Expose "new" ap_parse_form_data() function instead of requiring

mod_request for any module that may want to parse form data...

git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/trunk@1072099 13f79535-47bb-0310-9956-ffa450edef68

1 files changed, 219 insertions, 3 deletions

diff --git a/server/util.c b/server/util.c
index b61aa8eb80..c443c942a3 100644
--- a/server/util.c
+++ b/server/util.c
@@ -2147,7 +2147,7 @@ AP_DECLARE(int) ap_request_has_body(request_rec *r)
     char *estr;
     const char *cls;
     int has_body;
-            
+
     has_body = (!r->header_only
                 && (r->kept_body
                     || apr_table_get(r->headers_in, "Transfer-Encoding")
@@ -2168,7 +2168,7 @@ AP_DECLARE_NONSTD(apr_status_t) ap_pool_cleanup_set_null(void *data_)
 }
 
 AP_DECLARE(apr_status_t) ap_str2_alnum(const char *src, char *dest) {
-    
+
     for ( ; *src; src++, dest++)
     {
         if (!apr_isprint(*src))
@@ -2180,7 +2180,7 @@ AP_DECLARE(apr_status_t) ap_str2_alnum(const char *src, char *dest) {
     }
     *dest = '\0';
     return APR_SUCCESS;
-    
+
 }
 
