FastCGI binding or implementation?
Hello! Does anyone know if there's a D2 binding or implementation for fastcgi as of yet? Thanks! Jeremy Sandell
Re: FastCGI binding or implementation?
On Fri, Oct 14, 2011 at 09:14:19PM -0400, Jeremy Sandell wrote: >Does anyone know if there's a D2 binding or implementation for fastcgi as > of yet? My cgi.d wraps the C library for fast cgi https://github.com/adamdruppe/misc-stuff-including-D-programming-language-web-stuff compile with -version=fastcgi
Re: FastCGI binding or implementation?
Den 15-10-2011 03:14, Jeremy Sandell skrev: Hello! Does anyone know if there's a D2 binding or implementation for fastcgi as of yet? Thanks! Jeremy Sandell Adam Ruppe has a cgi module with some fastcgi in it afaik: http://arsdnet.net/dcode/cgi.d /Jonas
Re: FastCGI binding or implementation?
I'm working on a SCGI implementation for my job. SCGI is quite easy to handle and well-supported by servers. The only missing part is multipart POST (used for file post). get, post, cookie, headers (etc..) just work. I have to ask permission to publish my code, btw it's a very easy protocol to implement. Il giorno ven, 14/10/2011 alle 21.14 -0400, Jeremy Sandell ha scritto: > Hello! > > > >Does anyone know if there's a D2 binding or implementation for > fastcgi as of yet? > > > Thanks! > Jeremy Sandell
Re: FastCGI binding or implementation?
On 2011-10-17 10:14, Andrea Fontana wrote: I'm working on a SCGI implementation for my job. SCGI is quite easy to handle and well-supported by servers. The only missing part is multipart POST (used for file post). get, post, cookie, headers (etc..) just work. I have to ask permission to publish my code, btw it's a very easy protocol to implement. What about the rest of the HTTP methods? Il giorno ven, 14/10/2011 alle 21.14 -0400, Jeremy Sandell ha scritto: Hello! Does anyone know if there's a D2 binding or implementation for fastcgi as of yet? Thanks! Jeremy Sandell -- /Jacob Carlborg
Re: FastCGI binding or implementation?
I handle request on different threads. I do some pre-processing on scgi data and I fill a struct: request.get[] request.post[] request.cookie[] request.headers[string] then I call a virtual function (to override on subclasses) like: do(request, output); where user fill output struct in a way like: output.data ~= "hello world"; output.status = 200 output.cookies = bla bla and then if is method != "head" i send headers + data, else just "headers". btw 99% of usage is get, post, head. == Quotato dall`articolo Jacob Carlborg (d...@me.com) > On 2011-10-17 10:14, Andrea Fontana wrote: > > I'm working on a SCGI implementation for my job. SCGI is quite easy to > > handle and well-supported by servers. > > The only missing part is multipart POST (used for file post). > > > > get, post, cookie, headers (etc..) just work. > > > > I have to ask permission to publish my code, btw it's a very easy > > protocol to implement. > What about the rest of the HTTP methods? > > Il giorno ven, 14/10/2011 alle 21.14 -0400, Jeremy Sandell ha scritto: > >> Hello! > >> > >> > >> Does anyone know if there's a D2 binding or implementation for fastcgi > >> as of yet? > >> > >> > >> Thanks! > >> Jeremy Sandell
Re: FastCGI binding or implementation?
On 2011-10-17 16:01, Andrea Fontana wrote: I handle request on different threads. I do some pre-processing on scgi data and I fill a struct: request.get[] request.post[] request.cookie[] request.headers[string] then I call a virtual function (to override on subclasses) like: do(request, output); where user fill output struct in a way like: output.data ~= "hello world"; output.status = 200 output.cookies = bla bla and then if is method != "head" i send headers + data, else just "headers". btw 99% of usage is get, post, head. Yes, but if you want to write a web site that is RESTful you need the other HTTP methods as well, at least PUT and DELETE. BTW, what about creating something like Rack but for D. Rack is a low level interface in front of the web server which web frameworks can be built on top. http://rack.github.com/ -- /Jacob Carlborg
Re: FastCGI binding or implementation?
