From: Nick Kew <niq@apache.org>
Date: Tue, 3 Oct 2006 07:41:27 +0000 (+0000)
Subject: Support regexp-based manipulation of HTTP headers
X-Git-Tag: 2.3.0~2096
X-Git-Url: http://git.ipfire.org/cgi-bin/gitweb.cgi?a=commitdiff_plain;h=4034b23cc0d02ed8dcca0c72c77f3a86244cb5a0;p=thirdparty%2Fapache%2Fhttpd.git

Support regexp-based manipulation of HTTP headers


git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/trunk@452330 13f79535-47bb-0310-9956-ffa450edef68
---

diff --git a/CHANGES b/CHANGES
index ef9b5436a7d..21c5fc007b9 100644
--- a/CHANGES
+++ b/CHANGES
@@ -2,6 +2,8 @@
 Changes with Apache 2.3.0
   [Remove entries to the current 2.0 and 2.2 section below, when backported]
 
+  *) mod_headers: support regexp-based editing of HTTP headers [Nick Kew]
+
   *) mod_echo: Fix precedence problem in if statement. PR 40658.
      [Larry Cipriani <lvc lucent.com>]
 
diff --git a/docs/manual/mod/mod_headers.xml b/docs/manual/mod/mod_headers.xml
index 7491245aa8f..3c325b034fa 100644
--- a/docs/manual/mod/mod_headers.xml
+++ b/docs/manual/mod/mod_headers.xml
@@ -144,6 +144,15 @@ is available only in Apache 2.0</compatibility>
          MyHeader: D=3775428 t=991424704447256 mytext
        </example>
       </li>
+      <li>Enable DAV to work with Apache running HTTP through SSL hardware
+      (<a href="http://svn.haxx.se/users/archive-2006-03/0549.shtml"
+      >problem description</a>) by replacing <var>https:</var> with
+      <var>http:</var> in the <var>Destination</var> header:
+      <example>
+        RequestHeader edit Destination ^https: http: early
+      </example>
+      </li>
+
     </ol>
 </section>
 
@@ -151,13 +160,13 @@ is available only in Apache 2.0</compatibility>
 <name>RequestHeader</name>
 <description>Configure HTTP request headers</description>
 <syntax>RequestHeader set|append|add|unset <var>header</var>
-[<var>value</var>] [early|env=[!]<var>variable</var>]</syntax>
+[<var>value</var>] [<var>replacement</var>] [early|env=[!]<var>variable</var>]</syntax>
 <contextlist><context>server config</context><context>virtual host</context>
 <context>directory</context><context>.htaccess</context></contextlist>
 <override>FileInfo</override>
 
 <usage>
-    <p>This directive can replace, merge or remove HTTP request
+    <p>This directive can replace, merge, change or remove HTTP request
     headers. The header is modified just before the content handler
     is run, allowing incoming headers to be modified. The action it
     performs is determined by the first argument. This can be one
@@ -186,18 +195,28 @@ is available only in Apache 2.0</compatibility>
     <dd>The request header of this name is removed, if it exists. If
     there are multiple headers of the same name, all will be removed.
     <var>value</var> must be omitted.</dd>
+
+    <dt><code>edit</code></dt>
+    <dd>If this request header exists, its value is transformed according
+    to a <glossary ref="regex">regular expression</glossary>
+    search-and-replace.  The <var>value</var> argument is a <glossary
+    ref="regex">regular expression</glossary>, and the <var>replacement</var>
+    is a replacement string, which may contain backreferences.</dd>
     </dl>
 
     <p>This argument is followed by a header name, which can
     include the final colon, but it is not required. Case is
     ignored. For <code>add</code>, <code>append</code> and
-    <code>set</code> a <var>value</var> is given as the third argument. If
+    <code>set</code> a <var>value</var> is given as the third argument. If a
     <var>value</var> contains spaces, it should be surrounded by double
     quotes. For unset, no <var>value</var> should be given.
     <var>value</var> may be a character string, a string containing format
     specifiers or a combination of both. The supported format specifiers
     are the same as for the <directive module="mod_headers">Header</directive>,
-    please have a look there for details.</p>
+    please have a look there for details.  For <code>edit</code> both
+    a <var>value</var> and a <var>replacement</var> are required, and are
+    a <glossary ref="regex">regular expression</glossary> and a
+    replacement string respectively.</p>
 
