VladPolskiy/apache-http-server
1
0
Fork 0
mirror of https://github.com/apache/httpd.git synced 2025-08-03 16:33:59 +00:00
Code Issues Packages Projects Releases Wiki Activity

Joe Orton says:

Browse Source 
This patch is a first pass at making util_xml.c independent of mod_dav:

* Rename dav_* to ap_xml_* for exported interface
* Remove dav_* from private functions
* Move text_header handling in from dav/main/util.c
* Move XML tree -> text conversion in
* Add include/util_xml.h header from parts of mod_dav.h

Submitted by: Joe Orton <joe@orton.demon.co.uk>
Reviewed by: Greg Stein


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/trunk@85760 13f79535-47bb-0310-9956-ffa450edef68
This commit is contained in:
Greg Stein
2000-07-03 22:28:38 +00:00
parent a9799bce1f
commit ec10e40b85
2 changed files with 668 additions and 57 deletions
Download Patch File Download Diff File Expand all files Collapse all files

505
server/util_xml.c
View File

@ -64,52 +64,55 @@
#include "http_protocol.h"
#include "http_log.h"
/* ### need to fix this... */
#include "../modules/dav/main/mod_dav.h"
#include "util_xml.h"
#define DEBUG_CR "\r\n"
#define AP_XML_READ_BLOCKSIZE 2048 /* used for reading input blocks */
/* errors related to namespace processing */
#define DAV_NS_ERROR_UNKNOWN_PREFIX (DAV_NS_ERROR_BASE)
#define AP_XML_NS_ERROR_UNKNOWN_PREFIX (AP_XML_NS_ERROR_BASE)
/* test for a namespace prefix that begins with [Xx][Mm][Ll] */
#define DAV_NS_IS_RESERVED(name) \
#define AP_XML_NS_IS_RESERVED(name) \
( (name[0] == 'X' || name[0] == 'x') && \
(name[1] == 'M' || name[1] == 'm') && \
(name[2] == 'L' || name[2] == 'l') )
/* content for parsing */
typedef struct dav_xml_ctx {
dav_xml_doc *doc; /* the doc we're parsing */
typedef struct ap_xml_ctx {
ap_xml_doc *doc; /* the doc we're parsing */
ap_pool_t *p; /* the pool we allocate from */
dav_xml_elem *cur_elem; /* current element */
ap_xml_elem *cur_elem; /* current element */
int error; /* an error has occurred */
/* errors may be DAV_NS_ERROR_* or other errors defined here (none yet) */
/* errors may be AP_XML_NS_ERROR_* or other private errors which will
be defined here (none yet) */
} dav_xml_ctx;
} ap_xml_ctx;
/* struct for scoping namespace declarations */
typedef struct dav_xml_ns_scope {
typedef struct ap_xml_ns_scope {
const char *prefix; /* prefix used for this ns */
int ns; /* index into namespace table */
int emptyURI; /* the namespace URI is the empty string */
struct dav_xml_ns_scope *next; /* next scoped namespace */
} dav_xml_ns_scope;
/* ### need a similar mechanism for xml:lang values */
struct ap_xml_ns_scope *next; /* next scoped namespace */
} ap_xml_ns_scope;
/* return namespace table index for a given prefix */
static int dav_find_prefix(dav_xml_ctx *ctx, const char *prefix)
static int find_prefix(ap_xml_ctx *ctx, const char *prefix)
{
dav_xml_elem *elem = ctx->cur_elem;
ap_xml_elem *elem = ctx->cur_elem;
/*
** Walk up the tree, looking for a namespace scope that defines this
** prefix.
*/
for (; elem; elem = elem->parent) {
dav_xml_ns_scope *ns_scope = elem->ns_scope;
ap_xml_ns_scope *ns_scope = elem->ns_scope;
for (ns_scope = elem->ns_scope; ns_scope; ns_scope = ns_scope->next) {
if (strcmp(prefix, ns_scope->prefix) == 0) {
@ -120,7 +123,7 @@ static int dav_find_prefix(dav_xml_ctx *ctx, const char *prefix)
** to mean "no namespace." We just found the prefix
** refers to an empty URI, so return "no namespace."
*/
return DAV_NS_NONE;
return AP_XML_NS_NONE;
}
return ns_scope->ns;
@ -136,19 +139,19 @@ static int dav_find_prefix(dav_xml_ctx *ctx, const char *prefix)
* has "no namespace". We have a reserved value for this.
