va_list ap; struct roffsu mysu, *su; + char numbuf[16]; struct tag *t; const char *attr; - char *s; + char *arg1, *arg2; double v; - int i, have_style; + int i, have_style, tflags; - /* Push this tags onto the stack of open scopes. */ + tflags = htmltags[tag].flags; - if ( ! (HTML_NOSTACK & htmltags[tag].flags)) { + /* 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; @@ -465,19 +511,22 @@ print_otag(struct html *h, enum htmltag tag, const cha /* Print out the tag name and attributes. */ - printf("<%s", htmltags[tag].name); + print_byte(h, '<'); + print_word(h, htmltags[tag].name); va_start(ap, fmt); have_style = 0; while (*fmt != '\0') { if (*fmt == 's') { - printf(" style=\""); have_style = 1; fmt++; break; } - s = va_arg(ap, char *); + + /* Parse a non-style attribute and its arguments. */ + + arg1 = va_arg(ap, char *); switch (*fmt++) { case 'c': attr = "class"; @@ -489,59 +538,92 @@ print_otag(struct html *h, enum htmltag tag, const cha attr = "id"; break; case '?': - attr = s; - s = va_arg(ap, char *); + attr = arg1; + arg1 = va_arg(ap, char *); break; default: abort(); } - printf(" %s=\"", attr); + 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 'M': - print_href(h, s, va_arg(ap, char *), 1); + case 'I': + print_href(h, arg1, NULL, 0); fmt++; break; - case 'I': - print_href(h, s, NULL, 0); + case 'M': + print_href(h, arg1, arg2, 1); fmt++; break; case 'R': - putchar('#'); + print_byte(h, '#'); + print_encode(h, arg1, NULL, 1); fmt++; - /* FALLTHROUGH */ + 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, s, NULL, 1); + print_encode(h, arg1, NULL, 1); break; } - putchar('"'); + print_byte(h, '"'); } /* Print out styles. */ - s = NULL; - su = &mysu; 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': - s = va_arg(ap, char *); + arg1 = va_arg(ap, char *); break; case 'u': su = va_arg(ap, struct roffsu *); break; case 'v': i = va_arg(ap, int); + su = &mysu; SCALE_VS_INIT(su, i); break; case 'w': - s = va_arg(ap, char *); - a2width(s, su); + if ((arg2 = va_arg(ap, char *)) == NULL) + break; + su = &mysu; + a2width(arg2, su); + if (*fmt == '+') { + /* Increase to make even bold text fit. */ + su->scale *= 1.2; + /* Add padding. */ + su->scale += 3.0; + fmt++; + } + if (*fmt == '-') { + su->scale *= -1.0; + fmt++; + } break; default: abort(); @@ -572,34 +654,57 @@ print_otag(struct html *h, enum htmltag tag, const cha attr = "min-width"; break; case '?': - printf("%s: %s;", s, va_arg(ap, char *)); - continue; + attr = arg1; + arg1 = va_arg(ap, char *); + break; default: abort(); } - 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; - printf("%s: %.2f%s;", attr, v, roffscales[su->unit]); + if (su == NULL && arg1 == NULL) + continue; + + 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) - putchar('"'); + if (have_style == 2) + print_byte(h, '"'); va_end(ap); /* Accommodate for "well-formed" singleton escaping. */ if (HTML_AUTOCLOSE & htmltags[tag].flags) - putchar('/'); + print_byte(h, '/'); - putchar('>'); + print_byte(h, '>'); - h->flags |= HTML_NOSPACE; + if (tflags & HTML_NLBEGIN) + print_endline(h); + else + h->flags |= HTML_NOSPACE; - if ((HTML_AUTOCLOSE | HTML_CLRLINE) & htmltags[tag].flags) - putchar('\n'); + if (tflags & HTML_INDENT) + h->indent++; + if (tflags & HTML_NOINDENT) + h->noindent++; return t; } @@ -607,6 +712,7 @@ print_otag(struct html *h, enum htmltag tag, const cha static void print_ctag(struct html *h, struct tag *tag) { + int tflags; /* * Remember to close out and nullify the current @@ -617,35 +723,43 @@ print_ctag(struct html *h, struct tag *tag) if (tag == h->tblt) h->tblt = NULL; - printf("", htmltags[tag->tag].name); - if (HTML_CLRLINE & htmltags[tag->tag].flags) { - h->flags |= HTML_NOSPACE; - putchar('\n'); - } + tflags = htmltags[tag->tag].flags; - h->tags.head = tag->next; + 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); + + h->tag = tag->next; free(tag); } void print_gen_decls(struct html *h) { - - puts(""); + print_word(h, ""); + print_endline(h); } void print_text(struct html *h, const char *word) { - - if ( ! (HTML_NOSPACE & h->flags)) { - /* Manage keeps! */ + 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); @@ -661,6 +775,7 @@ print_text(struct html *h, const char *word) print_otag(h, TAG_I, ""); break; default: + print_indent(h); break; } @@ -685,7 +800,7 @@ print_tagq(struct html *h, const struct tag *until) { struct tag *tag; - while ((tag = h->tags.head) != NULL) { + while ((tag = h->tag) != NULL) { print_ctag(h, tag); if (until && tag == until) return; @@ -697,7 +812,7 @@ 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; print_ctag(h, tag); @@ -709,10 +824,123 @@ print_paragraph(struct html *h) { struct tag *t; - t = print_otag(h, TAG_DIV, "c", "spacer"); + t = print_otag(h, TAG_DIV, "c", "Pp"); print_tagq(h, t); } + +/*********************************************************************** + * Low level output functions. + * They implement line breaking using a short static buffer. + ***********************************************************************/ + +/* + * Buffer one HTML output byte. + * If the buffer is full, flush and deactivate it and start a new line. + * If the buffer is inactive, print directly. + */ +static void +print_byte(struct html *h, char c) +{ + if ((h->flags & HTML_BUFFER) == 0) { + putchar(c); + h->col++; + return; + } + + if (h->col + h->bufcol < sizeof(h->buf)) { + h->buf[h->bufcol++] = c; + return; + } + + 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 +print_endline(struct html *h) +{ + if (h->col == 0) + return; + + 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; +} + +/* + * Flush the HTML output buffer. + * If it is inactive, activate it. + */ +static void +print_endword(struct html *h) +{ + if (h->noindent) { + print_byte(h, ' '); + return; + } + + if ((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; +} + +/* + * 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) +{ + size_t i; + + if (h->col) + return; + + if (h->noindent == 0) { + h->col = h->indent * 2; + for (i = 0; i < h->col; i++) + putchar(' '); + } + h->flags &= ~HTML_NOSPACE; +} + +/* + * 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++); +} /* * Calculate the scaling unit passed in a `-width' argument. This uses