On Mon, Oct 17, 2011 at 2:11 PM, Jacob Carlborg wrote: > On 2011-10-17 16:01, Andrea Fontana wrote: > >> I handle request on different threads. I do some pre-processing on >> scgi data and I fill a struct: >> >> request.get[] >> request.post[] >> request.cookie[] >> request.headers[string] >> >> then I call a virtual function (to override on subclasses) like: >> >> do(request, output); >> >> where user fill output struct in a way like: >> >> output.data ~= "hello world"; >> output.status = 200 >> output.cookies = bla bla >> >> and then if is method != "head" i send headers + data, else just >> "headers". >> >> btw 99% of usage is get, post, head. >> > > Yes, but if you want to write a web site that is RESTful you need the other > HTTP methods as well, at least PUT and DELETE. > > BTW, what about creating something like Rack but for D. Rack is a low level > interface in front of the web server which web frameworks can be built on > top. > > http://rack.github.com/ > > -- > /Jacob Carlborg > Yes, this is exactly why I was wondering whether FastCGI had been implemented (though SCGI works for me as well) - so that I could write something on top of it, in much the same way I would using (for example) WSGI in Python. I also agree with you re: supporting all of the HTTP methods. Just because the most common ones are GET, POST, and HEAD doesn't mean we should leave out the others; both PUT and DELETE are quite useful. Best regards, Jeremy Sandell
Re: FastCGI binding or implementation?
I've been having trouble with my news postings, so forgive me if I said this before. But my cgi.d module supports FastCGI via the C library, with the same D interface as normal CGI or an embedded server: https://github.com/adamdruppe/misc-stuff-including-D-programming-language-web-stuff Just compile with -version=fastcgi to use that.
Re: FastCGI binding or implementation?
On 18.10.2011 19:24, Jeremy Sandell wrote:
On Mon, Oct 17, 2011 at 2:11 PM, Jacob Carlborg mailto:d...@me.com>> wrote:
On 2011-10-17 16:01, Andrea Fontana wrote:
I handle request on different threads. I do some pre-processing on
scgi data and I fill a struct:
request.get[]
request.post[]
request.cookie[]
request.headers[string]
then I call a virtual function (to override on subclasses) like:
do(request, output);
where user fill output struct in a way like:
output.data ~= "hello world";
output.status = 200
output.cookies = bla bla
and then if is method != "head" i send headers + data, else just
"headers".
btw 99% of usage is get, post, head.
Yes, but if you want to write a web site that is RESTful you need
the other HTTP methods as well, at least PUT and DELETE.
BTW, what about creating something like Rack but for D. Rack is a
low level interface in front of the web server which web frameworks
can be built on top.
http://rack.github.com/
--
/Jacob Carlborg
Yes, this is exactly why I was wondering whether FastCGI had been
implemented (though SCGI works for me as well) - so that I could write
something on top of it, in much the same way I would using (for example)
WSGI in Python.
I also agree with you re: supporting all of the HTTP methods. Just
because the most common ones are GET, POST, and HEAD doesn't mean we
should leave out the others; both PUT and DELETE are quite useful.
Best regards,
Jeremy Sandell
Adam D. Ruppe has a wrapper for libfcgi at github. I started
implementing fcgi, but it's basically just a very, very limited
ugly-hack prototype. Doubt you'll get much use of it, but I'll attach it
here anyway.
module fcgi;
import std.socket, std.stdio, std.socketstream;
import std.datetime;
import std.concurrency;
import core.thread;
import std.process;
import std.conv;
private import std.c.windows.windows, std.c.windows.winsock;
///
// PROTOCOL END
///
S toStruct(S, T)(T[] buf) {
static ubyte[S.sizeof] buf2;
assert(buf.length >= S.sizeof);
buf2 = buf[0..S.sizeof];
return cast(S)(buf2);
}
/// Who is responsible for closing the socket?
