=================================================================== RCS file: /cvs/docbook2mdoc/parse.c,v retrieving revision 1.6 retrieving revision 1.17 diff -u -p -r1.6 -r1.17 --- docbook2mdoc/parse.c 2019/03/28 15:05:40 1.6 +++ docbook2mdoc/parse.c 2019/04/07 14:49:26 1.17 @@ -1,4 +1,4 @@ -/* $Id: parse.c,v 1.6 2019/03/28 15:05:40 schwarze Exp $ */ +/* $Id: parse.c,v 1.17 2019/04/07 14:49:26 schwarze Exp $ */ /* * Copyright (c) 2014 Kristaps Dzonsons * Copyright (c) 2019 Ingo Schwarze @@ -30,6 +30,14 @@ * The implementation of the DocBook parser. */ +enum pstate { + PARSE_ELEM, + PARSE_TAG, + PARSE_ARG, + PARSE_SQ, + PARSE_DQ +}; + /* * Global parse state. * Keep this as simple and small as possible. @@ -44,6 +52,7 @@ struct parse { int nline; /* Line number of next token. */ int ncol; /* Column number of next token. */ int del; /* Levels of nested nodes being deleted. */ + int spc; /* Whitespace before the next element. */ int attr; /* The most recent attribute is valid. */ int warn; }; @@ -69,10 +78,11 @@ static const struct element elements[] = { { "citerefentry", NODE_CITEREFENTRY }, { "citetitle", NODE_CITETITLE }, { "cmdsynopsis", NODE_CMDSYNOPSIS }, - { "code", NODE_CODE }, + { "code", NODE_LITERAL }, { "colspec", NODE_COLSPEC }, { "command", NODE_COMMAND }, { "constant", NODE_CONSTANT }, + { "contrib", NODE_CONTRIB }, { "copyright", NODE_COPYRIGHT }, { "date", NODE_DATE }, { "editor", NODE_EDITOR }, @@ -80,9 +90,10 @@ static const struct element elements[] = { { "emphasis", NODE_EMPHASIS }, { "entry", NODE_ENTRY }, { "envar", NODE_ENVAR }, + { "errorname", NODE_ERRORNAME }, { "fieldsynopsis", NODE_FIELDSYNOPSIS }, { "filename", NODE_FILENAME }, - { "firstname", NODE_IGNORE }, + { "firstname", NODE_PERSONNAME }, { "firstterm", NODE_FIRSTTERM }, { "footnote", NODE_FOOTNOTE }, { "funcdef", NODE_FUNCDEF }, @@ -97,7 +108,7 @@ static const struct element elements[] = { { "indexterm", NODE_DELETE }, { "info", NODE_INFO }, { "informalequation", NODE_INFORMALEQUATION }, - { "informaltable", NODE_INFORMALTABLE }, + { "informaltable", NODE_TABLE }, { "inlineequation", NODE_INLINEEQUATION }, { "itemizedlist", NODE_ITEMIZEDLIST }, { "keysym", NODE_KEYSYM }, @@ -122,7 +133,7 @@ static const struct element elements[] = { { "option", NODE_OPTION }, { "orderedlist", NODE_ORDEREDLIST }, { "orgname", NODE_ORGNAME }, - { "othername", NODE_IGNORE }, + { "othername", NODE_PERSONNAME }, { "para", NODE_PARA }, { "paramdef", NODE_PARAMDEF }, { "parameter", NODE_PARAMETER }, @@ -160,11 +171,14 @@ static const struct element elements[] = { { "sect2", NODE_SECTION }, { "section", NODE_SECTION }, { "sgmltag", NODE_SGMLTAG }, + { "simpara", NODE_PARA }, { "simplelist", NODE_SIMPLELIST }, { "spanspec", NODE_SPANSPEC }, - { "structname", NODE_STRUCTNAME }, + { "structfield", NODE_PARAMETER }, + { "structname", NODE_TYPE }, { "subtitle", NODE_SUBTITLE }, - { "surname", NODE_IGNORE }, + { "surname", NODE_PERSONNAME }, + { "symbol", NODE_CONSTANT }, { "synopsis", NODE_SYNOPSIS }, { "table", NODE_TABLE }, { "tbody", NODE_TBODY }, @@ -177,7 +191,7 @@ static const struct element elements[] = { { "trademark", NODE_IGNORE }, { "type", NODE_TYPE }, { "ulink", NODE_ULINK }, - { "userinput", NODE_USERINPUT }, + { "userinput", NODE_LITERAL }, { "variablelist", NODE_VARIABLELIST }, { "varlistentry", NODE_VARLISTENTRY }, { "varname", NODE_VARNAME }, @@ -188,6 +202,62 @@ static const struct element elements[] = { { NULL, NODE_IGNORE } }; +struct entity { + const char *name; + const char *roff; +}; + +/* + * XML character entity references found in the wild. + * Those that don't