spool_reqbody_cl(): mention in commentary that any data we
read into memory before reaching our threshold will not be
written to disk
tweak an existing comment, since "also" does not apply to
the situation where we chunk due to an environment variable
setting
git-svn-id: https://svn.apache.org/repos/asf/httpd/httpd/branches/proxy-reqbody@151147 13f79535-47bb-0310-9956-ffa450edef68
diff --git a/modules/proxy/proxy_http.c b/modules/proxy/proxy_http.c
index e3b7f8f..9b7628f 100644
--- a/modules/proxy/proxy_http.c
+++ b/modules/proxy/proxy_http.c
@@ -538,7 +538,10 @@
apr_brigade_length(input_brigade, 1, &bytes);
if (bytes_spooled + bytes > MAX_MEM_SPOOL) {
- /* can't spool any more in memory; write latest brigade to disk */
+ /* can't spool any more in memory; write latest brigade to disk;
+ * what we read into memory before reaching our threshold will
+ * remain there; we just write this and any subsequent data to disk
+ */
if (tmpfile == NULL) {
const char *temp_dir;
char *template;
@@ -649,8 +652,8 @@
* . if client sent C-L and there are no input resource filters, the
* the body size can't change so we send the same CL and stream the
* body
- * . if client used chunked or proxy-sendchunks is set, we'll also
- * use chunked
+ * . if client used chunked or proxy-sendchunks is set, we'll use
+ * chunked
*
* normal case:
* we have to compute content length by reading the entire request
