1PSGI::FAQ(3) User Contributed Perl Documentation PSGI::FAQ(3)
2
6 PSGI::FAQ - Frequently Asked Questions and answers
7
9 General
10 How do you pronounce PSGI?
11 We read it simply P-S-G-I.
13 So what is this?
15 PSGI is an interface between web servers and perl-based web
17 applications akin to what CGI does for web servers and CGI scripts.
18 Why do we need this?
20 Perl has CGI as a core module that somewhat abstracts the difference
22 between CGI, mod_perl and FastCGI. However, most web application
23 framework developers (e.g. Catalyst and Jifty) usually avoid using it
24 to maximize the performance and to access low-level APIs. So they end
25 up writing adapters for all of those different environments, some of
26 which may be well tested while others are not.
27 PSGI allows web application framework developers to only write an
29 adapter for PSGI. End users can choose from among all the backends
30 that support the PSGI interface.
31 You said PSGI is similar to CGI. How is the PSGI interface different
33 from CGI?
34 The PSGI interface is intentionally designed to be very similar to CGI
36 so that supporting PSGI in addition to CGI would be extremely easy.
37 Here's a highlight of the key differences between CGI and PSGI:
38 • In CGI, servers are the actual web servers written in any languages
40 but mostly in C, and script is a script that can be written in any
41 language such as C, Perl, Shell scripts, Ruby or Python.
42 In PSGI, servers are still web servers, but they're perl processes
44 that are usually embedded in the web server (like mod_perl) or a
45 perl daemon process called by a web server (like FastCGI), or an
46 entirely perl based web server. And PSGI application is a perl code
47 reference.
48 • In CGI, we use STDIN, STDERR, and environment variables to read
50 parameters and the HTTP request body and to send errors from the
51 application.
52 In PSGI, we use the $env hash references and the psgi.input and
54 psgi.errors streams to pass that data between servers and
55 applications.
56 • In CGI, applications are supposed to print HTTP headers and body to
58 STDOUT to pass it back to the web server.
59 In PSGI, applications are supposed to return a HTTP status code,
61 headers, and body (as an array ref or a filehandle-like object) to
62 the application as an array reference.
63 My framework already does CGI, FCGI and mod_perl. Why do I want to
65 support PSGI?
66 There are many benefits for the web application framework to support
68 PSGI.
69 • You can stop writing code to support many web server environments.
71 Plack has a lot of well-tested server adapters to environments such
73 as CGI, FastCGI and mod_perl. There are also many new web servers
74 built to support the PSGI standard interface, such as Starman,
75 Starlet and Twiggy. Once your framework supports PSGI, there's
76 nothing you need to do to run your application on these new web
77 servers. You can get that for free.
78 Also, even if your framework already supports most server
80 environments like discussed above, you can now drop these code in
81 favor of only supporting PSGI. This is what Jifty and Catalyst have
82 done, when they implemented the PSGI support. Less code means less
83 bugs :)
84 • Your framework can now use all of Plack middleware components.
86 Just search for "Plack::Middleware" on CPAN and you'll see hundreds
88 of PSGI compatible middleware components. They're often newly
89 created, but also extracted from plugins for certain web frameworks
90 such as Catalyst. By supporting PSGI interface, your framework can
91 make use of all of these useful middleware, such as session
92 management, content caching, URL rewriting and debug panel to name
93 just a few.
94 • You can test the application using the consistent Plack::Test
96 interface.
97 Any PSGI application can be tested using Plack::Test, either
99 through a mock request or a live server implementation. There's
100 also Test::WWW::Mechanize::PSGI to allow Mechanize-style testing.
101 I'm writing a web application. What's the benefit of PSGI for me?
103 If the framework you're using supports PSGI, that means your
105 application can run on any of existing and future PSGI implementations.
106 You can provide a ".psgi" file that returns PSGI application, the end
107 users of your application should be able to configure and run your
108 application in a bunch of different ways.
109 But I'm writing a web application in CGI and it works well. Should I
111 switch to PSGI?
112 If you're writing a web application with a plain CGI.pm and without
114 using any web frameworks, you're limiting your application in the plain
115 CGI environments, along with mod_perl and FastCGI with some tweaks. If
116 you're the only one developer and user of your application then that's
117 probably fine.
118 One day you want to deploy your application in a shared hosting
120 environment for your clients, or run your server in the standalone mode
121 rather than as a CGI script, or distribute your application as open
122 source software. Limiting your application in the CGI environment by
123 using CGI.pm will bite you then.
124 You can start using one of PSGI compatible frameworks (either full-
126 stack ones or micro ones), or use Plack::Request if you are anti
127 frameworks, to make your application PSGI aware, to be more future
128 proof.
129 Even if you ignore PSGI today and write applications in plain CGI, you
131 can always later switch to PSGI with the CGI::PSGI wrapper.
132 What should I do to support PSGI?
134 If you're a web server developer, write a PSGI implementation that
136 calls a PSGI application. Also join the development on Plack, the PSGI
137 toolkit and utilities, to add a server adapter for your web server.