enum SocketLifetimeOwner {
application = 1,
server = 2
}
alias void delegate(Request) RequestCallback;
alias void delegate(ushort, int, ProtocolStatus) EndRequestHandler;
class Request {
RequestId _id;
Params _params;
Role_role;
SocketLifetimeOwner _socketOwner;
EndRequestHandler _endRequestHandler;
ubyte[] _input;
ListenThread _server;
bool _ended;
this(RequestId id, ListenThread server, EndRequestHandler
endRequestHandler) {
_id = id;
_endRequestHandler = endRequestHandler;
_server = server;
}
@property ubyte[] input() pure nothrow @safe {
return _input;
}
@property RequestId id() pure const nothrow @safe {
return _id;
}
@property void id(RequestId id) pure nothrow @safe {
_id = id;
}
@property const(Params) params() pure const nothrow @safe {
return _params;
}
private @property void params(Params params) pure nothrow @safe {
_params = params;
}
@property Role role() pure const nothrow @safe {
return _role;
}
private @property void role(Role role) pure nothrow @safe {
_role = role;
}
@property SocketLifetimeOwner socketOwner() pure const nothrow @safe {
return _socketOwner;
}
@property void socketOwner(SocketLifetimeOwner owner) pure nothrow @safe {
_socketOwner = owner;
}
void write(T)(const(T[]) data, RecordType method = RecordType.stdOut) {
assert(!_ended);
// Break into chunks of 65535
auto socket = _server._active;
auto head = RecordHead(method, _id,
cast(ushort)(data.length));
socket.send(cast(ubyte[RecordHead.sizeof])head);
socket.send(cast(ubyte[])data);
head.contentLength = 0;
socket.send(cast(ubyte[RecordHead.sizeof])head);
}
void end() {
_endRequestHandler(_id, 0, ProtocolStatus.requestComplete);
_ended = true;
}
}
class Params {
private string[string] _params;
private this(string[string] params) {
_params = params;
}
private this(in ubyte[] raw) pure nothrow @trusted {
int i;
do {
auto keyLen = cast(ubyte) raw[i++];
auto valLen = cast(ubyte) raw[i++];
auto key= cast(string)raw[i .. i+keyLen];
i += keyLen;
auto value = cast(string)raw[i .. i+valLen];
i += valLen;
_params[key] = value;
} while(i < raw.length);
}
string opDispatch(string key)() const pure nothrow @safe {
auto v = key in _params;
Re: FastCGI binding or implementation?
Tale time, only tangentially on topic. Today, I was coincidentally switching one of my work apps from standard CGI to Fast CGI. Almost trivial. Set up Apache, then build the program with -version=fastcgi. Done. Well, not 100% done. I had a piece of static data in the app that worked correctly before but now means subsequent requests were out of it. Changed that to an instance variable, and boom, works perfectly. * If you are using Fast CGI, avoid static variables. The next thing is speed. From principles, there's very little reason for FastCGI to actually be faster than normal CGI - the startup costs are insignificant next to the total app runtime and network lag. (Startup is maybe 5 ms on the live server, with runtime close to 50ms and ping to the user another 100ms. The "cost" of CGI is roundoff error in the actual deployment.) My benchmarks supported this for the kind of loads we had before. But now, the number of concurrent users is picking up. The CGI still performed very well, though every so often, users complained about lag on some resources. I ran a benchmark comparing cgi to fast cgi with a very large number of concurrent users. It showed better availability and about a 15% speed boost under this load. Since Apache restarts it when it segfaults, reliability ought not to be affected, though it's too soon to say for sure. So, I changed the makefile to say "-version=fastcgi" and soon realized I must search for static variables - found just one, so easy fix, and we're up on fastcgi. ... but that 15% in the benchmark hasn't translated to a big change in the live environment yet. Been several hours now, and we've been trying to force the availability issue, and failed so far. Looks like a win, but not a very big one. Speed on the whole - unaffected. The difference is roundoff error once you factor in network lag and such again. So, how can we speed up the application? The key here is client side caching. Using my cgi.d, there's a function: cgi.setCache(true); which tells it simply to cache the response forever. It makes an expiration date long in the future. Set that for any content which changes infrequently - css, javascript, images, any kind of (conceptually) pure or static data, etc. Now, your code doesn't run again and the user doesn't hit the network again. What was 150ms is now < 1ms. The users will feel the difference. You might set even data that changes often to cache for a few minutes. Odds are the user doesn't need it to revalidate on the server every minute. cgi.d's setResponseExpires can help here, just set it a little bit in the future. If the user hits a link to go back to a page then, it will load from cache most the time, making navigating the site feel snappy. Until the time expires, then it's wait again, but IMO some cache is better than none. Remember: you can cache AJAX responses too. What if the resource actually does change? You'll want to change the link. When compiling, there's a __TIMESTAMP__ special token. A quick and dirty method is to use that __TIMESTAMP__ on your resource URLs in your html so every time you recompile, it invalidates the user's cache. A better way might be to hash the content at compile time, but I haven't written code that can do this well enough for real work yet. * Caching makes a much bigger difference than just about any other technique. You'll still want fast code for the cold cache users, but as they browse your site, a good cache policy can shave full seconds off the experience. The fastest code is running no code at all.
Re: FastCGI binding or implementation?
