version 1.18, 2019/04/07 17:00:56 |
version 1.34, 2019/04/12 04:39:24 |
|
|
*/ |
*/ |
#include <assert.h> |
#include <assert.h> |
#include <ctype.h> |
#include <ctype.h> |
|
#include <errno.h> |
|
#include <fcntl.h> |
|
#include <libgen.h> |
#include <stdarg.h> |
#include <stdarg.h> |
#include <stdio.h> |
#include <stdio.h> |
#include <stdlib.h> |
#include <stdlib.h> |
|
|
struct parse { |
struct parse { |
const char *fname; /* Name of the input file. */ |
const char *fname; /* Name of the input file. */ |
struct ptree *tree; /* Complete parse result. */ |
struct ptree *tree; /* Complete parse result. */ |
|
struct pnode *doctype; |
struct pnode *cur; /* Current node in the tree. */ |
struct pnode *cur; /* Current node in the tree. */ |
enum nodeid ncur; /* Type of the current node. */ |
enum nodeid ncur; /* Type of the current node. */ |
int line; /* Line number in the input file. */ |
int line; /* Line number in the input file. */ |
|
|
int nline; /* Line number of next token. */ |
int nline; /* Line number of next token. */ |
int ncol; /* Column number of next token. */ |
int ncol; /* Column number of next token. */ |
int del; /* Levels of nested nodes being deleted. */ |
int del; /* Levels of nested nodes being deleted. */ |
int spc; /* Whitespace before the next element. */ |
int flags; |
int attr; /* The most recent attribute is valid. */ |
#define PFLAG_WARN (1 << 0) /* Print warning messages. */ |
int warn; |
#define PFLAG_SPC (1 << 1) /* Whitespace before the next element. */ |
|
#define PFLAG_ATTR (1 << 2) /* The most recent attribute is valid. */ |
|
#define PFLAG_EEND (1 << 3) /* This element is self-closing. */ |
}; |
}; |
|
|
struct element { |
struct element { |
Line 66 static const struct element elements[] = { |
|
Line 72 static const struct element elements[] = { |
|
{ "acronym", NODE_IGNORE }, |
{ "acronym", NODE_IGNORE }, |
{ "affiliation", NODE_AFFILIATION }, |
{ "affiliation", NODE_AFFILIATION }, |
{ "anchor", NODE_DELETE }, |
{ "anchor", NODE_DELETE }, |
|
{ "appendix", NODE_APPENDIX }, |
{ "application", NODE_APPLICATION }, |
{ "application", NODE_APPLICATION }, |
{ "arg", NODE_ARG }, |
{ "arg", NODE_ARG }, |
|
{ "article", NODE_SECTION }, |
{ "author", NODE_AUTHOR }, |
{ "author", NODE_AUTHOR }, |
{ "authorgroup", NODE_AUTHORGROUP }, |
{ "authorgroup", NODE_AUTHORGROUP }, |
{ "blockquote", NODE_BLOCKQUOTE }, |
{ "blockquote", NODE_BLOCKQUOTE }, |
{ "book", NODE_BOOK }, |
{ "book", NODE_SECTION }, |
{ "bookinfo", NODE_BOOKINFO }, |
{ "bookinfo", NODE_BOOKINFO }, |
{ "caution", NODE_CAUTION }, |
{ "caution", NODE_CAUTION }, |
{ "chapter", NODE_SECTION }, |
{ "chapter", NODE_SECTION }, |
Line 85 static const struct element elements[] = { |
|
Line 93 static const struct element elements[] = { |
|
{ "contrib", NODE_CONTRIB }, |
{ "contrib", NODE_CONTRIB }, |
{ "copyright", NODE_COPYRIGHT }, |
{ "copyright", NODE_COPYRIGHT }, |
{ "date", NODE_DATE }, |
{ "date", NODE_DATE }, |
|
{ "!doctype", NODE_DOCTYPE }, |
|
{ "!DOCTYPE", NODE_DOCTYPE }, |
{ "editor", NODE_EDITOR }, |
{ "editor", NODE_EDITOR }, |
{ "email", NODE_EMAIL }, |
{ "email", NODE_EMAIL }, |
{ "emphasis", NODE_EMPHASIS }, |
{ "emphasis", NODE_EMPHASIS }, |
|
{ "!ENTITY", NODE_ENTITY }, |
{ "entry", NODE_ENTRY }, |
{ "entry", NODE_ENTRY }, |
{ "envar", NODE_ENVAR }, |
{ "envar", NODE_ENVAR }, |
{ "errorname", NODE_ERRORNAME }, |
{ "errorname", NODE_ERRORNAME }, |