*/
if (*prefix == '\0') {
return DAV_NS_NONE;
return AP_XML_NS_NONE;
}
/* not found */
return DAV_NS_ERROR_UNKNOWN_PREFIX;
return AP_XML_NS_ERROR_UNKNOWN_PREFIX;
}
static void dav_start_handler(void *userdata, const char *name, const char **attrs)
static void start_handler(void *userdata, const char *name, const char **attrs)
{
dav_xml_ctx *ctx = userdata;
dav_xml_elem *elem;
dav_xml_attr *attr;
dav_xml_attr *prev;
ap_xml_ctx *ctx = userdata;
ap_xml_elem *elem;
ap_xml_attr *attr;
ap_xml_attr *prev;
char *colon;
const char *quoted;
char *elem_name;
@ -201,7 +204,7 @@ static void dav_start_handler(void *userdata, const char *name, const char **att
attr = attr->next) {
if (strncmp(attr->name, "xmlns", 5) == 0) {
const char *prefix = &attr->name[5];
dav_xml_ns_scope *ns_scope;
ap_xml_ns_scope *ns_scope;
/* test for xmlns:foo= form and xmlns= form */
if (*prefix == ':')
@ -213,12 +216,12 @@ static void dav_start_handler(void *userdata, const char *name, const char **att
}
/* quote the URI before we ever start working with it */
quoted = dav_quote_string(ctx->p, attr->value, 1);
quoted = ap_xml_quote_string(ctx->p, attr->value, 1);
/* build and insert the new scope */
ns_scope = ap_pcalloc(ctx->p, sizeof(*ns_scope));
ns_scope->prefix = prefix;
ns_scope->ns = dav_insert_uri(ctx->doc->namespaces, quoted);
ns_scope->ns = ap_xml_insert_uri(ctx->doc->namespaces, quoted);
ns_scope->emptyURI = *quoted == '\0';
ns_scope->next = elem->ns_scope;
elem->ns_scope = ns_scope;
@ -233,7 +236,7 @@ static void dav_start_handler(void *userdata, const char *name, const char **att
}
else if (strcmp(attr->name, "xml:lang") == 0) {
/* save away the language (in quoted form) */
elem->lang = dav_quote_string(ctx->p, attr->value, 1);
elem->lang = ap_xml_quote_string(ctx->p, attr->value, 1);
/* remove this attribute from the element */
if (prev == NULL)
@ -253,7 +256,7 @@ static void dav_start_handler(void *userdata, const char *name, const char **att
** If an xml:lang attribute didn't exist (lang==NULL), then copy the
** language from the parent element (if present).
**
** NOTE: dav_elem_size() *depends* upon this pointer equality.
** NOTE: elem_size() *depends* upon this pointer equality.
*/
if (elem->lang == NULL && elem->parent != NULL)
elem->lang = elem->parent->lang;
@ -266,17 +269,17 @@ static void dav_start_handler(void *userdata, const char *name, const char **att
* be found. Either it will be "no namespace", or a default
* namespace URI has been specified at some point.
*/
elem->ns = dav_find_prefix(ctx, "");
elem->ns = find_prefix(ctx, "");
}
else if (DAV_NS_IS_RESERVED(elem->name)) {
elem->ns = DAV_NS_NONE;
else if (AP_XML_NS_IS_RESERVED(elem->name)) {
elem->ns = AP_XML_NS_NONE;
}
else {
*colon = '\0';
elem->ns = dav_find_prefix(ctx, elem->name);
elem->ns = find_prefix(ctx, elem->name);
elem->name = colon + 1;
if (DAV_NS_IS_ERROR(elem->ns)) {
if (AP_XML_NS_IS_ERROR(elem->ns)) {
ctx->error = elem->ns;
return;
}
@ -285,7 +288,7 @@ static void dav_start_handler(void *userdata, const char *name, const char **att
/* adjust all remaining attributes' namespaces */
for (attr = elem->attr; attr; attr = attr->next) {
/*
* dav_xml_attr defines this as "const" but we dup'd it, so we
* ap_xml_attr defines this as "const" but we dup'd it, so we
* know that we can change it. a bit hacky, but the existing
* structure def is best.