138 If you're a web application framework developer, write an adapter for
140 PSGI. Now you're freed from supporting all different server
141 environments.
142 If you're a web application developer (or a web application framework
144 user), choose the framework that supports PSGI, or ask the author to
145 support it. :) If your application is a large scale installable
146 application that doesn't use any existing frameworks (e.g. WebGUI or
147 Movable Type) you're considered as a framework developer instead from
148 the PSGI point of view. So, writing an adapter for PSGI on your
149 application would make more sense.
150 Is PSGI faster than (my framework)?
152 Again, PSGI is not an implementation, but there's a potential for a
154 very fast PSGI implementation that preloads everything and runs fully
155 optimized code as a preforked standalone with XS parsers, an event-
156 based tiny web server written in C and embedded perl that supports
157 PSGI, or a plain-old CGI.pm based backend that doesn't load any modules
158 at all and runs pretty quickly without eating so much memory under the
159 CGI environment.
160 There are prefork web server implementations such as Starman and
162 Starlet, as well as fully asynchronous event based implementations such
163 as Twiggy, Corona or Feersum. They're pretty fast and they include
164 adapters for Plack so you can run with the plackup utility.
165 Users of your framework can choose which backend is the best for their
167 needs. You, as a web application framework developer, don't need to
168 think about lots of different users with different needs.
169 Plack
171 What is Plack? What is the difference between PSGI and Plack?
172 PSGI is a specification, so there's no software or module called PSGI.
174 End users will need to choose one of the PSGI server implementations to
175 run PSGI applications on. Plack is a set of PSGI utilities and contains
176 the reference PSGI server HTTP::Server::PSGI, as well as Web server
177 adapters for CGI, FastCGI and mod_perl.
178 Plack also has useful APIs and helpers on top of PSGI, such as
180 Plack::Request to provide a nice object-oriented API on request
181 objects, plackup that allows you to run an PSGI application from the
182 command line and configure it using "app.psgi" (a la Rack's Rackup),
183 and Plack::Test that allows you to test your application using standard
184 HTTP::Request and HTTP::Response pair through mocked HTTP or live HTTP
185 servers. See Plack for details.
186 What kind of server backends would be available?
188 In Plack, we already support most web servers like Apache2, and also
190 the ones that supports standard CGI or FastCGI, but also try to support
191 special web servers that can embed perl, like Perlbal or nginx. We
192 think it would be really nice if Apache module mod_perlite and Google
193 AppEngine supported PSGI too, so that you could run your PSGI/Plack
194 based perl app in the cloud.
195 Ruby is Rack and JavaScript is Jack. Why is it not called Pack?
197 Well Pack indeed is a cute name, but Perl has a built-in function pack
199 so it's a little confusing, especially when speaking instead of
200 writing.
201 What namespaces should I use to implement PSGI support?
203 Do not use the PSGI:: namespace to implement PSGI backends or adapters.
205 The PSGI namespace is reserved for PSGI specifications and reference
207 unit tests that implementors have to pass. It should not be used by
208 particular implementations.
209 If you write a plugin or an extension to support PSGI for an
211 (imaginary) web application framework called "Camper", name the code
212 such as "Camper::Engine::PSGI".
213 If you write a web server that supports PSGI interface, then name it
215 however you want. You can optionally support Plack::Handler's abstract
216 interface or write an adapter for it, which is:
217 my $server = Plack::Handler::FooBar->new(%opt);
219 $server->run($app);
220 By supporting this "new" and "run" in your server, it becomes plackup
222 compatible, so users can run your app via "plackup". You're recommended
223 to, but not required to follow this API, in which case you have to
224 provide your own PSGI app launcher.
225 I have a CGI or mod_perl application that I want to run on PSGI/Plack.
227 What should I do?
228 You have several choices:
230 CGI::PSGI
232 If you have a web application (or framework) that uses CGI.pm to
233 handle query parameters, CGI::PSGI can help you migrate to PSGI.
234 You'll need to change how you create CGI objects and how to return
235 the response headers and body, but the rest of your code will work
236 unchanged.
237 CGI::Emulate::PSGI and CGI::Compile
239 If you have a dead old CGI script that you want to change as little
240 as possible (or even no change at all), then CGI::Emulate::PSGI and
241 CGI::Compile can compile and wrap them up as a PSGI application.
242 Compared to CGI::PSGI, this might be less efficient because of
244 STDIN/STDOUT capturing and environment variable mangling, but
245 should work with any CGI implementation, not just CGI.pm, and
246 CGI::Compile does the job of compiling a CGI script into a code
247 reference just like mod_perl's Registry does.
248 Plack::Request and Plack::Response
250 If you have an HTTP::Engine based application (framework), or want
251 to write an app from scratch and need a better interface than CGI,
252 or you're used to Apache::Request, then Plack::Request and
253 Plack::Response might be what you want. It gives you a nice
254 Request/Response object API on top of the PSGI env hash and
255 response array.
256 NOTE: Don't forget that whenever you have a CGI script that runs once
258 and exits, and you turn it into a persistent process, it may have
259 cleanup that needs to happen after every request -- variables that need
260 to be reset, files that need to be closed or deleted, etc. PSGI can do
261 nothing about that (you have to fix it) except give you this friendly
262 reminder.