have an exact mandoc_char(7) representation + * are approximated, and the desired codepoint is given as a comment. + * Encoding them as \\[u...] would leave -Tascii out in the cold. + */ +static const struct entity entities[] = { + { "alpha", "\\(*a" }, + { "amp", "&" }, + { "apos", "'" }, + { "auml", "\\(:a" }, + { "beta", "\\(*b" }, + { "circ", "^" }, /* U+02C6 */ + { "copy", "\\(co" }, + { "dagger", "\\(dg" }, + { "Delta", "\\(*D" }, + { "eacute", "\\('e" }, + { "emsp", "\\ " }, /* U+2003 */ + { "gt", ">" }, + { "hairsp", "\\^" }, + { "kappa", "\\(*k" }, + { "larr", "\\(<-" }, + { "ldquo", "\\(lq" }, + { "le", "\\(<=" }, + { "lowbar", "_" }, + { "lsqb", "[" }, + { "lt", "<" }, + { "mdash", "\\(em" }, + { "minus", "\\-" }, + { "ndash", "\\(en" }, + { "nbsp", "\\ " }, + { "num", "#" }, + { "oslash", "\\(/o" }, + { "ouml", "\\(:o" }, + { "percnt", "%" }, + { "quot", "\\(dq" }, + { "rarr", "\\(->" }, + { "rArr", "\\(rA" }, + { "rdquo", "\\(rq" }, + { "reg", "\\(rg" }, + { "rho", "\\(*r" }, + { "rsqb", "]" }, + { "sigma", "\\(*s" }, + { "shy", "\\&" }, /* U+00AD */ + { "tau", "\\(*t" }, + { "tilde", "\\[u02DC]" }, + { "times", "\\[tmu]" }, + { "uuml", "\\(:u" }, + { NULL, NULL } +}; + static void error_msg(struct parse *p, const char *fmt, ...) { @@ -225,6 +295,7 @@ static void xml_char(struct parse *ps, const char *p, int sz) { struct pnode *dat; + size_t newsz; if (ps->del > 0) return; @@ -240,6 +311,7 @@ xml_char(struct parse *ps, const char *p, int sz) exit(1); } dat->node = NODE_TEXT; + dat->spc = ps->spc; dat->parent = ps->cur; TAILQ_INIT(&dat->childq); TAILQ_INIT(&dat->attrq); @@ -254,24 +326,79 @@ xml_char(struct parse *ps, const char *p, int sz) /* Append to the current text node. */ assert(sz >= 0); - ps->cur->b = realloc(ps->cur->b, ps->cur->bsz + sz + 1); + newsz = ps->cur->bsz + (ps->cur->bsz && ps->spc) + sz; + ps->cur->b = realloc(ps->cur->b, newsz + 1); if (ps->cur->b == NULL) { perror(NULL); exit(1); } + if (ps->cur->bsz && ps->spc) + ps->cur->b[ps->cur->bsz++] = ' '; memcpy(ps->cur->b + ps->cur->bsz, p, sz); - ps->cur->bsz += sz; - ps->cur->b[ps->cur->bsz] = '\0'; + ps->cur->b[ps->cur->bsz = newsz] = '\0'; ps->cur->real = ps->cur->b; + ps->spc = 0; } +/* + * Close out the text node and strip trailing whitespace, if one is open. + */ static void -pnode_trim(struct pnode *pn) +pnode_closetext(struct parse *p) { - assert(pn->node == NODE_TEXT); - for (; pn->bsz > 0; pn->b[--pn->bsz] = '\0') - if (isspace((unsigned char)pn->b[pn->bsz - 1]) == 0) + struct pnode *n; + + if ((n = p->cur) == NULL || n->node != NODE_TEXT) + return; + p->cur = n->parent; + while (n->bsz > 0 && isspace((unsigned char)n->b[n->bsz - 1])) { + n->b[--n->bsz] = '\0'; + p->spc = 1; + } +} + +static void +xml_entity(struct parse *p, const char *name) +{ + const struct entity *entity; + struct pnode *dat; + + if (p->del > 0) + return; + + if (p->cur == NULL) { + error_msg(p, "discarding entity before document: &%s;", name); + return; + } + + pnode_closetext(p); + + if (p->tree->flags & TREE_CLOSED && p->cur == p->tree->root) + warn_msg(p, "entity after end of document: &%s;", name); + + for (entity = entities; entity->name != NULL; entity++) + if (strcmp(name, entity->name) == 0) break; + + if (entity->roff == NULL) { + error_msg(p, "unknown entity &%s;", name); + return; + } + + /* Create, append, and close out an entity node. */ + if ((dat = calloc(1, sizeof(*dat))) == NULL || + (dat->b = dat->real = strdup(entity->roff)) == NULL) { + perror(NULL); + exit(1); + } + dat->node = NODE_ESCAPE; + dat->bsz = strlen(dat->b); + dat->spc = p->spc; + dat->parent = p->cur; + TAILQ_INIT(&dat->childq); + TAILQ_INIT(&dat->attrq); + TAILQ_INSERT_TAIL(&p->cur->childq, dat, child); + p->spc = 0; } /* @@ -295,11 +422,7 @@ xml_elem_start(struct parse *ps, const char *name) return; } - /* Close out the text node, if there is one. */ - if (ps->cur != NULL && ps->cur->node == NODE_TEXT) { - pnode_trim(ps->cur); - ps->cur = ps->cur->parent; - } + pnode_closetext(ps); for (elem = elements; elem->name != NULL; elem++) if (strcmp(elem->name, name) == 0) @@ -333,7 +456,50 @@ xml_elem_start(struct parse *ps, const char *name) perror(NULL); exit(1); } - dat->node = elem->node; + + /* + * Nodes that begin a new macro or request line or start by + * printing text always want whitespace before themselves. + */ + + switch (dat->node = elem->node) { + case NODE_AUTHORGROUP: + case NODE_BOOKINFO: + case NODE_CAUTION: + case NODE_EDITOR: + case NODE_ENTRY: + case NODE_FUNCDEF: + case NODE_FUNCPROTOTYPE: + case NODE_INFORMALEQUATION: + case NODE_INLINEEQUATION: + case NODE_ITEMIZEDLIST: + case NODE_LEGALNOTICE: + case NODE_LITERALLAYOUT: + case NODE_NOTE: + case NODE_ORDEREDLIST: + case NODE_PARA: + case NODE_PREFACE: + case NODE_PROGRAMLISTING: + case NODE_REFMETA: + case NODE_REFNAMEDIV: + case NODE_REFSYNOPSISDIV: + case NODE_ROW: + case NODE_SBR: + case NODE_SCREEN: + case NODE_SECTION: + case NODE_SYNOPSIS: + case NODE_TGROUP: + case NODE_TIP: + case NODE_TITLE: + case NODE_VARIABLELIST: + case NODE_VARLISTENTRY: + case NODE_WARNING: + dat->spc = 1; + break; + default: + dat->spc = ps->spc; + break; + } dat->parent = ps->cur; TAILQ_INIT(&dat->childq); TAILQ_INIT(&dat->attrq); @@ -404,11 +570,8 @@ xml_elem_end(struct parse *ps, const char *name) return; } - /* Close out the text node, if there is one. */ - if (ps->del == 0 && ps->cur != NULL && ps->cur->node == NODE_TEXT) { - pnode_trim(ps->cur); - ps->cur = ps->cur->parent; - } + if (ps->del == 0) + pnode_closetext(ps); if (name != NULL) { for (elem = elements; elem->name != NULL; elem++) @@ -443,6 +606,7 @@ xml_elem_end(struct parse *ps, const char *name) ps->tree->flags |= TREE_CLOSED; else ps->cur = ps->cur->parent; + ps->spc = 0; break; } assert(ps->del == 0); @@ -476,6 +640,19 @@ parse_free(struct parse *p) free(p); } +static void +increment(struct parse *p, char *b, size_t *pend, int refill) +{ + if (refill) { + if (b[*pend] == '\n') { + p->nline++; + p->ncol = 1; + } else + p->ncol++; + } + ++*pend; +} + /* * Advance the pend pointer to the next character in the charset. * If the charset starts with a space, it stands for any whitespace. @@ -486,7 +663,7 @@ parse_free(struct parse *p) */ static int advance(struct parse *p, char *b, size_t rlen, size_t *pend, - const char *charset) + const char *charset, int refill) { int space; @@ -496,197 +673,221 @@ advance(struct parse *p, char *b, size_t rlen, size_t } else space = 0; - p->nline = p->line; - p->ncol = p->col; + if (refill) { + p->nline = p->line; + p->ncol = p->col; + } while (*pend < rlen) { - if (b[*pend] == '\n') { - p->nline++; - p->ncol = 1; - } else - p->ncol++; if (space && isspace((unsigned char)b[*pend])) break; if (strchr(charset, b[*pend]) != NULL) break; - ++*pend; + increment(p, b, pend, refill); } if (*pend == rlen) { b[rlen] = '\0'; - return 1; + return refill; } else return 0; } -struct ptree * -parse_file(struct parse *p, int fd, const char *fname) +size_t +parse_string(struct parse *p, char *b, size_t rlen, + enum pstate *pstate, int refill) { - char b[4096]; - ssize_t rsz; /* Return value from read(2). */ - size_t rlen; /* Number of bytes in b[]. */ + char *cp; size_t poff; /* Parse offset in