*/
@ -297,17 +300,17 @@ static void dav_start_handler(void *userdata, const char *name, const char **att
* Attributes do NOT use the default namespace. Therefore,
* we place them into the "no namespace" category.
*/
attr->ns = DAV_NS_NONE;
attr->ns = AP_XML_NS_NONE;
}
else if (DAV_NS_IS_RESERVED(attr->name)) {
attr->ns = DAV_NS_NONE;
else if (AP_XML_NS_IS_RESERVED(attr->name)) {
attr->ns = AP_XML_NS_NONE;
}
else {
*colon = '\0';
attr->ns = dav_find_prefix(ctx, attr->name);
attr->ns = find_prefix(ctx, attr->name);
attr->name = colon + 1;
if (DAV_NS_IS_ERROR(attr->ns)) {
if (AP_XML_NS_IS_ERROR(attr->ns)) {
ctx->error = attr->ns;
return;
}
@ -315,9 +318,9 @@ static void dav_start_handler(void *userdata, const char *name, const char **att
}
}
static void dav_end_handler(void *userdata, const char *name)
static void end_handler(void *userdata, const char *name)
{
dav_xml_ctx *ctx = userdata;
ap_xml_ctx *ctx = userdata;
/* punt once we find an error */
if (ctx->error)
@ -327,11 +330,11 @@ static void dav_end_handler(void *userdata, const char *name)
ctx->cur_elem = ctx->cur_elem->parent;
}
static void dav_cdata_handler(void *userdata, const char *data, int len)
static void cdata_handler(void *userdata, const char *data, int len)
{
dav_xml_ctx *ctx = userdata;
dav_xml_elem *elem;
dav_text_header *hdr;
ap_xml_ctx *ctx = userdata;
ap_xml_elem *elem;
ap_text_header *hdr;
const char *s;
/* punt once we find an error */
@ -350,13 +353,13 @@ static void dav_cdata_handler(void *userdata, const char *data, int len)
hdr = &elem->last_child->following_cdata;
}
dav_text_append(ctx->p, hdr, s);
ap_text_append(ctx->p, hdr, s);
}
int dav_parse_input(request_rec * r, dav_xml_doc **pdoc)
API_EXPORT(int) ap_xml_parse_input(request_rec * r, ap_xml_doc **pdoc)
{
int result;
dav_xml_ctx ctx =
ap_xml_ctx ctx =
{0};
XML_Parser parser;
@ -372,7 +375,7 @@ int dav_parse_input(request_rec * r, dav_xml_doc **pdoc)
ctx.doc = ap_pcalloc(ctx.p, sizeof(*ctx.doc));
ctx.doc->namespaces = ap_make_array(ctx.p, 5, sizeof(const char *));
dav_insert_uri(ctx.doc->namespaces, "DAV:");
ap_xml_insert_uri(ctx.doc->namespaces, "DAV:");
/* ### we should get the encoding from Content-Encoding */
parser = XML_ParserCreate(NULL);
@ -383,8 +386,8 @@ int dav_parse_input(request_rec * r, dav_xml_doc **pdoc)
}
XML_SetUserData(parser, (void *) &ctx);
XML_SetElementHandler(parser, dav_start_handler, dav_end_handler);
XML_SetCharacterDataHandler(parser, dav_cdata_handler);
XML_SetElementHandler(parser, start_handler, end_handler);
XML_SetCharacterDataHandler(parser, cdata_handler);
if (ap_should_client_block(r)) {
long len;
@ -392,13 +395,13 @@ int dav_parse_input(request_rec * r, dav_xml_doc **pdoc)
char end;
int rv;
size_t total_read = 0;
size_t limit_xml_body = dav_get_limit_xml_body(r);
size_t limit_xml_body = 1000000; /* ### fix this */
/* allocate our working buffer */
buffer = ap_palloc(r->pool, DAV_READ_BLOCKSIZE);
buffer = ap_palloc(r->pool, AP_XML_READ_BLOCKSIZE);
/* read the body, stuffing it into the parser */