Line 101 static const struct element elements[] = { |
|
Line 112 static const struct element elements[] = { |
|
{ "funcsynopsis", NODE_FUNCSYNOPSIS }, |
{ "funcsynopsis", NODE_FUNCSYNOPSIS }, |
{ "funcsynopsisinfo", NODE_FUNCSYNOPSISINFO }, |
{ "funcsynopsisinfo", NODE_FUNCSYNOPSISINFO }, |
{ "function", NODE_FUNCTION }, |
{ "function", NODE_FUNCTION }, |
|
{ "glossary", NODE_VARIABLELIST }, |
|
{ "glossdef", NODE_IGNORE }, |
|
{ "glossdiv", NODE_IGNORE }, |
|
{ "glossentry", NODE_VARLISTENTRY }, |
|
{ "glosslist", NODE_VARIABLELIST }, |
{ "glossterm", NODE_GLOSSTERM }, |
{ "glossterm", NODE_GLOSSTERM }, |
{ "group", NODE_GROUP }, |
{ "group", NODE_GROUP }, |
{ "holder", NODE_HOLDER }, |
{ "holder", NODE_HOLDER }, |
Line 197 static const struct element elements[] = { |
|
Line 213 static const struct element elements[] = { |
|
{ "varname", NODE_VARNAME }, |
{ "varname", NODE_VARNAME }, |
{ "warning", NODE_WARNING }, |
{ "warning", NODE_WARNING }, |
{ "wordasword", NODE_WORDASWORD }, |
{ "wordasword", NODE_WORDASWORD }, |
{ "xi:include", NODE_DELETE_WARN }, |
{ "xi:include", NODE_INCLUDE }, |
{ "year", NODE_YEAR }, |
{ "year", NODE_YEAR }, |
{ NULL, NODE_IGNORE } |
{ NULL, NODE_IGNORE } |
}; |
}; |
Line 258 static const struct entity entities[] = { |
|
Line 274 static const struct entity entities[] = { |
|
{ NULL, NULL } |
{ NULL, NULL } |
}; |
}; |
|
|
|
static size_t parse_string(struct parse *, char *, size_t, |
|
enum pstate *, int); |
|
static void parse_fd(struct parse *, int); |
|
|
|
|
static void |
static void |
|
fatal(struct parse *p) |
|
{ |
|
fprintf(stderr, "%s:%d:%d: FATAL: ", p->fname, p->line, p->col); |
|
perror(NULL); |
|
exit(6); |
|
} |
|
|
|
static void |
error_msg(struct parse *p, const char *fmt, ...) |
error_msg(struct parse *p, const char *fmt, ...) |
{ |
{ |
va_list ap; |
va_list ap; |
|
|
fprintf(stderr, "%s:%d:%d: ", p->fname, p->line, p->col); |
fprintf(stderr, "%s:%d:%d: ERROR: ", p->fname, p->line, p->col); |
va_start(ap, fmt); |
va_start(ap, fmt); |
vfprintf(stderr, fmt, ap); |
vfprintf(stderr, fmt, ap); |
va_end(ap); |
va_end(ap); |
fputc('\n', stderr); |
fputc('\n', stderr); |
p->tree->flags |= TREE_FAIL; |
p->tree->flags |= TREE_ERROR; |
} |
} |
|
|
static void |
static void |
Line 276 warn_msg(struct parse *p, const char *fmt, ...) |
|
Line 305 warn_msg(struct parse *p, const char *fmt, ...) |
|
{ |
{ |
va_list ap; |
va_list ap; |
|
|
if (p->warn == 0) |
if ((p->flags & PFLAG_WARN) == 0) |
return; |
return; |
|
|
fprintf(stderr, "%s:%d:%d: warning: ", p->fname, p->line, p->col); |
fprintf(stderr, "%s:%d:%d: WARNING: ", p->fname, p->line, p->col); |
va_start(ap, fmt); |
va_start(ap, fmt); |
vfprintf(stderr, fmt, ap); |
vfprintf(stderr, fmt, ap); |
va_end(ap); |
va_end(ap); |
fputc('\n', stderr); |
fputc('\n', stderr); |
|
p->tree->flags |= TREE_WARN; |
} |
} |
|
|
/* |
/* |
Line 292 warn_msg(struct parse *p, const char *fmt, ...) |
|
Line 322 warn_msg(struct parse *p, const char *fmt, ...) |
|
* Otherwise, create a new one as a child of the current node. |
* Otherwise, create a new one as a child of the current node. |
*/ |
*/ |
static void |
static void |
xml_char(struct parse *ps, const char *p, int sz) |
xml_char(struct parse *p, const char *word, int sz) |
{ |
{ |
struct pnode *dat; |
struct pnode *n; |
size_t newsz; |
size_t oldsz, newsz; |
|
|
if (ps->del > 0) |
