=================================================================== RCS file: /cvs/mandoc/html.c,v retrieving revision 1.196 retrieving revision 1.229 diff -u -p -r1.196 -r1.229 --- mandoc/html.c 2017/01/18 19:22:21 1.196 +++ mandoc/html.c 2018/05/25 20:23:51 1.229 @@ -1,7 +1,7 @@ -/* $Id: html.c,v 1.196 2017/01/18 19:22:21 schwarze Exp $ */ +/* $Id: html.c,v 1.229 2018/05/25 20:23:51 schwarze Exp $ */ /* * Copyright (c) 2008-2011, 2014 Kristaps Dzonsons - * Copyright (c) 2011-2015, 2017 Ingo Schwarze + * 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 @@ -22,14 +22,17 @@ #include #include #include +#include #include #include #include #include #include -#include "mandoc.h" #include "mandoc_aux.h" +#include "mandoc_ohash.h" +#include "mandoc.h" +#include "roff.h" #include "out.h" #include "html.h" #include "manconf.h" @@ -58,6 +61,7 @@ static const struct htmldata htmltags[TAG_MAX] = { {"meta", HTML_NOSTACK | HTML_AUTOCLOSE | HTML_NLALL}, {"title", HTML_NLAROUND}, {"div", HTML_NLAROUND}, + {"div", 0}, {"h1", HTML_NLAROUND}, {"h2", HTML_NLAROUND}, {"span", 0}, @@ -65,7 +69,7 @@ static const struct htmldata htmltags[TAG_MAX] = { {"br", HTML_NOSTACK | HTML_AUTOCLOSE | HTML_NLALL}, {"a", 0}, {"table", HTML_NLALL | HTML_INDENT}, - {"tbody", HTML_NLALL | HTML_INDENT}, + {"colgroup", HTML_NLALL | HTML_INDENT}, {"col", HTML_NOSTACK | HTML_AUTOCLOSE | HTML_NLALL}, {"tr", HTML_NLALL | HTML_INDENT}, {"td", HTML_NLAROUND}, @@ -75,8 +79,9 @@ static const struct htmldata htmltags[TAG_MAX] = { {"dl", HTML_NLALL | HTML_INDENT}, {"dt", HTML_NLAROUND}, {"dd", HTML_NLAROUND | HTML_INDENT}, - {"blockquote", HTML_NLALL | HTML_INDENT}, {"pre", HTML_NLALL | HTML_NOINDENT}, + {"var", 0}, + {"cite", 0}, {"b", 0}, {"i", 0}, {"code", 0}, @@ -85,6 +90,7 @@ static const struct htmldata htmltags[TAG_MAX] = { {"math", HTML_NLALL | HTML_INDENT}, {"mrow", 0}, {"mi", 0}, + {"mn", 0}, {"mo", 0}, {"msup", 0}, {"msub", 0}, @@ -113,11 +119,17 @@ static const char *const roffscales[SCALE_MAX] = { "ex", /* SCALE_FS */ }; +/* Avoid duplicate HTML id= attributes. */ +static struct ohash id_unique; + static void a2width(const char *, struct roffsu *); -static void html_endline(struct html *); -static void html_indent(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 *); + 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); @@ -130,13 +142,15 @@ 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; if (outopts->fragment) h->oflags |= HTML_FRAGMENT; + mandoc_ohash_init(&id_unique, 4, 0); + return h; } @@ -145,15 +159,22 @@ 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); } - free(h); + + cp = ohash_first(&id_unique, &slot); + while (cp != NULL) { + free(cp); + cp = ohash_next(&id_unique, &slot); + } + ohash_delete(&id_unique); } void @@ -162,28 +183,41 @@ print_gen_head(struct html *h) struct tag *t; 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 default style-sheet. + * Print a minimal embedded style sheet. */ t = print_otag(h, TAG_STYLE, ""); print_text(h, "table.head, table.foot { width: 100%; }"); - html_endline(h); + print_endline(h); print_text(h, "td.head-rtitle, td.foot-os { text-align: right; }"); - html_endline(h); + print_endline(h); print_text(h, "td.head-vol { text-align: center; }"); - html_endline(h); - print_text(h, "table.foot td { width: 50%; }"); - html_endline(h); - print_text(h, "table.head td { width: 33%; }"); - html_endline(h); - print_text(h, "div.spacer { margin: 1em 0; }"); + 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); - - if (h->style) - print_otag(h, TAG_LINK, "?h??", "rel", "stylesheet", - h->style, "type", "text/css", "media", "all"); } static void @@ -236,6 +270,56 @@ print_metaf(struct html *h, enum mandoc_esc deco) } } +char * +html_make_id(const struct roff_node *n, int unique) +{ + const struct roff_node *nch; + char *buf, *bufs, *cp; + unsigned int slot; + int suffix; + + 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; + + /* http://www.w3.org/TR/html5/dom.html#the-id-attribute */ + + for (cp = buf; *cp != '\0'; cp++) + if (*cp == ' ') + *cp = '_'; + + if (unique == 0) + return buf; + + /* Avoid duplicate HTML id= attributes. */ + + 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) { @@ -288,27 +372,27 @@ 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; @@ -321,16 +405,19 @@ print_escape(char c) static int print_encode(struct html *h, const char *p, const char *pend, int norecurse) { - size_t sz; - int c, len, nospace; + char numbuf[16]; + struct tag *t; const char *seq; + size_t sz; + int c, len, breakline, nospace; enum mandoc_esc esc; - static const char rejs[9] = { '\\', '<', '>', '&', '"', + static const char rejs[10] = { ' ', '\\', '<', '>', '&', '"', ASCII_NBRSP, ASCII_HYPH, ASCII_BREAK, '\0' }; if (pend == NULL) pend = strchr(p, '\0'); + breakline = 0; nospace = 0; while (p < pend) { @@ -340,19 +427,32 @@ 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++) + print_byte(h, *p); - fwrite(p, 1, sz, stdout); - p += (int)sz; + 