while ((len = ap_get_client_block(r, buffer, DAV_READ_BLOCKSIZE)) > 0) {
while ((len = ap_get_client_block(r, buffer, AP_XML_READ_BLOCKSIZE)) > 0) {
total_read += len;
if (limit_xml_body && total_read > limit_xml_body) {
ap_log_rerror(APLOG_MARK, APLOG_ERR | APLOG_NOERRNO, 0, r,
@ -426,7 +429,7 @@ int dav_parse_input(request_rec * r, dav_xml_doc **pdoc)
if (ctx.error) {
switch (ctx.error) {
case DAV_NS_ERROR_UNKNOWN_PREFIX:
case AP_XML_NS_ERROR_UNKNOWN_PREFIX:
ap_log_rerror(APLOG_MARK, APLOG_ERR | APLOG_NOERRNO, 0, r,
"An undefined namespace prefix was used.");
break;
@ -468,3 +471,391 @@ int dav_parse_input(request_rec * r, dav_xml_doc **pdoc)
/* Apache will supply a default error, plus whatever was logged. */
return HTTP_BAD_REQUEST;
}
API_EXPORT(void) ap_text_append(ap_pool_t * p, ap_text_header *hdr,
const char *text)
{
ap_text *t = ap_palloc(p, sizeof(*t));
t->text = text;
t->next = NULL;
if (hdr->first == NULL) {
/* no text elements yet */
hdr->first = hdr->last = t;
}
else {
/* append to the last text element */
hdr->last->next = t;
hdr->last = t;
}
}
/* ---------------------------------------------------------------
**
** XML UTILITY FUNCTIONS
*/
/*
** ap_xml_quote_string: quote an XML string
**
** Replace '<', '>', and '&' with '&lt;', '&gt;', and '&amp;'.
** If quotes is true, then replace '"' with '&quot;'.
**
** quotes is typically set to true for XML strings that will occur within
** double quotes -- attribute values.
*/
API_EXPORT(const char *) ap_xml_quote_string(ap_pool_t *p, const char *s,
int quotes)
{
const char *scan;
int len = 0;
int extra = 0;
char *qstr;
char *qscan;
char c;
for (scan = s; (c = *scan) != '\0'; ++scan, ++len) {
if (c == '<' || c == '>')
extra += 3; /* &lt; or &gt; */
else if (c == '&')
extra += 4; /* &amp; */
else if (quotes && c == '"')
extra += 5; /* &quot; */
}
/* nothing to do? */
if (extra == 0)
return s;
qstr = ap_palloc(p, len + extra + 1);
for (scan = s, qscan = qstr; (c = *scan) != '\0'; ++scan) {
if (c == '<') {
*qscan++ = '&';
*qscan++ = 'l';
*qscan++ = 't';
*qscan++ = ';';
}
else if (c == '>') {
*qscan++ = '&';
*qscan++ = 'g';
*qscan++ = 't';
*qscan++ = ';';
}
else if (c == '&') {
*qscan++ = '&';
*qscan++ = 'a';
*qscan++ = 'm';
*qscan++ = 'p';
*qscan++ = ';';
}
else if (quotes && c == '"') {
*qscan++ = '&';
*qscan++ = 'q';
*qscan++ = 'u';
*qscan++ = 'o';
*qscan++ = 't';
*qscan++ = ';';
}
else {
*qscan++ = c;
}
}
*qscan = '\0';
return qstr;
}
/* how many characters for the given integer? */
#define AP_XML_NS_LEN(ns) ((ns) < 10 ? 1 : (ns) < 100 ? 2 : (ns) < 1000 ? 3 : \
(ns) < 10000 ? 4 : (ns) < 100000 ? 5 : \
(ns) < 1000000 ? 6 : (ns) < 10000000 ? 7 : \
(ns) < 100000000 ? 8 : (ns) < 1000000000 ? 9 : 10)
static int text_size(const ap_text *t)
{
int size = 0;
for (; t; t = t->next)
size += strlen(t->text);
return size;
}
static size_t elem_size(const ap_xml_elem *elem, int style,
ap_array_header_t *namespaces, int *ns_map)
{
size_t size;
if (style == AP_XML_X2T_FULL || style == AP_XML_X2T_FULL_NS_LANG) {
const ap_xml_attr *attr;
size = 0;
if (style == AP_XML_X2T_FULL_NS_LANG) {
int i;