On 2011-10-18 19:24, Jeremy Sandell wrote: On Mon, Oct 17, 2011 at 2:11 PM, Jacob Carlborg mailto:d...@me.com>> wrote: On 2011-10-17 16:01, Andrea Fontana wrote: I handle request on different threads. I do some pre-processing on scgi data and I fill a struct: request.get[] request.post[] request.cookie[] request.headers[string] then I call a virtual function (to override on subclasses) like: do(request, output); where user fill output struct in a way like: output.data ~= "hello world"; output.status = 200 output.cookies = bla bla and then if is method != "head" i send headers + data, else just "headers". btw 99% of usage is get, post, head. Yes, but if you want to write a web site that is RESTful you need the other HTTP methods as well, at least PUT and DELETE. BTW, what about creating something like Rack but for D. Rack is a low level interface in front of the web server which web frameworks can be built on top. http://rack.github.com/ -- /Jacob Carlborg Yes, this is exactly why I was wondering whether FastCGI had been implemented (though SCGI works for me as well) - so that I could write something on top of it, in much the same way I would using (for example) WSGI in Python. I also agree with you re: supporting all of the HTTP methods. Just because the most common ones are GET, POST, and HEAD doesn't mean we should leave out the others; both PUT and DELETE are quite useful. Best regards, Jeremy Sandell Although I have no idea if the rest of the 9 HTTP methods are useful, e.g. trace, options, connect and patch. -- /Jacob Carlborg
Re: FastCGI binding or implementation?
On 2011-10-18 20:21, Adam D. Ruppe wrote: Tale time, only tangentially on topic. Today, I was coincidentally switching one of my work apps from standard CGI to Fast CGI. Almost trivial. Set up Apache, then build the program with -version=fastcgi. Done. Well, not 100% done. I had a piece of static data in the app that worked correctly before but now means subsequent requests were out of it. Changed that to an instance variable, and boom, works perfectly. * If you are using Fast CGI, avoid static variables. The next thing is speed. From principles, there's very little reason for FastCGI to actually be faster than normal CGI - the startup costs are insignificant next to the total app runtime and network lag. (Startup is maybe 5 ms on the live server, with runtime close to 50ms and ping to the user another 100ms. The "cost" of CGI is roundoff error in the actual deployment.) My benchmarks supported this for the kind of loads we had before. But now, the number of concurrent users is picking up. The CGI still performed very well, though every so often, users complained about lag on some resources. I ran a benchmark comparing cgi to fast cgi with a very large number of concurrent users. It showed better availability and about a 15% speed boost under this load. Since Apache restarts it when it segfaults, reliability ought not to be affected, though it's too soon to say for sure. So, I changed the makefile to say "-version=fastcgi" and soon realized I must search for static variables - found just one, so easy fix, and we're up on fastcgi. ... but that 15% in the benchmark hasn't translated to a big change in the live environment yet. Been several hours now, and we've been trying to force the availability issue, and failed so far. Looks like a win, but not a very big one. Speed on the whole - unaffected. The difference is roundoff error once you factor in network lag and such again. So, how can we speed up the application? The key here is client side caching. Using my cgi.d, there's a function: cgi.setCache(true); which tells it simply to cache the response forever. It makes an expiration date long in the future. Set that for any content which changes infrequently - css, javascript, images, any kind of (conceptually) pure or static data, etc. Now, your code doesn't run again and the user doesn't hit the network again. What was 150ms is now< 1ms. The users will feel the difference. You might set even data that changes often to cache for a few minutes. Odds are the user doesn't need it to revalidate on the server every minute. cgi.d's setResponseExpires can help here, just set it a little bit in the future. If the user hits a link to go back to a page then, it will load from cache most the time, making navigating the site feel snappy. Until the time expires, then it's wait again, but IMO some cache is better than none. Remember: you can cache AJAX responses too. What if the resource actually does change? You'll want to change the link. When compiling, there's a __TIMESTAMP__ special token. A quick and dirty method is to use that __TIMESTAMP__ on your resource URLs in your html so every time you recompile, it invalidates the user's cache. A better way might be to hash the content at compile time, but I haven't written code that can do this well enough for real work yet. Yeah, that's how Rails 3.1 does it now. Rails 3.1 inserts a hash of the content in the file name instead of a time stamp after the question mark. * Caching makes a much bigger difference than just about any other technique. You'll still want fast code for the cold cache users, but as they browse your site, a good cache policy can shave full seconds off the experience. The fastest code is running no code at all. Why not just cache the generated HTML and let Apache handle it. Then it doesn't even need to start the application if the cache is available. -- /Jacob Carlborg
Re: FastCGI binding or implementation?
