| .gitignore | ||
| example.c | ||
| help.c | ||
| help.h | ||
| LICENSE | ||
| Makefile | ||
| README.md | ||
| webserver.c | ||
| webserver.h | ||
Serwer
Standalone webserver written in C99
What is this shit?
Well, it's standalone webserver written entirely in C99 (no libs required).
Why would someone need this?
I've no idea. But I'm going to use this for my next project.
Why is it's name to weird?
Because W for web... Get it? ... And it's the Polish word for server. Well next time I'll ask you for a catchy name, smartypants.
Is this ... thing ... even finished?
No. There are many functions missing and the code needs some revision (and comments I suppose). There is a working mode though (see example.c for an example).
If you want to help me with this project, feel free to fork and submit-pull requests. I also appreciate opinions, suggestions and complaints.
I don't get it. Is this a webserver for a lib for webservers?
Both, I guess. Although as I'm thinking about it, it might actually be more of a library...
Okay...? So, how can I use this thing?
See exmaple.c.
Basically you need at least one handler.
/**
* @param method contains the method-type (see method_t); can be stringified with const char* ws_strm(method_t)
* @param host the hostname (e.g. "localhost:8080" or "www.google.com")
* @param path the requested URI
* @param requestHeaders the headers of the request (see headers_t)
* @param responseHeaders well... guess. (Hint: You can modify these.)
* @param request "file" stream of the request (can be read with e.g. fscanf(2))
* @param respinse "file" stream for the response (can be written with w.g. fprintf(2))
*/
int not_found(method_t method, const char* host, const char* path, headers_t requestHeaders,
headers_t* responseHeaders, stream_t request, stream_t response) {
// The definition of this function is defined as handler_t (don't confuse with handle_t) in /webserver.h.
// You can write a response like this.
fprintf(response, "<!DOCTYPE html>\n");
fprintf(response, "<html><head><title>Not Found</title></head><body>");
fprintf(response, "<h1>404 - File Not Found</h1>");
fprintf(response, "<h1>The requested file (%s) was not found on this server.</h1>", path);
fprintf(response, "</body></html>");
return 404; // This is the HTTP status code by the way.
}
Then you can setup a server.
srvoptions_t options = {
.mode = LINEAR, // the mode to use
.timeout = 30, // the connection timeout (in this case useless, because of LINEAR)
.maxconnections = 5, // the maximum number of connections
.loglevel = LOG_DEBUG // the log level for the logfile
};
// webserver_t ws_create(const char* name, const char* host (can be NULL for default), const char* port, FILE* logfile, srvoptions_t options);
webserver_t server = ws_create("server_name", NULL, "8080", stderr, options);
Now you can add you handlers.
handle_t handle = {
.host = ANY, // the hostname to listen on (ANY can be used)
.path = "/*.c", // yes, you can use wildcards (ANY can still be used)
.handler = ¬_found // function pointer of the handler
};
ws_handle_add(&server, handle);
Done. Now you can start it.
ws_run(&server);
Simple as that. : )
Any questions?
... Oh, you were talking to me?
-.-
Is there something missing?
Well, where can I start?
- the other modes (LINEAR is working, but it's not useable in practice).
- SSL/HTTPS support
- management for cookies and sessions
- index-handlers
- cgi-handlers
Okay, anything else to do?
Yes:
- extracting the ws_*_run-functions plus their own dependencies to other files (I hate files with more than 1000 lines. D: )
- comments (Yes I know, I'm lazy on that.)
- some security features would be nice (timeouts, ...)