assert(sz > 0); |
|
if (p->del > 0) |
return; |
return; |
|
|
if (ps->cur == NULL) { |
if ((n = p->cur) == NULL) { |
error_msg(ps, "discarding text before document: %.*s", sz, p); |
error_msg(p, "discarding text before document: %.*s", sz, word); |
return; |
return; |
} |
} |
|
|
if (ps->cur->node != NODE_TEXT) { |
if (n->node != NODE_TEXT) { |
if ((dat = calloc(1, sizeof(*dat))) == NULL) { |
if ((n = pnode_alloc(p->cur)) == NULL) |
perror(NULL); |
fatal(p); |
exit(1); |
n->node = NODE_TEXT; |
} |
n->spc = (p->flags & PFLAG_SPC) != 0; |
dat->node = NODE_TEXT; |
p->cur = n; |
dat->spc = ps->spc; |
|
dat->parent = ps->cur; |
|
TAILQ_INIT(&dat->childq); |
|
TAILQ_INIT(&dat->attrq); |
|
TAILQ_INSERT_TAIL(&ps->cur->childq, dat, child); |
|
ps->cur = dat; |
|
} |
} |
|
|
if (ps->tree->flags & TREE_CLOSED && |
if (p->tree->flags & TREE_CLOSED && n->parent == p->tree->root) |
ps->cur->parent == ps->tree->root) |
warn_msg(p, "text after end of document: %.*s", sz, word); |
warn_msg(ps, "text after end of document: %.*s", sz, p); |
|
|
|
/* Append to the current text node. */ |
/* Append to the current text node. */ |
|
|
assert(sz >= 0); |
oldsz = n->b == NULL ? 0 : strlen(n->b); |
newsz = ps->cur->bsz + (ps->cur->bsz && ps->spc) + sz; |
newsz = oldsz + sz; |
ps->cur->b = realloc(ps->cur->b, newsz + 1); |
if (oldsz && (p->flags & PFLAG_SPC)) |
if (ps->cur->b == NULL) { |
newsz++; |
perror(NULL); |
if ((n->b = realloc(n->b, newsz + 1)) == NULL) |
exit(1); |
fatal(p); |
} |
if (oldsz && (p->flags & PFLAG_SPC)) |
if (ps->cur->bsz && ps->spc) |
n->b[oldsz++] = ' '; |
ps->cur->b[ps->cur->bsz++] = ' '; |
memcpy(n->b + oldsz, word, sz); |
memcpy(ps->cur->b + ps->cur->bsz, p, sz); |
n->b[newsz] = '\0'; |
ps->cur->b[ps->cur->bsz = newsz] = '\0'; |
p->flags &= ~PFLAG_SPC; |
ps->cur->real = ps->cur->b; |
|
ps->spc = 0; |
|
} |
} |
|
|
/* |
/* |
|
|
pnode_closetext(struct parse *p) |
pnode_closetext(struct parse *p) |
{ |
{ |
struct pnode *n; |
struct pnode *n; |
|
char *cp; |
|
|
if ((n = p->cur) == NULL || n->node != NODE_TEXT) |
if ((n = p->cur) == NULL || n->node != NODE_TEXT) |
return; |
return; |
p->cur = n->parent; |
p->cur = n->parent; |
while (n->bsz > 0 && isspace((unsigned char)n->b[n->bsz - 1])) { |
for (cp = strchr(n->b, '\0'); |
n->b[--n->bsz] = '\0'; |
cp > n->b && isspace((unsigned char)cp[-1]); |
p->spc = 1; |
*--cp = '\0') |
} |
p->flags |= PFLAG_SPC; |
} |
} |
|
|
static void |
static void |
xml_entity(struct parse *p, const char *name) |
xml_entity(struct parse *p, const char *name) |
{ |
{ |
const struct entity *entity; |
const struct entity *entity; |
struct pnode *dat; |
struct pnode *n; |
|
const char *ccp; |
|
char *cp; |
|
enum pstate pstate; |
|
|
if (p->del > 0) |
if (p->del > 0) |
return; |
return; |
Line 381 xml_entity(struct parse *p, const char *name) |
|
Line 407 xml_entity(struct parse *p, const char *name) |
|
break; |
break; |
|
|
if (entity->roff == NULL) { |
if (entity->roff == NULL) { |
|
if (p->doctype != NULL) { |
|
TAILQ_FOREACH(n, &p->doctype->childq, child) { |
|
if ((ccp = pnode_getattr_raw(n, |
|
ATTRKEY_NAME, NULL)) == NULL || |
|
strcmp(ccp, name) != 0) |
|
continue; |
|
if ((ccp = pnode_getattr_raw(n, |
|
ATTRKEY_SYSTEM, NULL)) != NULL) { |
|
parse_file(p, -1, ccp); |
|
p->flags &= ~PFLAG_SPC; |
|
return; |
|
} |
|
if ((ccp = pnode_getattr_raw(n, |
|
ATTRKEY_DEFINITION, NULL)) == NULL) |
|
continue; |
|