     <p>The <directive>RequestHeader</directive> directive may be followed by
     an additional argument, which may be used to specify conditions under
@@ -271,6 +290,13 @@ is available only in Apache 2.0</compatibility>
     response headers. <var>header</var> may be a 
     <glossary ref="regex">regular expression</glossary>.
     <var>value</var> must be omitted.</dd>
+
+    <dt><code>edit</code></dt>
+    <dd>If this request header exists, its value is transformed according
+    to a <glossary ref="regex">regular expression</glossary>
+    search-and-replace.  The <var>value</var> argument is a <glossary
+    ref="regex">regular expression</glossary>, and the <var>replacement</var>
+    is a replacement string, which may contain backreferences.</dd>
     </dl>
 
     <p>This argument is followed by a <var>header</var> name, which
@@ -322,6 +348,10 @@ is available only in Apache 2.0</compatibility>
       more efficient than <code>%s</code>.</p>
     </note> 
 
+    <p>For <code>edit</code> there is both a <var>value</var> argument
+    which is a <glossary ref="regex">regular expression</glossary>,
+    and an additional <var>replacement</var> string.</p>
+
     <p>The <directive>Header</directive> directive may be followed by an
     an additional argument, which may be used to specify conditions under
     which the action will be taken, or may be the keyword <code>early</code>
diff --git a/modules/metadata/mod_headers.c b/modules/metadata/mod_headers.c
index bbeb146124d..34174a1809c 100644
--- a/modules/metadata/mod_headers.c
+++ b/modules/metadata/mod_headers.c
@@ -89,7 +89,8 @@ typedef enum {
     hdr_set = 's',              /* set (replace old value) */
     hdr_append = 'm',           /* append (merge into any old value) */
     hdr_unset = 'u',            /* unset header */
-    hdr_echo = 'e'              /* echo headers from request to response */
+    hdr_echo = 'e',             /* echo headers from request to response */
+    hdr_edit = 'r'              /* change value by regexp */
 } hdr_actions;
 
 /*
@@ -119,6 +120,7 @@ typedef struct {
     apr_array_header_t *ta;   /* Array of format_tag structs */
     ap_regex_t *regex;
     const char *condition_var;
+    const char *subs;
 } header_entry;
 
 /* echo_do is used for Header echo to iterate through the request headers*/
@@ -348,6 +350,7 @@ static char *parse_format_string(apr_pool_t *p, header_entry *hdr, const char *s
 
     /* No string to parse with unset and echo commands */
     if (hdr->action == hdr_unset ||
+        hdr->action == hdr_edit ||
         hdr->action == hdr_echo) {
         return NULL;
     }
@@ -368,7 +371,8 @@ static APR_INLINE const char *header_inout_cmd(cmd_parms *cmd,
                                                const char *action,
                                                const char *hdr,
                                                const char *value,
-                                               const char* envclause)
+                                               const char *subs,
+                                               const char *envclause)
 {
     headers_conf *dirconf = indirconf;
     const char *condition_var = NULL;
@@ -392,10 +396,29 @@ static APR_INLINE const char *header_inout_cmd(cmd_parms *cmd,
         new->action = hdr_unset;
     else if (!strcasecmp(action, "echo"))
         new->action = hdr_echo;
+    else if (!strcasecmp(action, "edit"))
+        new->action = hdr_edit;
     else
-        return "first argument must be 'add', 'set', 'append', 'unset' or "
-               "'echo'.";
+        return "first argument must be 'add', 'set', 'append', 'unset', "
+               "'echo' or 'edit'.";
 
