This patch fixes three things:
- older libexpat does not know about enum XML_Status
- as in my patch for http-fetch, do not rely on a curl result in
free()d data
- calloc the new_lock structure
Signed-off-by: Johannes Schindelin <Johannes.Schindelin@gmx.de>
---
I verified in a small scale test that it works now. The buglets
showed only on my iBook, not on any Linux box I tried.
The calloc'ing of the new_lock structure is necessary, because
activelock_cdata() relies on several members being 0 until
set by start_activelock_element().
The problem I was experiencing with a crash in
cleanup_multi_handle() was due to the curl result not being
interpreted correctly.
Unfortunately, valgrind does not work on macosx, so it could
not help me.
Note: http-push shares a lot of code with http-fetch. It might
be a good idea to refactor them (after some more out-ironing).
http-push.c | 14 ++++++++++++--
1 files changed, 12 insertions(+), 2 deletions(-)
Show changes to http-push.c +12 −2
diff --git a/http-push.c b/http-push.c
index c10067c..5532d02 100644
--- a/http-push.c
+++ b/http-push.c
@@ -28,6 +28,15 @@ static const char http_push_usage[] =
#define NO_CURL_EASY_DUPHANDLE
#endif
+#ifndef XML_STATUS_OK
+enum XML_Status {
+ XML_STATUS_OK = 1,
+ XML_STATUS_ERROR = 0
+};
+#define XML_STATUS_OK 1
+#define XML_STATUS_ERROR 0
+#endif
+
#define RANGE_HEADER_SIZE 30
/* DAV method names and request body templates */
@@ -728,11 +737,12 @@ void process_curl_messages(void)
slot->curl != curl_message->easy_handle)
slot = slot->next;
if (slot != NULL) {
+ int curl_result = curl_message->data.result;
curl_multi_remove_handle(curlm, slot->curl);
active_requests--;
slot->done = 1;
slot->in_use = 0;
- slot->curl_result = curl_message->data.result;
+ slot->curl_result = curl_result;
curl_easy_getinfo(slot->curl,
CURLINFO_HTTP_CODE,
&slot->http_code);
@@ -1230,7 +1240,7 @@ struct active_lock *lock_remote(char *fi
in_buffer.posn = 0;
in_buffer.buffer = in_data;
- new_lock = xmalloc(sizeof(*new_lock));
+ new_lock = xcalloc(1, sizeof(*new_lock));
new_lock->owner = NULL;
new_lock->token = NULL;
new_lock->timeout = -1;
Johannes Schindelin