Extending Webserver.r

I should clarify things, I don't want to deal with CGI issues. All I've been doing at work for the last 3 months is writing CGI scripts in Rebol. It's gotten to the point where some of the things I'd like to do would be best accomplished with a small custom HTTP server. I think I solved my problem. I went digging around in the system/ schemes/http and looked at how it read in the http headers: while [(line: pick port/sub-port 1) <> ""] [append headers join line "^/"] I put in a line like that, and then extracted the content-length. I used the content-length in a read-io loop to get the form data. I have to say that this is most assuredly non-intuitive. I would never have thought to use pick like that. I'd kill for a really good tutorial on the ins and outs of ports and networking in Rebol.

I do, that's why I'm biting the bullet and writing a web server. ;) :Eric

Eric, Writing the web server might be good experience. These notes might help (or not). ================================================================ I had a little bit of experience with this about a year ago after Carl released the webserver.r I wanted a rebol instance that would handle rebol scripts without having to launch a new instance of rebol ala CGI approach. This would give me faster response time in the mode that I was working in then which was the demonstration of some web prototypes. I wanted to be able to run a webserver, the web applications, and all of the html pages and graphics for a reasonable sized application prototype from a diskette. Hence REBOL. I made a few changes to the webserver.r code which I think they are all in the while block with comments to indicate the goal. For me, these changes made sense 'cause I was just trying to get the REBOL listener on port 80 to recognize that the URL pointed to a REBOL script - then knowing that, run it. So it this was fairly intuitive. ==================================================================== while [on] [ http-port: first wait listen-port request: first http-port file: "index.html" rebol: false ; turn this boolean off mime: "text/plain" parse request ["get" ["http" | "/ " | copy file to " "]] parse file [thru "." [ "htm" (mime: "text/html") | "html" (mime: "text/html") | "gif" (mime: "image/gif") | "jpg" (mime: "image/jpeg") | "r" (rebol: true ; being asked to process a REBOL script mime: "text/html") ; assuming that want to emit html from REBOL script ] ] if not none? (parts: parse file "?") ; Look for GET arguments [file: pick parts 1 ; plug the GET arguments back into the query-string ; so that the REBOL script will be able to get them system/options/cgi/query-string: pick parts 2] print ["retrieving:" file] any [ if not exists? web-dir/:file [send-error 404 file] either rebol ; if the url points in the get points to a rebol script, then just do it. [data: copy "" print: func [some_text] [append data some_text] if error? try [do web-dir/:file] [send-error 400 file] ] [if error? try [data: read/binary web-dir/:file] [send-error 400 file]] send-page data mime ] close http-port ] ============================================================================ ========= This did what I wanted and worked about an hour after I started it. The howevers : a. may not be what you want b. only handles low usage Cal Dixon ([deadzaphod--hotmail--com]) released a different version on the REBOL List a few days later. He updated it in January and February 2000. Cal's would: a. handle PUTs as well as GETs b. handle concurrent access but I did not run it since my few lines above met my needs. I just checked the REBOL script site and could not find it. The source that Cal produced was: ============================================================================ ============ REBOL [ Title: "REBOL Web Server" File: %webserv.r Author: "Cal Dixon" Email: [deadzaphod--hotmail--com] Date: 26-Feb-2000 Purpose: { A Simple HTTP-Server that can run REBOL CGI scripts } Notes: { 0.0.0.3: This version redirects all i/o to the web browser so 'read-io on system/ports/input can be used to get POSTed data, etc.. 0.0.0.4: Now has better error checking and passes content-length as a string like it should 0.0.0.5: Can now send multiple files at once 0.0.0.6: Now patches 'print and 'prin to work correctly and passes all http headers to CGIs also translates access to a folder to %index.html in that folder. Also handles the HTTP HEAD method in addition to GET and sends the Last-Modified header 0.0.0.7: Added logging in Extended Common Log Format - but for CGI scripts the number of bytes sent is recorded as 1, due to current limitations of this program } Version: 0.0.0.7 Category: 'web ] wwwpath: %www/ ; change this to where the files are... port: 80 ; change this to whatever port the server should listen to logfile: %webserv.log ; the name of the logfile or set to none secure none system/options/quiet: false e: {<HTML><HEAD><TITLE>404 Not Found</TITLE></HEAD><BODY>Page not found.