if ((cp = strdup(ccp)) == NULL) |
|
fatal(p); |
|
pstate = PARSE_ELEM; |
|
parse_string(p, cp, strlen(cp), &pstate, 0); |
|
p->flags &= ~PFLAG_SPC; |
|
free(cp); |
|
return; |
|
} |
|
} |
error_msg(p, "unknown entity &%s;", name); |
error_msg(p, "unknown entity &%s;", name); |
return; |
return; |
} |
} |
|
|
/* Create, append, and close out an entity node. */ |
/* Create, append, and close out an entity node. */ |
if ((dat = calloc(1, sizeof(*dat))) == NULL || |
if ((n = pnode_alloc(p->cur)) == NULL || |
(dat->b = dat->real = strdup(entity->roff)) == NULL) { |
(n->b = strdup(entity->roff)) == NULL) |
perror(NULL); |
fatal(p); |
exit(1); |
n->node = NODE_ESCAPE; |
} |
n->spc = (p->flags & PFLAG_SPC) != 0; |
dat->node = NODE_ESCAPE; |
p->flags &= ~PFLAG_SPC; |
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; |
|
} |
} |
|
|
/* |
/* |
* Begin an element. |
* Begin an element. |
*/ |
*/ |
static void |
static void |
xml_elem_start(struct parse *ps, const char *name) |
xml_elem_start(struct parse *p, const char *name) |
{ |
{ |
const struct element *elem; |
const struct element *elem; |
struct pnode *dat; |
struct pnode *n; |
|
|
if (*name == '!' || *name == '?') |
|
return; |
|
|
|
/* |
/* |
* An ancestor is excluded from the tree; |
* An ancestor is excluded from the tree; |
* keep track of the number of levels excluded. |
* keep track of the number of levels excluded. |
*/ |
*/ |
if (ps->del > 0) { |
if (p->del > 0) { |
ps->del++; |
if (*name != '!' && *name != '?') |
|
p->del++; |
return; |
return; |
} |
} |
|
|
pnode_closetext(ps); |
pnode_closetext(p); |
|
|
for (elem = elements; elem->name != NULL; elem++) |
for (elem = elements; elem->name != NULL; elem++) |
if (strcmp(elem->name, name) == 0) |
if (strcmp(elem->name, name) == 0) |
break; |
break; |
|
|
if (elem->name == NULL) |
if (elem->name == NULL) { |
error_msg(ps, "unknown element <%s>", name); |
if (*name == '!' || *name == '?') |
|
return; |
|
error_msg(p, "unknown element <%s>", name); |
|
} |
|
|
ps->ncur = elem->node; |
p->ncur = elem->node; |
|
|
switch (ps->ncur) { |
switch (p->ncur) { |
case NODE_DELETE_WARN: |
case NODE_DELETE_WARN: |
warn_msg(ps, "skipping element <%s>", name); |
warn_msg(p, "skipping element <%s>", name); |
/* FALLTHROUGH */ |
/* FALLTHROUGH */ |
case NODE_DELETE: |
case NODE_DELETE: |
ps->del = 1; |
p->del = 1; |
/* FALLTHROUGH */ |
/* FALLTHROUGH */ |
case NODE_IGNORE: |
case NODE_IGNORE: |
return; |
return; |
case NODE_INLINEEQUATION: |
case NODE_INLINEEQUATION: |
ps->tree->flags |= TREE_EQN; |
p->tree->flags |= TREE_EQN; |
break; |
break; |
default: |
default: |
break; |
break; |
} |
} |
|
|
if (ps->tree->flags & TREE_CLOSED && ps->cur->parent == NULL) |
if (p->tree->flags & TREE_CLOSED && p->cur->parent == NULL) |
warn_msg(ps, "element after end of document: <%s>", name); |
warn_msg(p, "element after end of document: <%s>", name); |
|
|
if ((dat = calloc(1, sizeof(*dat))) == NULL) { |
if ((n = pnode_alloc(p->cur)) == NULL) |
perror(NULL); |
fatal(p); |
exit(1); |
|
} |
|
|
|
/* |
/* |
* Nodes that begin a new macro or request line or start by |
* Nodes that begin a new macro or request line or start by |
* printing text always want whitespace before themselves. |
* printing text always want whitespace before themselves. |
*/ |
*/ |
|
|
switch (dat->node = elem->node) { |
switch (n->node = elem->node) { |
|
case NODE_DOCTYPE: |
|