/*
** The outer element will contain xmlns:ns%d="%s" attributes
** and an xml:lang attribute, if applicable.
*/
for (i = namespaces->nelts; i--;) {
/* compute size of: ' xmlns:ns%d="%s"' */
size += (9 + AP_XML_NS_LEN(i) + 2 +
strlen(AP_XML_GET_URI_ITEM(namespaces, i)) + 1);
}
if (elem->lang != NULL) {
/* compute size of: ' xml:lang="%s"' */
size += 11 + strlen(elem->lang) + 1;
}
}
if (elem->ns == AP_XML_NS_NONE) {
/* compute size of: <%s> */
size += 1 + strlen(elem->name) + 1;
}
else {
int ns = ns_map ? ns_map[elem->ns] : elem->ns;
/* compute size of: <ns%d:%s> */
size += 3 + AP_XML_NS_LEN(ns) + 1 + strlen(elem->name) + 1;
}
if (AP_XML_ELEM_IS_EMPTY(elem)) {
/* insert a closing "/" */
size += 1;
}
else {
/*
* two of above plus "/":
* <ns%d:%s> ... </ns%d:%s>
* OR <%s> ... </%s>
*/
size = 2 * size + 1;
}
for (attr = elem->attr; attr; attr = attr->next) {
if (attr->ns == AP_XML_NS_NONE) {
/* compute size of: ' %s="%s"' */
size += 1 + strlen(attr->name) + 2 + strlen(attr->value) + 1;
}
else {
/* compute size of: ' ns%d:%s="%s"' */
size += 3 + AP_XML_NS_LEN(attr->ns) + 1 + strlen(attr->name) + 2 + strlen(attr->value) + 1;
}
}
/*
** If the element has an xml:lang value that is *different* from
** its parent, then add the thing in: ' xml:lang="%s"'.
**
** NOTE: we take advantage of the pointer equality established by
** the parsing for "inheriting" the xml:lang values from parents.
*/
if (elem->lang != NULL &&
(elem->parent == NULL || elem->lang != elem->parent->lang)) {
size += 11 + strlen(elem->lang) + 1;
}
}
else if (style == AP_XML_X2T_LANG_INNER) {
/*
* This style prepends the xml:lang value plus a null terminator.
* If a lang value is not present, then we insert a null term.
*/
size = elem->lang ? strlen(elem->lang) + 1 : 1;
}
else
size = 0;
size += text_size(elem->first_cdata.first);
for (elem = elem->first_child; elem; elem = elem->next) {
/* the size of the child element plus the CDATA that follows it */
size += (elem_size(elem, AP_XML_X2T_FULL, NULL, ns_map) +
text_size(elem->following_cdata.first));
}
return size;
}
static char *write_text(char *s, const ap_text *t)
{
for (; t; t = t->next) {
size_t len = strlen(t->text);
memcpy(s, t->text, len);
s += len;
}
return s;
}
static char *write_elem(char *s, const ap_xml_elem *elem, int style,
ap_array_header_t *namespaces, int *ns_map)
{
const ap_xml_elem *child;
size_t len;
int ns;
if (style == AP_XML_X2T_FULL || style == AP_XML_X2T_FULL_NS_LANG) {
int empty = AP_XML_ELEM_IS_EMPTY(elem);
const ap_xml_attr *attr;
if (elem->ns == AP_XML_NS_NONE) {
len = sprintf(s, "<%s", elem->name);
}
else {
ns = ns_map ? ns_map[elem->ns] : elem->ns;
len = sprintf(s, "<ns%d:%s", ns, elem->name);
}
s += len;
for (attr = elem->attr; attr; attr = attr->next) {
if (attr->ns == AP_XML_NS_NONE)
len = sprintf(s, " %s=\"%s\"", attr->name, attr->value);
else
len = sprintf(s, " ns%d:%s=\"%s\"", attr->ns, attr->name, attr->value);
s += len;
}
/* add the xml:lang value if necessary */
if (elem->lang != NULL &&
(style == AP_XML_X2T_FULL_NS_LANG ||
elem->parent == NULL ||
elem->lang != elem->parent->lang)) {