</BODY></HTML>} cgi-obj: make system/options/cgi [ context: func [] [ return 'context ] ] listen: open/lines/direct join tcp://: port inport: system/ports/input outport: system/ports/output queue: [] ; these replacements for 'print and 'prin should work better for CGI scripts prin: func [ out /local data ] [ data: replace/all (reform out) newline "^M^J" write-io system/ports/output data length? data return ] print: func [ out /local data ] [ data: replace/all (reform out) newline "^M^J" data: append data "^M^J" write-io system/ports/output data length? data return ] quit: halt: func [] [throw] www-send: func [ conn data ] [ write-io conn data length? data ] either logfile [ write-log: func [ entry ] [ write/append logfile join to-string entry newline ] ][ write-log: func [ ignorethisvalue ] [] ] get-http-headers: func [ conn /local line buffer a b c ] [ buffer: copy [] while [ ((line: first conn) <> "") and not none? line ] [ a: copy/part line b: find line ":" c: trim next b insert buffer reduce [ a c ] ] return buffer ] handle-cgi: func [ conn request query headers /local cd ] [ system/options/cgi: make cgi-obj compose [ server-software: "REBOL Web Server" server-name: (read dns://) gateway-interface: "CGI/1.1" server-protocol: "HTTP/1.0" server-port: "80" query-string: (query) request-method: (pick request 1) script-name: (first parse (pick request 2) "?") Content-Type: (select headers "Content-Type") Content-Length: (select headers "Content-Length") other-headers: (headers) ] cd: what-dir system/ports/output: conn system/ports/input: conn www-send conn "HTTP/1.0 200 OK^/" if error? try [ catch [ do file-path ] ] [] system/ports/input: inport system/ports/output: outport change-dir cd close conn ] content-type?: func [ filename [string! file!] ] [ switch/default next find/last to-string filename "." [ "txt" [ return "text/plain" ] "gif" [ return "image/gif" ] "jpg" [ return "image/jpeg" ] "png" [ return "image/png" ] "mov" [ return "video/quicktime" ] "tif" [ return "image/tiff" ] "tiff" [ return "image/tiff" ] "wav" [ return "audio/wav" ] "xml" [ return "text/xml" ] "xsl" [ return "text/xml" ] "mid" [ return "audio/midi" ] "r" [ return none ] ] [ return "text/html" ] ] process-queue: func [ /local connection data file conn newqueue ] [ newqueue: copy [] foreach connection queue [ set [ conn file ] connection data: copy/part file 2048 file: skip file 2048 write-io conn data length? data either tail? file [ close conn ] [ insert/only newqueue reduce [ conn file ] ] ] queue: newqueue ] send-header: func [ conn result content-type data-length ] [ www-send conn rejoin [ "HTTP/1.0 " result newline "Content-Type: " content-type newline "Content-Length: " data-length newline "Date: " to-idate now newline "Last-Modified: " to-idate modified? file-path "^/^/" ] ] translate-request-to-resource: func [ file /local file-path ] [ if (last file) = #"/" [ append file "index.html" ] file-path: clean-path join wwwpath to-file next file if none? find file-path clean-path wwwpath [ file-path: clean-path join wwwpath "index.html" ] if dir? file-path [ append file-path "/index.html" ] return file-path ] http-log: func [ host request status bytes /extended headers /local when agent referer] [ when: rejoin [ replace/all copy/part mold now 11 "-" "/" replace skip mold now 11 "/" ":" ] replace when "-" " -" either (agent: select headers "User-Agent") [ agent: join {"} [ agent {"} ] ][ agent: "-" ] either (referer: select headers "Referer") [ referer: join {"} [ referer {"} ] ][ referer: "-" ] reform [ host "- -" rejoin [ "[" when "]" ] mold form request status bytes either extended [ reform [ referer agent ] ][ "" ] ] ] handle-new-connections: func [ /local data conn http-headers ] [ if none? wait reduce [ listen 0 ] [ return ] request: parse first (conn: first listen) none if (length? queue) > 3000 [ insert conn "HTTP/1.0 503 Server Overloaded^/" close conn return ] ; refuse connections if server is overloaded request-method: pick request 1 set [ file urlquery ] parse (pick request 2) "?" file-path: translate-request-to-resource file http-headers: get-http-headers conn either exists? file-path [ either none? content: content-type? file-path [ write-log http-log/extended conn/host request 200 1 http-headers handle-cgi conn request urlquery http-headers return ] [ write-log http-log/extended conn/host request 200 size? file-path http-headers set [ responce data ] reduce [ "200 OK" (data: read/binary file-path) ] ] ] [ write-log http-log/extended conn/host request 404 0 http-headers set [ responce content data file-path ] reduce [ "404 Not Found" text/html e %. ] ] send-header conn responce content length? data if request-method = "HEAD" [ close conn return ] insert/only queue reduce [ conn data ] ] forever [ if ( zero? ( length? queue ) ) [ wait listen ] handle-new-connections process-queue ]

What's "intuitive" in C, might seem "non-intuitive" in rebol. And vice versa. Even tho' C remains my "first" language for the time being, rebol will probably eventually take C's place and part of the reason is rebol's sheer productivity and it's closer resemblance to human language. I' m reading Elan's book. and as I progress through it, all rebol syntax is becoming more intuitive.

Amen to that! Am writing a server using my current legacy of C/C++ resources. I will then probably use rebol scripts as an interface. I would think that the more I learn of ports and networking, and the more I learn of rebol, then the use of ports and networking in rebol will become clearer.

regards Tim

Thanks, lots of good code in there. My next set of questions was going to revolve around handling multiple connections. :Eric