case NODE_ENTITY: |
|
case NODE_SBR: |
|
p->flags |= PFLAG_EEND; |
|
/* FALLTHROUGH */ |
|
case NODE_APPENDIX: |
case NODE_AUTHORGROUP: |
case NODE_AUTHORGROUP: |
|
case NODE_BLOCKQUOTE: |
case NODE_BOOKINFO: |
case NODE_BOOKINFO: |
case NODE_CAUTION: |
case NODE_CAUTION: |
case NODE_EDITOR: |
case NODE_EDITOR: |
Line 484 xml_elem_start(struct parse *ps, const char *name) |
|
Line 533 xml_elem_start(struct parse *ps, const char *name) |
|
case NODE_REFNAMEDIV: |
case NODE_REFNAMEDIV: |
case NODE_REFSYNOPSISDIV: |
case NODE_REFSYNOPSISDIV: |
case NODE_ROW: |
case NODE_ROW: |
case NODE_SBR: |
|
case NODE_SCREEN: |
case NODE_SCREEN: |
case NODE_SECTION: |
case NODE_SECTION: |
case NODE_SYNOPSIS: |
case NODE_SYNOPSIS: |
Line 494 xml_elem_start(struct parse *ps, const char *name) |
|
Line 542 xml_elem_start(struct parse *ps, const char *name) |
|
case NODE_VARIABLELIST: |
case NODE_VARIABLELIST: |
case NODE_VARLISTENTRY: |
case NODE_VARLISTENTRY: |
case NODE_WARNING: |
case NODE_WARNING: |
dat->spc = 1; |
n->spc = 1; |
break; |
break; |
default: |
default: |
dat->spc = ps->spc; |
n->spc = (p->flags & PFLAG_SPC) != 0; |
break; |
break; |
} |
} |
dat->parent = ps->cur; |
p->cur = n; |
TAILQ_INIT(&dat->childq); |
if (n->node == NODE_DOCTYPE) { |
TAILQ_INIT(&dat->attrq); |
if (p->doctype == NULL) |
|
p->doctype = n; |
if (ps->cur != NULL) |
else |
TAILQ_INSERT_TAIL(&ps->cur->childq, dat, child); |
error_msg(p, "duplicate doctype"); |
|
} else if (n->parent == NULL && p->tree->root == NULL) |
ps->cur = dat; |
p->tree->root = n; |
if (ps->tree->root == NULL) |
|
ps->tree->root = dat; |
|
} |
} |
|
|
static void |
static void |
xml_attrkey(struct parse *ps, const char *name) |
xml_attrkey(struct parse *p, const char *name) |
{ |
{ |
struct pattr *attr; |
struct pattr *a; |
|
const char *value; |
enum attrkey key; |
enum attrkey key; |
|
|
if (ps->del > 0 || *name == '\0') |
if (p->del > 0 || p->ncur == NODE_IGNORE || *name == '\0') |
return; |
return; |
|
|
|
if ((p->ncur == NODE_DOCTYPE || p->ncur == NODE_ENTITY) && |
|
TAILQ_FIRST(&p->cur->attrq) == NULL) { |
|
value = name; |
|
name = "NAME"; |
|
} else |
|
value = NULL; |
|
|
if ((key = attrkey_parse(name)) == ATTRKEY__MAX) { |
if ((key = attrkey_parse(name)) == ATTRKEY__MAX) { |
ps->attr = 0; |
p->flags &= ~PFLAG_ATTR; |
return; |
return; |
} |
} |
if ((attr = calloc(1, sizeof(*attr))) == NULL) { |
if ((a = calloc(1, sizeof(*a))) == NULL) |
perror(NULL); |
fatal(p); |
exit(1); |
|
|
a->key = key; |
|
a->val = ATTRVAL__MAX; |
|
if (value == NULL) { |
|
a->rawval = NULL; |
|
p->flags |= PFLAG_ATTR; |
|
} else { |
|
if ((a->rawval = strdup(value)) == NULL) |
|
fatal(p); |
|
p->flags &= ~PFLAG_ATTR; |
} |
} |
attr->key = key; |
TAILQ_INSERT_TAIL(&p->cur->attrq, a, child); |
attr->val = ATTRVAL__MAX; |
if (p->ncur == NODE_ENTITY && key == ATTRKEY_NAME) |
attr->rawval = NULL; |
xml_attrkey(p, "DEFINITION"); |
TAILQ_INSERT_TAIL(&ps->cur->attrq, attr, child); |
|
ps->attr = 1; |
|
} |
} |
|
|
static void |
static void |
xml_attrval(struct parse *ps, const char *name) |
xml_attrval(struct parse *p, const char *name) |
{ |
{ |
struct pattr *attr; |
struct pattr *a; |
|
|
if (ps->del > 0 || ps->attr == 0) |
if (p->del > 0 || p->ncur == NODE_IGNORE || |
|
(p->flags & PFLAG_ATTR) == 0) |
return; |
return; |
if ((attr = TAILQ_LAST(&ps->cur->attrq, pattrq)) == NULL) |
if ((a = TAILQ_LAST(&p->cur->attrq, pattrq)) == NULL) |
