Add benchmarks

[CapacitorSet]

As mentioned here.

A simple, state-of-the-art tool you could use is ApacheBench; I'll try it out and post results here.

[Glorf]

Hi, @CapacitorSet, thanks for your post
I've found that it's not easily applicable as I don't support HTTP/1.0 and its keepalive headers yet
I'll investigate it further today and publish results when I'll succeed

[CapacitorSet]

I did a quick and dirty fix:

diff --git a/http.c b/http.c
index 7334de0..0ea35f3 100644
--- a/http.c
+++ b/http.c
@@ -100,16 +100,18 @@ void parse_request_line(s_string *bareLine, s_http_request *request) {

         s_string http11 = create_string("HTTP/1.1", 8);

-        if(compare_string(&protocol, &http11)) {
+        // if(compare_string(&protocol, &http11)) {
             message_log("Requested resource: ", DEBUG);
             string_log(&request->resource, DEBUG);
             request->status = OK;
+        /*
         }
         else {
             message_log("HTTP version unsupported", DEBUG);
             request->status = HTTP_VERSION_NOT_SUPPORTED;
             return;
         }
+        */

         delete_string(http11);
     }
diff --git a/worker.c b/worker.c
index 3f8fb79..f1420e9 100644
--- a/worker.c
+++ b/worker.c
@@ -93,6 +93,7 @@ int create_worker() {
                     message_log("Error 500", ERR);
                     cli_connection->currentRequest->status = BAD_REQUEST;
                 }
+                close_client_connection(cli_connection);
             }
             else if(event_queue[i].events & EPOLLOUT) { //one of connected clients is ready to read
                 if(cli_connection->response_buffer.size > 0) { //there is some data remaining so let's send it

[CapacitorSet]

With this patch I did a simple run, which is far from a complete benchmark but a starting point:

Nginx (ab -H 'Host: xxxxx' -c 100 -n 100000 http://127.0.0.1/index.html):

Server Hostname:        127.0.0.1
Server Port:            80

Document Path:          /index.html
Document Length:        2588 bytes

Concurrency Level:      100
Time taken for tests:   10.543 seconds
Complete requests:      100000
Failed requests:        0
Write errors:           0
Total transferred:      282200000 bytes
HTML transferred:       258800000 bytes
Requests per second:    9484.95 [#/sec] (mean)
Time per request:       10.543 [ms] (mean)
Time per request:       0.105 [ms] (mean, across all concurrent requests)
Transfer rate:          26139.18 [Kbytes/sec] received

Connection Times (ms)
              min  mean[+/-sd] median   max
Connect:        0    5   0.7      5       8
Processing:     1    6   1.0      6      15
Waiting:        0    4   1.2      4      13
Total:          2   11   0.7     10      19
WARNING: The median and mean for the total time are not within a normal deviation
        These results are probably not that reliable.

Percentage of the requests served within a certain time (ms)
  50%     10
  66%     11
  75%     11
  80%     11
  90%     11
  95%     12
  98%     12
  99%     12
 100%     19 (longest request)

Lear ( ab -c 100 -n 100000 http://127.0.0.1:9000/index.html):

Server Software:
Server Hostname:        127.0.0.1
Server Port:            9000

Document Path:          /index.html
Document Length:        2588 bytes

Concurrency Level:      100
Time taken for tests:   10.344 seconds
Complete requests:      100000
Failed requests:        0
Write errors:           0
Total transferred:      262900000 bytes
HTML transferred:       258800000 bytes
Requests per second:    9667.88 [#/sec] (mean)
Time per request:       10.344 [ms] (mean)
Time per request:       0.103 [ms] (mean, across all concurrent requests)
Transfer rate:          24821.15 [Kbytes/sec] received

Connection Times (ms)
              min  mean[+/-sd] median   max
Connect:        0    4   0.7      4       8
Processing:     1    6   1.0      6      11
Waiting:        0    5   1.1      5      10
Total:          5   10   1.2     10      15

Percentage of the requests served within a certain time (ms)
  50%     10
  66%     11
  75%     11
  80%     11
  90%     12
  95%     12
  98%     13
  99%     13
 100%     15 (longest request)

[Glorf]

Thanks! It looks quite impressive. I'll try to eliminate major bottlenecks and give it a try again :)

[Glorf]

@CapacitorSet, the server should be benchmarkable by ab now, but some of my changes severly degraded performance. I'd like to tweak it during next days and create some visual result to add to readme

[Glorf]

I've added sample ab results to README, and will update it with every new commit (I'm currently profiling the server for better performance). While serving small files, it appears to be slightly better than nginx for single-threaded, and MUCH better for multi-threaded environments
For big files, it's much worse than nginx in any case.
Also, please run any other test you'd like