+    if (new->action == hdr_edit) {
+        if (subs == NULL) {
+            return "Header edit requires a match and a substitution";
+        }
+        new->regex = ap_pregcomp(cmd->pool, value, AP_REG_EXTENDED);
+        if (new->regex == NULL) {
+            return "Header edit regex could not be compiled";
+        }
+        new->subs = subs;
+    }
+    else {
+        /* there's no subs, so envclause is really that argument */
+        if (envclause != NULL) {
+            return "Too many arguments to directive";
+        }
+        envclause = subs;
+    }
     if (new->action == hdr_unset) {
         if (value) {
             if (envclause) {
@@ -465,6 +488,7 @@ static const char *header_cmd(cmd_parms *cmd, void *indirconf,
     const char *hdr;
     const char *val;
     const char *envclause;
+    const char *subs;
 
     action = ap_getword_conf(cmd->pool, &args);
     if (cmd->info == &hdr_out) {
@@ -478,6 +502,7 @@ static const char *header_cmd(cmd_parms *cmd, void *indirconf,
     }
     hdr = ap_getword_conf(cmd->pool, &args);
     val = *args ? ap_getword_conf(cmd->pool, &args) : NULL;
+    subs = *args ? ap_getword_conf(cmd->pool, &args) : NULL;
     envclause = *args ? ap_getword_conf(cmd->pool, &args) : NULL;
 
     if (*args) {
@@ -485,7 +510,7 @@ static const char *header_cmd(cmd_parms *cmd, void *indirconf,
                            " has too many arguments", NULL);
     }
 
-    return header_inout_cmd(cmd, indirconf, action, hdr, val, envclause);
+    return header_inout_cmd(cmd, indirconf, action, hdr, val, subs, envclause);
 }
 
 /*
@@ -512,6 +537,26 @@ static char* process_tags(header_entry *hdr, request_rec *r)
     }
     return str ? str : "";
 }
+static const char *process_regexp(header_entry *hdr, const char *value,
+                                  apr_pool_t *pool)
+{
+    unsigned int nmatch = 10;
+    ap_regmatch_t pmatch[10];
+    const char *subs;
+    char *ret;
+    int diffsz;
+    if (ap_regexec(hdr->regex, value, nmatch, pmatch, 0)) {
+        /* no match, nothing to do */
+        return value;
+    }
+    subs = ap_pregsub(pool, hdr->subs, value, nmatch, pmatch);
+    diffsz = strlen(subs) - (pmatch[0].rm_eo - pmatch[0].rm_so);
+    ret = apr_palloc(pool, strlen(value) + 1 + diffsz);
+    memcpy(ret, value, pmatch[0].rm_so);
+    strcpy(ret + pmatch[0].rm_so, subs);
+    strcat(ret, value + pmatch[0].rm_eo);
+    return ret;
+}
 
 static int echo_header(echo_do *v, const char *key, const char *val)
 {
@@ -528,7 +573,9 @@ static int echo_header(echo_do *v, const char *key, const char *val)
 static void do_headers_fixup(request_rec *r, apr_table_t *headers,
                              apr_array_header_t *fixup, int early)
 {
+    echo_do v;
     int i;
+    const char *val;
 
     for (i = 0; i < fixup->nelts; ++i) {
         header_entry *hdr = &((header_entry *) (fixup->elts))[i];
@@ -568,14 +615,18 @@ static void do_headers_fixup(request_rec *r, apr_table_t *headers,
             apr_table_unset(headers, hdr->header);
             break;
         case hdr_echo:
-        {
-            echo_do v;
             v.r = r;
             v.hdr = hdr;
             apr_table_do((int (*) (void *, const char *, const char *))
                          echo_header, (void *) &v, r->headers_in, NULL);
             break;
-        }
+        case hdr_edit:
+            val = apr_table_get(headers, hdr->header);
+            if (val != NULL) {
+                apr_table_setn(headers, hdr->header,
+                               process_regexp(hdr, val, r->pool));
+            }
+            break;
         }
     }
 }