return; |
return; |
if ((attr->val = attrval_parse(name)) == ATTRVAL__MAX && |
if ((a->val = attrval_parse(name)) == ATTRVAL__MAX && |
(attr->rawval = strdup(name)) == NULL) { |
(a->rawval = strdup(name)) == NULL) |
perror(NULL); |
fatal(p); |
exit(1); |
p->flags &= ~PFLAG_ATTR; |
} |
|
} |
} |
|
|
/* |
/* |
Line 556 xml_attrval(struct parse *ps, const char *name) |
|
Line 618 xml_attrval(struct parse *ps, const char *name) |
|
* If we're at a text node, roll that one up first. |
* If we're at a text node, roll that one up first. |
*/ |
*/ |
static void |
static void |
xml_elem_end(struct parse *ps, const char *name) |
xml_elem_end(struct parse *p, const char *name) |
{ |
{ |
const struct element *elem; |
const struct element *elem; |
|
struct pnode *n; |
|
const char *cp; |
enum nodeid node; |
enum nodeid node; |
|
|
/* |
/* |
* An ancestor is excluded from the tree; |
* An ancestor is excluded from the tree; |
* keep track of the number of levels excluded. |
* keep track of the number of levels excluded. |
*/ |
*/ |
if (ps->del > 1) { |
if (p->del > 1) { |
ps->del--; |
p->del--; |
return; |
return; |
} |
} |
|
|
if (ps->del == 0) |
if (p->del == 0) |
pnode_closetext(ps); |
pnode_closetext(p); |
|
|
if (name != NULL) { |
if (name != NULL) { |
for (elem = elements; elem->name != NULL; elem++) |
for (elem = elements; elem->name != NULL; elem++) |
Line 579 xml_elem_end(struct parse *ps, const char *name) |
|
Line 643 xml_elem_end(struct parse *ps, const char *name) |
|
break; |
break; |
node = elem->node; |
node = elem->node; |
} else |
} else |
node = ps->ncur; |
node = p->ncur; |
|
|
switch (node) { |
switch (node) { |
case NODE_DELETE_WARN: |
case NODE_DELETE_WARN: |
case NODE_DELETE: |
case NODE_DELETE: |
if (ps->del > 0) |
if (p->del > 0) |
ps->del--; |
p->del--; |
break; |
break; |
case NODE_IGNORE: |
case NODE_IGNORE: |
break; |
break; |
|
case NODE_INCLUDE: |
|
n = p->cur; |
|
p->cur = p->cur->parent; |
|
cp = pnode_getattr_raw(n, ATTRKEY_HREF, NULL); |
|
if (cp == NULL) |
|
error_msg(p, "<xi:include> element " |
|
"without href attribute"); |
|
else |
|
parse_file(p, -1, cp); |
|
pnode_unlink(n); |
|
p->flags &= ~PFLAG_SPC; |
|
break; |
|
case NODE_DOCTYPE: |
|
case NODE_SBR: |
|
p->flags &= ~PFLAG_EEND; |
|
/* FALLTHROUGH */ |
default: |
default: |
if (ps->cur == NULL || node != ps->cur->node) { |
if (p->cur == NULL || node != p->cur->node) { |
warn_msg(ps, "element not open: </%s>", name); |
warn_msg(p, "element not open: </%s>", name); |
break; |
break; |
} |
} |
|
|
Line 602 xml_elem_end(struct parse *ps, const char *name) |
|
Line 682 xml_elem_end(struct parse *ps, const char *name) |
|
* obviously better than discarding it or crashing. |
* obviously better than discarding it or crashing. |
*/ |
*/ |
|
|
if (ps->cur->parent == NULL) |
if (p->cur->parent != NULL || node == NODE_DOCTYPE) { |
ps->tree->flags |= TREE_CLOSED; |
p->cur = p->cur->parent; |
else |
if (p->cur != NULL) |
ps->cur = ps->cur->parent; |
p->ncur = p->cur->node; |
ps->spc = 0; |
} else |
|
p->tree->flags |= TREE_CLOSED; |
|
p->flags &= ~PFLAG_SPC; |
break; |
break; |
} |
} |
assert(ps->del == 0); |
assert(p->del == 0); |
} |
} |
|
|
struct parse * |
struct parse * |
Line 624 parse_alloc(int warn) |
|
Line 706 parse_alloc(int warn) |
|
free(p); |
free(p); |
return NULL; |
return NULL; |
} |
} |
p->warn = warn; |
if (warn) |
|