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; + } if (p >= pend) break; - if (print_escape(*p++)) + if (*p == ' ') { + print_endword(h); + p++; continue; + } + if (print_escape(h, *p++)) + continue; + esc = mandoc_escape(&p, &seq, &len); if (ESCAPE_ERROR == esc) break; @@ -394,6 +494,9 @@ print_encode(struct html *h, const char *p, const char if (c <= 0) continue; break; + case ESCAPE_BREAK: + breakline = 1; + continue; case ESCAPE_NOSPACE: if ('\0' == *p) nospace = 1; @@ -409,10 +512,11 @@ 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), "&#x%.4X;", c); + print_word(h, numbuf); + } else if (print_escape(h, c) == 0) + print_byte(h, c); } return nospace; @@ -428,7 +532,7 @@ 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') || @@ -447,35 +551,36 @@ 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, tflags; tflags = htmltags[tag].flags; - /* Push this tags onto the stack of open scopes. */ + /* 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 (tflags & HTML_NLBEFORE) - html_endline(h); - if (h->flags & HTML_NLDONE) - html_indent(h); + print_endline(h); + if (h->col == 0) + print_indent(h); else if ((h->flags & HTML_NOSPACE) == 0) { if (h->flags & HTML_KEEP) - printf(" "); + print_word(h, " "); else { if (h->flags & HTML_PREKEEP) h->flags |= HTML_KEEP; - putchar(' '); + print_endword(h); } } @@ -486,19 +591,22 @@ 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"; @@ -510,59 +618,96 @@ 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); - SCALE_VS_INIT(su, i); - break; case 'w': - s = va_arg(ap, char *); - a2width(s, su); + if ((arg2 = va_arg(ap, char *)) != NULL) { + su = &mysu; + a2width(arg2, su); + } + if (*fmt == '*') { + if (su != NULL && su->unit == SCALE_EN && + su->scale > 5.9 && su->scale < 6.1) + su = NULL; + fmt++; + } + if (*fmt == '+') { + if (su != NULL) { + /* Make even bold text fit. */ + su->scale *= 1.2; + /* Add padding. */ + su->scale += 3.0; + } + fmt++; + } + if (*fmt == '-') { + if (su != NULL) + su->scale *= -1.0; + fmt++; + } break; default: abort(); @@ -571,9 +716,6 @@ print_otag(struct html *h, enum htmltag tag, const cha /* Second letter: style name. */ switch (*fmt++) { - case 'b': - attr = "margin-bottom"; - break; case 'h': attr = "height"; break; @@ -583,9 +725,6 @@ print_otag(struct html *h, enum htmltag tag, const cha case 'l': attr = "margin-left"; break; - case 't': - attr = "margin-top"; - break; case 'w': attr = "width"; break; @@ -593,32 +732,50 @@ 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, '>'); if (tflags & HTML_NLBEGIN) - html_endline(h); + print_endline(h); else h->flags |= HTML_NOSPACE; @@ -651,34 +808,62 @@ print_ctag(struct html *h, struct tag *tag) if (tflags & HTML_NOINDENT) h->noindent--; if (tflags & HTML_NLEND) - html_endline(h); - html_indent(h); - printf("", htmltags[tag->tag].name); + 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) - html_endline(h); + print_endline(h); - h->tags.head = tag->next; + h->tag = tag->next; free(tag); } void print_gen_decls(struct html *h) { + print_word(h, ""); + print_endline(h); +} - puts(""); - h->flags |= HTML_NLDONE; +void +print_gen_comment(struct html *h, struct roff_node *n) +{ + int wantblank; + + 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; } void print_text(struct html *h, const char *word) { - if ((h->flags & (HTML_NLDONE | HTML_NOSPACE)) == 0) { + 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); @@ -694,7 +879,7 @@ print_text(struct html *h, const char *word) print_otag(h, TAG_I, ""); break; default: - html_indent(h); + print_indent(h); break; } @@ -719,7 +904,7 @@ 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; @@ -731,7 +916,7 @@ 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); @@ -743,45 +928,125 @@ 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. + ***********************************************************************/ + /* - * If something was printed on the current output line, end it. - * Not to be called right after html_indent(). + * 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 -html_endline(struct html *h) +print_byte(struct html *h, char c) { - if (h->flags & HTML_NLDONE) + 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->flags |= HTML_NLDONE | HTML_NOSPACE; + 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 -html_indent(struct html *h) +print_indent(struct html *h) { - int i; + size_t i; - if ((h->flags & HTML_NLDONE) == 0) + if (h->col) return; - if (h->noindent == 0) - for (i = 0; i < h->indent * 2; i++) + if (h->noindent == 0) { + h->col = h->indent * 2; + for (i = 0; i < h->col; i++) putchar(' '); - h->flags &= ~(HTML_NLDONE | HTML_NOSPACE); + } + 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 * either a native scaling unit (e.g., 1i, 2m) or the string length of * the value. @@ -789,7 +1054,10 @@ html_indent(struct html *h) static void a2width(const char *p, struct roffsu *su) { - if (a2roffsu(p, su, SCALE_MAX) < 2) { + const char *end; + + 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)