 AP_DECLARE(apr_status_t) ap_pstr2_alnum(apr_pool_t *p, const char *src,
@@ -2192,3 +2192,219 @@ AP_DECLARE(apr_status_t) ap_pstr2_alnum(apr_pool_t *p, const char *src,
     *dest = new;
     return ap_str2_alnum(src, new);
 }
+
+/**
+ * Read the body and parse any form found, which must be of the
+ * type application/x-www-form-urlencoded.
+ *
+ * Name/value pairs are returned in an array, with the names as
+ * strings with a maximum length of HUGE_STRING_LEN, and the
+ * values as bucket brigades. This allows values to be arbitrarily
+ * large.
+ *
+ * All url-encoding is removed from both the names and the values
+ * on the fly. The names are interpreted as strings, while the
+ * values are interpreted as blocks of binary data, that may
+ * contain the 0 character.
+ *
+ * In order to ensure that resource limits are not exceeded, a
+ * maximum size must be provided. If the sum of the lengths of
+ * the names and the values exceed this size, this function
+ * will return HTTP_REQUEST_ENTITY_TOO_LARGE.
+ *
+ * An optional number of parameters can be provided, if the number
+ * of parameters provided exceeds this amount, this function will
+ * return HTTP_REQUEST_ENTITY_TOO_LARGE. If this value is negative,
+ * no limit is imposed, and the number of parameters is in turn
+ * constrained by the size parameter above.
+ *
+ * This function honours any kept_body configuration, and the
+ * original raw request body will be saved to the kept_body brigade
+ * if so configured, just as ap_discard_request_body does.
+ *
+ * NOTE: File upload is not yet supported, but can be without change
+ * to the function call.
+ */
+
+/* form parsing stuff */
+typedef enum {
+    FORM_NORMAL,
+    FORM_AMP,
+    FORM_NAME,
+    FORM_VALUE,
+    FORM_PERCENTA,
+    FORM_PERCENTB,
+    FORM_ABORT
+} ap_form_type_t;
+
+AP_DECLARE(int) ap_parse_form_data(request_rec *r, ap_filter_t *f,
+                                   apr_array_header_t **ptr,
+                                   apr_size_t num, apr_size_t usize)
+{
+    apr_bucket_brigade *bb = NULL;
+    int seen_eos = 0;
+    char buffer[HUGE_STRING_LEN + 1];
+    const char *ct;
+    apr_size_t offset = 0;
+    apr_ssize_t size;
+    ap_form_type_t state = FORM_NAME, percent = FORM_NORMAL;
+    ap_form_pair_t *pair = NULL;
+    apr_array_header_t *pairs = apr_array_make(r->pool, 4, sizeof(ap_form_pair_t));
+
+    char hi = 0;
+    char low = 0;
+
+    *ptr = pairs;
+
+    /* sanity check - we only support forms for now */
+    ct = apr_table_get(r->headers_in, "Content-Type");
+    if (!ct || strcmp("application/x-www-form-urlencoded", ct)) {
+        return ap_discard_request_body(r);
+    }
+
+    if (usize > APR_SIZE_MAX >> 1)
+        size = APR_SIZE_MAX >> 1;
+    else
+        size = usize;
+
+    if (!f) {
+        f = r->input_filters;
+    }
+
+    bb = apr_brigade_create(r->pool, r->connection->bucket_alloc);
+    do {
+        apr_bucket *bucket = NULL, *last = NULL;
+
+        int rv = ap_get_brigade(f, bb, AP_MODE_READBYTES,
+                                APR_BLOCK_READ, HUGE_STRING_LEN);
+        if (rv != APR_SUCCESS) {
+            apr_brigade_destroy(bb);
+            return (rv == AP_FILTER_ERROR) ? rv : HTTP_BAD_REQUEST;
+        }
+
+        for (bucket = APR_BRIGADE_FIRST(bb);
+             bucket != APR_BRIGADE_SENTINEL(bb);
+             last = bucket, bucket = APR_BUCKET_NEXT(bucket)) {
+            const char *data;
+            apr_size_t len, slide;
+
+            if (last) {
+                apr_bucket_delete(last);
+            }
+            if (APR_BUCKET_IS_EOS(bucket)) {
+                seen_eos = 1;
+                break;
+            }
+            if (bucket->length == 0) {
+                continue;
+            }
+
+            rv = apr_bucket_read(bucket, &data, &len, APR_BLOCK_READ);
+            if (rv != APR_SUCCESS) {
+                apr_brigade_destroy(bb);
+                return HTTP_BAD_REQUEST;
+            }
+
+            slide = len;
+            while (state != FORM_ABORT && slide-- > 0 && size >= 0 && num != 0) {
+                char c = *data++;
+                if ('+' == c) {
+                    c = ' ';
+                }
+                else if ('&' == c) {
+                    state = FORM_AMP;
+                }
+                if ('%' == c) {
+                    percent = FORM_PERCENTA;
+                    continue;
+                }
+                if (FORM_PERCENTA == percent) {
+                    if (c >= 'a') {
+                        hi = c - 'a' + 10;
+                    }
+                    else if (c >= 'A') {
+                        hi = c - 'A' + 10;
+                    }
+                    else if (c >= '0') {
+                        hi = c - '0';
+                    }
+                    hi = hi << 4;
+                    percent = FORM_PERCENTB;
+                    continue;
+                }
+                if (FORM_PERCENTB == percent) {
+                    if (c >= 'a') {
+                        low = c - 'a' + 10;
+                    }
+                    else if (c >= 'A') {
+                        low = c - 'A' + 10;
+                    }
+                    else if (c >= '0') {
+                        low = c - '0';
+                    }
+                    c = low | hi;
+                    percent = FORM_NORMAL;
+                }
+                switch (state) {
+                    case FORM_AMP:
+                        if (pair) {
+                            const char *tmp = apr_pmemdup(r->pool, buffer, offset);
+                            apr_bucket *b = apr_bucket_pool_create(tmp, offset, r->pool, r->connection->bucket_alloc);
+                            APR_BRIGADE_INSERT_TAIL(pair->value, b);
+                        }
+                        state = FORM_NAME;
+                        pair = NULL;
+                        offset = 0;
+                        num--;
+                        break;
+                    case FORM_NAME:
+                        if (offset < HUGE_STRING_LEN) {
+                            if ('=' == c) {
+                                buffer[offset] = 0;
+                                offset = 0;
+                                pair = (ap_form_pair_t *) apr_array_push(pairs);
+                                pair->name = apr_pstrdup(r->pool, buffer);
+                                pair->value = apr_brigade_create(r->pool, r->connection->bucket_alloc);
+                                state = FORM_VALUE;
+                            }
+                            else {
+                                buffer[offset++] = c;
+                                size--;
+                            }
+                        }
+                        else {
+                            state = FORM_ABORT;
+                        }
+                        break;
+                    case FORM_VALUE:
+                        if (offset >= HUGE_STRING_LEN) {
+                            const char *tmp = apr_pmemdup(r->pool, buffer, offset);
+                            apr_bucket *b = apr_bucket_pool_create(tmp, offset, r->pool, r->connection->bucket_alloc);
+                            APR_BRIGADE_INSERT_TAIL(pair->value, b);
+                            offset = 0;
+                        }
+                        buffer[offset++] = c;
+                        size--;
+                        break;
+                    default:
+                        break;
+                }
+            }
+
+        }
+
+        apr_brigade_cleanup(bb);
+    } while (!seen_eos);
+
+    if (FORM_ABORT == state || size < 0 || num == 0) {
+        return HTTP_REQUEST_ENTITY_TOO_LARGE;
+    }
+    else if (FORM_VALUE == state && pair && offset > 0) {
+        const char *tmp = apr_pmemdup(r->pool, buffer, offset);
+        apr_bucket *b = apr_bucket_pool_create(tmp, offset, r->pool, r->connection->bucket_alloc);
+        APR_BRIGADE_INSERT_TAIL(pair->value, b);
+    }
+
+    return OK;
+
+}