p->flags |= PFLAG_WARN; |
|
else |
|
p->flags &= ~PFLAG_WARN; |
return p; |
return p; |
} |
} |
|
|
Line 700 parse_string(struct parse *p, char *b, size_t rlen, |
|
Line 785 parse_string(struct parse *p, char *b, size_t rlen, |
|
size_t pend; /* Offset of the end of the current word. */ |
size_t pend; /* Offset of the end of the current word. */ |
int elem_end; |
int elem_end; |
|
|
p->spc = 0; |
|
pend = 0; |
pend = 0; |
for (;;) { |
for (;;) { |
|
|
Line 713 parse_string(struct parse *p, char *b, size_t rlen, |
|
Line 797 parse_string(struct parse *p, char *b, size_t rlen, |
|
if ((poff = pend) == rlen) |
if ((poff = pend) == rlen) |
break; |
break; |
if (isspace((unsigned char)b[pend])) { |
if (isspace((unsigned char)b[pend])) { |
p->spc = 1; |
p->flags |= PFLAG_SPC; |
increment(p, b, &pend, refill); |
increment(p, b, &pend, refill); |
continue; |
continue; |
} |
} |
Line 754 parse_string(struct parse *p, char *b, size_t rlen, |
|
Line 838 parse_string(struct parse *p, char *b, size_t rlen, |
|
b[pend - 1] = '\0'; |
b[pend - 1] = '\0'; |
elem_end = 1; |
elem_end = 1; |
} |
} |
|
if (p->flags & PFLAG_EEND) |
|
elem_end = 1; |
} |
} |
b[pend] = '\0'; |
b[pend] = '\0'; |
if (pend < rlen) |
if (pend < rlen) |
Line 765 parse_string(struct parse *p, char *b, size_t rlen, |
|
Line 851 parse_string(struct parse *p, char *b, size_t rlen, |
|
/* Look for an attribute name. */ |
/* Look for an attribute name. */ |
|
|
} else if (*pstate == PARSE_TAG) { |
} else if (*pstate == PARSE_TAG) { |
|
switch (p->ncur) { |
|
case NODE_DOCTYPE: |
|
if (b[pend] == '[') { |
|
*pstate = PARSE_ELEM; |
|
increment(p, b, &pend, refill); |
|
continue; |
|
} |
|
/* FALLTHROUGH */ |
|
case NODE_ENTITY: |
|
if (b[pend] == '"' || b[pend] == '\'') { |
|
*pstate = PARSE_ARG; |
|
continue; |
|
} |
|
break; |
|
default: |
|
break; |
|
} |
if (advance(p, b, rlen, &pend, " =>", refill)) |
if (advance(p, b, rlen, &pend, " =>", refill)) |
break; |
break; |
elem_end = 0; |
elem_end = 0; |
Line 775 parse_string(struct parse *p, char *b, size_t rlen, |
|
Line 878 parse_string(struct parse *p, char *b, size_t rlen, |
|
b[pend - 1] = '\0'; |
b[pend - 1] = '\0'; |
elem_end = 1; |
elem_end = 1; |
} |
} |
|
if (p->flags & PFLAG_EEND) |
|
elem_end = 1; |
break; |
break; |
case '=': |
case '=': |
*pstate = PARSE_ARG; |
*pstate = PARSE_ARG; |
Line 823 parse_string(struct parse *p, char *b, size_t rlen, |
|
Line 928 parse_string(struct parse *p, char *b, size_t rlen, |
|
if (b[++poff] == '/') { |
if (b[++poff] == '/') { |
elem_end = 1; |
elem_end = 1; |
poff++; |
poff++; |
} else |
} else { |
xml_elem_start(p, b + poff); |
xml_elem_start(p, b + poff); |
|
if (*pstate == PARSE_ELEM && |
|
p->flags & PFLAG_EEND) |
|
elem_end = 1; |
|
} |
if (elem_end) |
if (elem_end) |
xml_elem_end(p, b + poff); |
xml_elem_end(p, b + poff); |
|
|
|
/* Close a doctype. */ |
|
|
|
} else if (p->ncur == NODE_DOCTYPE && b[poff] == ']') { |
|
*pstate = PARSE_TAG; |
|
increment(p, b, &pend, refill); |
|
|
/* Process an entity. */ |
/* Process an entity. */ |
|
|
} else if (b[poff] == '&') { |
} else if (b[poff] == '&') { |
Line 841 parse_string(struct parse *p, char *b, size_t rlen, |
|
Line 956 parse_string(struct parse *p, char *b, size_t rlen, |
|
/* Process text up to the next tag, entity, or EOL. */ |
/* Process text up to the next tag, entity, or EOL. */ |
|
|
} else { |
} else { |