AFAIK other http methods have nothing special. You have just to
implement on your code:
if (request.method == "PUT")
{
...
...
}
if you need them. Am i wrong?
Il giorno mer, 19/10/2011 alle 08.36 +0200, Jacob Carlborg ha scritto:
> On 2011-10-18 19:24, Jeremy Sandell wrote:
> > On Mon, Oct 17, 2011 at 2:11 PM, Jacob Carlborg > > wrote:
> >
> > On 2011-10-17 16:01, Andrea Fontana wrote:
> >
> > I handle request on different threads. I do some pre-processing on
> > scgi data and I fill a struct:
> >
> > request.get[]
> > request.post[]
> > request.cookie[]
> > request.headers[string]
> >
> > then I call a virtual function (to override on subclasses) like:
> >
> > do(request, output);
> >
> > where user fill output struct in a way like:
> >
> > output.data ~= "hello world";
> > output.status = 200
> > output.cookies = bla bla
> >
> > and then if is method != "head" i send headers + data, else just
> > "headers".
> >
> > btw 99% of usage is get, post, head.
> >
> >
> > Yes, but if you want to write a web site that is RESTful you need
> > the other HTTP methods as well, at least PUT and DELETE.
> >
> > BTW, what about creating something like Rack but for D. Rack is a
> > low level interface in front of the web server which web frameworks
> > can be built on top.
> >
> > http://rack.github.com/
> >
> > --
> > /Jacob Carlborg
> >
> >
> > Yes, this is exactly why I was wondering whether FastCGI had been
> > implemented (though SCGI works for me as well) - so that I could write
> > something on top of it, in much the same way I would using (for example)
> > WSGI in Python.
> >
> > I also agree with you re: supporting all of the HTTP methods. Just
> > because the most common ones are GET, POST, and HEAD doesn't mean we
> > should leave out the others; both PUT and DELETE are quite useful.
> >
> > Best regards,
> > Jeremy Sandell
>
> Although I have no idea if the rest of the 9 HTTP methods are useful,
> e.g. trace, options, connect and patch.
>
Re: FastCGI binding or implementation?
Andrea Fontana: > other http methods have nothing special Indeed. The only thing that might get you is if the data's content type is different than the default. That's possible on POST too, though, so still nothing special. My cgi library has an enum to tell you what the requestMethod is, and it lists all the options in the standard. It, however, does not handle all possible content-types. It does x-www-form-urlencoded and multipart/form-data, so it can handle virtually all web forms out there - including file uploads - but if you want others, it'll take a minor modification. The best way to do it is probably to not attempt to parse it in the library at all, and just pass a range of raw data to the application.
Re: FastCGI binding or implementation?
Jacob Carlborg: > Why not just cache the generated HTML and let Apache handle it. That sounds hard... configuring Apache to do anything beyond the most trivial of tasks is a huge pain to me. It is easy to call "cgi.setCache(true);" though. > Then it doesn't even need to start the application if the cache is > available. It's better with a client side cache. If it's available, you don't have to go to the server at all! Server side caching is something I haven't done yet; it's never been useful to me.
Re: FastCGI binding or implementation?
On 2011-10-19 16:36, Adam Ruppe wrote:
Jacob Carlborg:
Why not just cache the generated HTML and let Apache handle it.
That sounds hard... configuring Apache to do anything beyond the most
trivial of tasks is a huge pain to me.
No, it's not that hard. Just add a couple of rewrite-rules that checks
if a request matches an already cached page and rewrite it to the cached
page. Something like this:
...
RailsAllowModRewrite On
RewriteEngine On
RewriteCond %{THE_REQUEST} ^(GET|HEAD)
RewriteCond %{REQUEST_URI} ^/([^.]+)$
RewriteCond %{DOCUMENT_ROOT}/cache/%1.html -f
RewriteRule ^/[^.]+$ /cache/%1.html [QSA,L]
RewriteCond %{THE_REQUEST} ^(GET|HEAD)
RewriteCond %{DOCUMENT_ROOT}/cache/index.html -f
RewriteRule ^/$ /cache/index.html [QSA,L]
I found the above at:
http://www.alfajango.com/blog/make-sure-your-rails-application-is-actually-caching-and-not-just-pretending/
It is easy to call "cgi.setCache(true);" though.
Then it doesn't even need to start the application if the cache is
available.
It's better with a client side cache. If it's available, you don't
have to go to the server at all!
Yes of course, that is preferred.
Server side caching is something I haven't done yet; it's never
been useful to me.
--
/Jacob Carlborg