b[]. */ size_t pend; /* Offset of the end of the current word. */ - int in_tag, in_arg, in_quotes, elem_end; + int elem_end; - p->fname = fname; - p->nline = 1; - p->ncol = 1; - rlen = 0; - in_tag = in_arg = in_quotes = 0; + p->spc = 0; + pend = 0; + for (;;) { - /* - * Read loop. - * - * We have to enter the read loop once more even on EOF - * because the previous token may have been incomplete, - * such that it asked for more input. - * Once rsz is 0, incomplete tokens will no longer ask - * for more input but instead use whatever there is, - * and then exit the read loop. - * The minus one on the size limit for read(2) is needed - * such that advance() can set b[rlen] to NUL when needed. - */ + /* Proceed to the next token, skipping whitespace. */ - while ((rsz = read(fd, b + rlen, sizeof(b) - rlen - 1)) >= 0) { - if ((rlen += rsz) == 0) + if (refill) { + p->line = p->nline; + p->col = p->ncol; + } + if ((poff = pend) == rlen) break; + if (isspace((unsigned char)b[pend])) { + p->spc = 1; + increment(p, b, &pend, refill); + continue; + } - /* Token loop. */ + /* + * The following four cases (ARG, TAG, and starting an + * entity or a tag) all parse a word or quoted string. + * If that extends beyond the read buffer and the last + * read(2) still got data, they all break out of the + * token loop to request more data from the read loop. + * + * Also, three of them detect self-closing tags, those + * ending with "/>", setting the flag elem_end and + * calling xml_elem_end() at the very end, after + * handling the attribute value, attribute name, or + * tag name, respectively. + */ - pend = 0; - for (;;) { + /* Parse an attribute value. */ - /* Proceed to the next token, skipping whitespace. */ - - p->line = p->nline; - p->col = p->ncol; - if ((poff = pend) == rlen) - break; - if (isspace((unsigned char)b[pend])) { - if (b[pend++] == '\n') { - p->nline++; - p->ncol = 1; - } else - p->ncol++; + if (*pstate >= PARSE_ARG) { + if (*pstate == PARSE_ARG && + (b[pend] == '\'' || b[pend] == '"')) { + *pstate = b[pend] == '"' ? + PARSE_DQ : PARSE_SQ; + increment(p, b, &pend, refill); continue; } + if (advance(p, b, rlen, &pend, + *pstate == PARSE_DQ ? "\"" : + *pstate == PARSE_SQ ? "'" : " >", refill)) + break; + *pstate = PARSE_TAG; + elem_end = 0; + if (b[pend] == '>') { + *pstate = PARSE_ELEM; + if (pend > 0 && b[pend - 1] == '/') { + b[pend - 1] = '\0'; + elem_end = 1; + } + } + b[pend] = '\0'; + if (pend < rlen) + increment(p, b, &pend, refill); + xml_attrval(p, b + poff); + if (elem_end) + xml_elem_end(p, NULL); - /* - * The following three cases (in_arg, in_tag, - * and starting a tag) all parse a word or - * quoted string. If that extends beyond the - * read buffer and the last read(2) still got - * data, they all break out of the token loop - * to request more data from the read loop. - * - * Also, they all detect self-closing tags, - * those ending with "/>", setting the flag - * elem_end and calling xml_elem_end() at the - * very end, after handling the attribute value, - * attribute name, or tag name, respectively. - */ + /* Look for an attribute name. */ - /* Parse an attribute value. */ - - if (in_arg) { - if (in_quotes == 0 && b[pend] == '"') { - in_quotes = 1; - p->ncol++; - pend++; - continue; + } else if (*pstate == PARSE_TAG) { + if (advance(p, b, rlen, &pend, " =>", refill)) + break; + elem_end = 0; + switch (b[pend]) { + case '>': + *pstate = PARSE_ELEM; + if (pend > 0 && b[pend - 1] == '/') { + b[pend - 1] = '\0'; + elem_end = 1; } - if (advance(p, b, rlen, &pend, - in_quotes ? "\"" : " >") && rsz > 0) - break; - in_arg = in_quotes = elem_end = 0; - if (b[pend] == '>') { - in_tag = 0; - if (pend > 0 && b[pend - 1] == '/') { - b[pend - 1] = '\0'; - elem_end = 1; - } - } - b[pend] = '\0'; - if (pend < rlen) - pend++; - xml_attrval(p, b + poff); - if (elem_end) - xml_elem_end(p, NULL); + break; + case '=': + *pstate = PARSE_ARG; + break; + default: + break; + } + b[pend] = '\0'; + if (pend < rlen) + increment(p, b, &pend, refill); + xml_attrkey(p, b + poff); + if (elem_end) + xml_elem_end(p, NULL); - /* Look for an attribute name. */ + /* Begin an opening or closing tag. */ - } else if (in_tag) { - if (advance(p, b, rlen, &pend, " =>") && - rsz > 0) - break; - elem_end = 0; - switch (b[pend]) { - case '>': - in_tag = 0; - if (pend > 0 && b[pend - 1] == '/') { - b[pend - 1] = '\0'; - elem_end = 1; - } - break; - case '=': - in_arg = 1; - break; - default: - break; - } - b[pend] = '\0'; - if (pend < rlen) - pend++; - xml_attrkey(p, b + poff); - if (elem_end) - xml_elem_end(p, NULL); + } else if (b[poff] == '<') { + if (advance(p, b, rlen, &pend, " >", refill)) + break; + if (pend > poff + 3 && + strncmp(b + poff, ""); + if (cp == NULL) { + if (refill) + break; + cp = b + rlen; } else - xml_elem_start(p, b + poff); - if (elem_end) - xml_elem_end(p, b + poff); + cp += 3; + while (b + pend < cp) + increment(p, b, &pend, refill); + continue; + } + elem_end = 0; + if (b[pend] != '>') + *pstate = PARSE_TAG; + else if (pend > 0 && b[pend - 1] == '/') { + b[pend - 1] = '\0'; + elem_end = 1; + } + b[pend] = '\0'; + if (pend < rlen) + increment(p, b, &pend, refill); + if (b[++poff] == '/') { + elem_end = 1; + poff++; + } else + xml_elem_start(p, b + poff); + if (elem_end) + xml_elem_end(p, b + poff); - /* Process text up to the next tag. */ + /* Process an entity. */ - } else { - if (advance(p, b, rlen, &pend, "<") == 0) - p->ncol--; - xml_char(p, b + poff, pend - poff); - } + } else if (b[poff] == '&') { + if (advance(p, b, rlen, &pend, ";", refill)) + break; + b[pend] = '\0'; + if (pend < rlen) + increment(p, b, &pend, refill); + xml_entity(p, b + poff + 1); + + /* Process text up to the next tag, entity, or EOL. */ + + } else { + advance(p, b, rlen, &pend, "<&", refill); + xml_char(p, b + poff, pend - poff); } + } + return poff; +} - /* Buffer exhausted; shift left and re-fill. */ +struct ptree * +parse_file(struct parse *p, int fd, const char *fname) +{ + char b[4096]; + ssize_t rsz; /* Return value from read(2). */ + size_t rlen; /* Number of bytes in b[]. */ + size_t poff; /* Parse offset in b[]. */ + enum pstate pstate; + p->fname = fname; + p->nline = 1; + p->ncol = 1; + pstate = PARSE_ELEM; + rlen = 0; + + /* + * Read loop. + * + * If the previous token was incomplete and asked for more + * input, we have to enter the read loop once more even on EOF. + * Once rsz is 0, incomplete tokens will no longer ask + * for more input but instead use whatever there is, + * and then exit the read loop. + * The minus one on the size limit for read(2) is needed + * such that advance() can set b[rlen] to NUL when needed. + */ + + while ((rsz = read(fd, b + rlen, sizeof(b) - rlen - 1)) >= 0 && + (rlen += rsz) > 0) { + poff = parse_string(p, b, rlen, &pstate, rsz > 0); + /* Buffer exhausted; shift left and re-fill. */ assert(poff > 0); - memmove(b, b + poff, rlen - poff); rlen -= poff; + memmove(b, b + poff, rlen); } if (rsz < 0) { perror(fname); p->tree->flags |= TREE_FAIL; } - if (p->cur != NULL && p->cur->node == NODE_TEXT) { - pnode_trim(p->cur); - p->cur = p->cur->parent; - } + pnode_closetext(p); if ((p->tree->flags & TREE_CLOSED) == 0) warn_msg(p, "document not closed"); return p->tree;