diff options
| author | thiagoftsm <thiagoftsm@gmail.com> | 2019-08-28 15:41:13 +0000 |
|---|---|---|
| committer | Chris Akritidis <43294513+cakrit@users.noreply.github.com> | 2019-08-28 17:41:13 +0200 |
| commit | e972d27c5496ef3a9b7b9c3af2ce0e916e74f600 (patch) | |
| tree | 0c1b7eeccdcf0b1754c1c8957352c8b35190569a /web/server | |
| parent | b9e3d71126c81199de00b21d9e9335f443b64bbe (diff) | |
url_registry: Remove of unecessary NULL (#6714)
Summary
With the recent changes on URL parsing, we caused an issue in the registry. Requests were not being logged, so the counters remained the same. The reason was that we modified the request buffer and the daemon couldn't access the cookie required to store the info.
After @cosmix review, I notice that there was a code that is not necessary on Netdata. I executed tests with the stress.sh and urls/request.sh scripts, this last I used during 4 hours with cron calling it each hour.
Diffstat (limited to 'web/server')
| -rw-r--r-- | web/server/web_client.c | 3 |
1 files changed, 0 insertions, 3 deletions
diff --git a/web/server/web_client.c b/web/server/web_client.c index abd37a8f49..908e3a6a96 100644 --- a/web/server/web_client.c +++ b/web/server/web_client.c @@ -1065,9 +1065,6 @@ static inline HTTP_VALIDATION http_request_validate(struct web_client *w) { // copy the URL - we are going to overwrite parts of it // TODO -- ideally we we should avoid copying buffers around strncpyz(w->last_url, w->decoded_url, NETDATA_WEB_REQUEST_URL_SIZE); - if (w->url_search_path && w->separator) { - *w->url_search_path = 0x00; - } #ifdef ENABLE_HTTPS if ( (!web_client_check_unix(w)) && (netdata_srv_ctx) ) { if ((w->ssl.conn) && ((w->ssl.flags & NETDATA_SSL_NO_HANDSHAKE) && (web_client_is_using_ssl_force(w) || web_client_is_using_ssl_default(w)) && (w->mode != WEB_CLIENT_MODE_STREAM)) ) { |