advance(p, b, rlen, &pend, "<&", refill); |
advance(p, b, rlen, &pend, |
|
p->ncur == NODE_DOCTYPE ? "<&]\n" : "<&\n", |
|
refill); |
xml_char(p, b + poff, pend - poff); |
xml_char(p, b + poff, pend - poff); |
|
if (b[pend] == '\n') |
|
pnode_closetext(p); |
} |
} |
} |
} |
return poff; |
return poff; |
} |
} |
|
|
struct ptree * |
|
parse_file(struct parse *p, int fd, const char *fname) |
/* |
|
* The 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. |
|
*/ |
|
static void |
|
parse_fd(struct parse *p, int fd) |
{ |
{ |
char b[4096]; |
char b[4096]; |
ssize_t rsz; /* Return value from read(2). */ |
ssize_t rsz; /* Return value from read(2). */ |
Line 857 parse_file(struct parse *p, int fd, const char *fname) |
|
Line 986 parse_file(struct parse *p, int fd, const char *fname) |
|
size_t poff; /* Parse offset in b[]. */ |
size_t poff; /* Parse offset in b[]. */ |
enum pstate pstate; |
enum pstate pstate; |
|
|
p->fname = fname; |
|
p->nline = 1; |
|
p->ncol = 1; |
|
pstate = PARSE_ELEM; |
|
rlen = 0; |
rlen = 0; |
|
pstate = PARSE_ELEM; |
/* |
|
* 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 && |
while ((rsz = read(fd, b + rlen, sizeof(b) - rlen - 1)) >= 0 && |
(rlen += rsz) > 0) { |
(rlen += rsz) > 0) { |
poff = parse_string(p, b, rlen, &pstate, rsz > 0); |
poff = parse_string(p, b, rlen, &pstate, rsz > 0); |
Line 883 parse_file(struct parse *p, int fd, const char *fname) |
|
Line 996 parse_file(struct parse *p, int fd, const char *fname) |
|
rlen -= poff; |
rlen -= poff; |
memmove(b, b + poff, rlen); |
memmove(b, b + poff, rlen); |
} |
} |
if (rsz < 0) { |
if (rsz < 0) |
perror(fname); |
error_msg(p, "read: %s", strerror(errno)); |
p->tree->flags |= TREE_FAIL; |
} |
|
|
|
/* |
|
* Open and parse a file. |
|
*/ |
|
struct ptree * |
|
parse_file(struct parse *p, int fd, const char *fname) |
|
{ |
|
const char *save_fname; |
|
int save_line, save_col; |
|
|
|
/* Save and initialize reporting data. */ |
|
|
|
save_fname = p->fname; |
|
save_line = p->nline; |
|
save_col = p->ncol; |
|
p->fname = fname; |
|
p->line = 0; |
|
p->col = 0; |
|
|
|
/* Open the file, unless it is already open. */ |
|
|
|
if (fd == -1 && (fd = open(fname, O_RDONLY, 0)) == -1) { |
|
error_msg(p, "open: %s", strerror(errno)); |
|
p->fname = save_fname; |
|
return p->tree; |
} |
} |
pnode_closetext(p); |
|
if ((p->tree->flags & TREE_CLOSED) == 0) |
/* |
warn_msg(p, "document not closed"); |
* After opening the starting file, change to the directory it |
|
* is located in, in case it wants to include any further files, |
|
* which are typically given with relative paths in DocBook. |
|
* Do this on a best-effort basis; don't complain about failure. |
|
*/ |
|
|
|
if (save_fname == NULL && (fname = dirname(fname)) != NULL && |
|
strcmp(fname, ".") != 0) |
|
(void)chdir(fname); |
|
|
|
/* Run the read loop. */ |
|
|
|
p->nline = 1; |
|
p->ncol = 1; |
|
parse_fd(p, fd); |
|
|
|
/* On the top level, finalize the parse tree. */ |
|
|
|
if (save_fname == NULL) { |
|
pnode_closetext(p); |
|
if (p->tree->root == NULL) |
|
error_msg(p, "empty document"); |
|
else if ((p->tree->flags & TREE_CLOSED) == 0) |
|
warn_msg(p, "document not closed"); |
|
pnode_unlink(p->doctype); |
|
} |
|
|
|
/* Clean up. */ |
|
|
|
if (fd != STDIN_FILENO) |
|
close(fd); |
|
p->fname = save_fname; |
|
p->nline = save_line; |
|
p->ncol = save_col; |
return p->tree; |
return p->tree; |
} |
} |