len = sprintf(s, " xml:lang=\"%s\"", elem->lang);
s += len;
}
/* add namespace definitions, if required */
if (style == AP_XML_X2T_FULL_NS_LANG) {
int i;
for (i = namespaces->nelts; i--;) {
len = sprintf(s, " xmlns:ns%d=\"%s\"", i,
AP_XML_GET_URI_ITEM(namespaces, i));
s += len;
}
}
/* no more to do. close it up and go. */
if (empty) {
*s++ = '/';
*s++ = '>';
return s;
}
/* just close it */
*s++ = '>';
}
else if (style == AP_XML_X2T_LANG_INNER) {
/* prepend the xml:lang value */
if (elem->lang != NULL) {
len = strlen(elem->lang);
memcpy(s, elem->lang, len);
s += len;
}
*s++ = '\0';
}
s = write_text(s, elem->first_cdata.first);
for (child = elem->first_child; child; child = child->next) {
s = write_elem(s, child, AP_XML_X2T_FULL, NULL, ns_map);
s = write_text(s, child->following_cdata.first);
}
if (style == AP_XML_X2T_FULL || style == AP_XML_X2T_FULL_NS_LANG) {
if (elem->ns == AP_XML_NS_NONE) {
len = sprintf(s, "</%s>", elem->name);
}
else {
ns = ns_map ? ns_map[elem->ns] : elem->ns;
len = sprintf(s, "</ns%d:%s>", ns, elem->name);
}
s += len;
}
return s;
}
API_EXPORT(void) ap_xml_quote_elem(ap_pool_t *p, ap_xml_elem *elem)
{
ap_text *scan_txt;
ap_xml_attr *scan_attr;
ap_xml_elem *scan_elem;
/* convert the element's text */
for (scan_txt = elem->first_cdata.first;
scan_txt != NULL;
scan_txt = scan_txt->next) {
scan_txt->text = ap_xml_quote_string(p, scan_txt->text, 0);
}
for (scan_txt = elem->following_cdata.first;
scan_txt != NULL;
scan_txt = scan_txt->next) {
scan_txt->text = ap_xml_quote_string(p, scan_txt->text, 0);
}
/* convert the attribute values */
for (scan_attr = elem->attr;
scan_attr != NULL;
scan_attr = scan_attr->next) {
scan_attr->value = ap_xml_quote_string(p, scan_attr->value, 1);
}
/* convert the child elements */
for (scan_elem = elem->first_child;
scan_elem != NULL;
scan_elem = scan_elem->next) {
ap_xml_quote_elem(p, scan_elem);
}
}
/* convert an element to a text string */
API_EXPORT(void) ap_xml_to_text(ap_pool_t * p, const ap_xml_elem *elem,
int style, ap_array_header_t *namespaces,
int *ns_map, const char **pbuf, size_t *psize)
{
/* get the exact size, plus a null terminator */
size_t size = elem_size(elem, style, namespaces, ns_map) + 1;
char *s = ap_palloc(p, size);
(void) write_elem(s, elem, style, namespaces, ns_map);
s[size - 1] = '\0';
*pbuf = s;
if (psize)
*psize = size;
}
API_EXPORT(const char *) ap_xml_empty_elem(ap_pool_t * p,
const ap_xml_elem *elem)
{
if (elem->ns == AP_XML_NS_NONE) {
/*
* The prefix (xml...) is already within the prop name, or
* the element simply has no prefix.
*/
return ap_psprintf(p, "<%s/>" DEBUG_CR, elem->name);
}
return ap_psprintf(p, "<ns%d:%s/>" DEBUG_CR, elem->ns, elem->name);
}
/* return the URI's (existing) index, or insert it and return a new index */
API_EXPORT(int) ap_xml_insert_uri(ap_array_header_t *uri_array,
const char *uri)
{
int i;
const char **pelt;
for (i = uri_array->nelts; i--;) {
if (strcmp(uri, AP_XML_GET_URI_ITEM(uri_array, i)) == 0)
return i;
}
pelt = ap_push_array(uri_array);
*pelt = uri; /* assume uri is const or in a pool */
return uri_array->nelts - 1;
}