263 HTTP::Engine
265 Why PSGI/Plack instead of HTTP::Engine?
266 HTTP::Engine was a great experiment, but it mixed the application
268 interface (the "request_handler" interface) with implementations, and
269 the monolithic class hierarchy and role based interfaces make it really
270 hard to write a new backend. We kept the existing HTTP::Engine and
271 broke it into three parts: The interface specification (PSGI),
272 Reference server implementations (Plack::Handler) and Standard APIs and
273 Tools (Plack).
274 Will HTTP::Engine be dead?
276 It won't be dead. HTTP::Engine will stay as it is and still be useful
278 if you want to write a micro webserver application rather than a
279 framework.
280 Do I have to rewrite my HTTP::Engine application to follow PSGI
282 interface?
283 No, you don't need to rewrite your existing HTTP::Engine application.
285 It can be easily turned into a PSGI application using
286 HTTP::Engine::Interface::PSGI.
287 Alternatively, you can use Plack::Request and Plack::Response which
289 gives compatible APIs to HTTP::Engine::Request and
290 HTTP::Engine::Response:
291 use Plack::Request;
293 use Plack::Response;
294 sub request_handler {
296 my $req = Plack::Request->new(shift);
297 my $res = Plack::Response->new;
298 # ...
299 return $res->finalize;
300 }
301 And this "request_handler" is a PSGI application now.
303 API Design
305 Keep in mind that most design choices made in the PSGI spec are to
306 minimize the requirements on backends so they can optimize things.
307 Adding a fancy interface or allowing flexibility in the PSGI layers
308 might sound catchy to end users, but it would just add things that
309 backends have to support, which would end up getting in the way of
310 optimizations, or introducing more bugs. What makes a fancy API to
311 attract web application developers is your framework, not PSGI.
312 Why a big env hash instead of objects with APIs?
314 The simplicity of the interface is the key that made WSGI and Rack
316 successful. PSGI is a low-level interface between backends and web
317 application framework developers. If we define an API on what type of
318 objects should be passed and which method they need to implement, there
319 will be so much duplicated code in the backends, some of which may be
320 buggy.
321 For instance, PSGI defines "$env->{SERVER_NAME}" as a string. What if
323 the PSGI spec required it to be an instance of Net::IP? Backend code
324 would have to depend on the Net::IP module, or have to write a mock
325 object that implements ALL of Net::IP's methods. Backends depending on
326 specific modules or having to reinvent lots of stuff is considered
327 harmful and that's why the interface is as minimal as possible.
328 Making a nice API for the end users is a job that web application
330 frameworks (adapter developers) should do, not something PSGI needs to
331 define.
332 Why is the application a code ref rather than an object with a ->call
334 method?
335 Requiring an object in addition to a code ref would make EVERY
337 backend's code a few lines more tedious, while requiring an object
338 instead of a code ref would make application developers write another
339 class and instanciate an object.
340 In other words, yes an object with a "call" method could work, but
342 again PSGI was designed to be as simple as possible, and making a code
343 reference out of class/object is no brainer but the other way round
344 always requires a few lines of code and possibly a new file.
345 Why are the headers returned as an array ref and not a hash ref?
347 Short: In order to support multiple headers (e.g. "Set-Cookie").
349 Long: In Python WSGI, the response header is a list of ("header_name",
351 "header_value") tuples i.e. "type(response_headers) is ListType" so
352 there can be multiple entries for the same header key. In Rack and
353 JSGI, a header value is a String consisting of lines separated by
354 ""\n"".
355 We liked Python's specification here, and since Perl hashes don't allow
357 multiple entries with the same key (unless it's "tie"d), using an array
358 reference to store "[ key => value, key => value ]" is the simplest
359 solution to keep both framework adapters and backends simple. Other
360 options, like allowing an array ref in addition to a plain scalar, make
361 either side of the code unnecessarily tedious.
362 I want to send Unicode content in the HTTP response. How can I do so?
364 PSGI mocks wire protocols like CGI, and the interface doesn't care too
366 much about the character encodings and string semantics. That means,
367 all the data on PSGI environment values, content body etc. are sent as
368 byte strings, and it is an application's responsibility to properly
369 decode or encode characters such that it's being sent over HTTP.
370 If you have a decoded string in your application and want to send them
372 in "UTF-8" as an HTTP body, you should use Encode module to encode it
373 to utf-8. Note that if you use one of PSGI-supporting frameworks,
374 chances are that they allow you to set Unicode text in the response
375 body and they do the encoding for you. Check the documentation of your
376 framework to see if that's the case.
377 This design decision was made so it gives more flexibility to PSGI
379 applications and frameworks, without putting complicated work into PSGI
380 web servers and interface specification itself.
381 No iterators support in $body?
383 We learned that WSGI and Rack really enjoy the benefit of Python and
385 Ruby's language beauty, which are iterable objects in Python or
386 iterators in Ruby.
387 Rack, for instance, expects the body as an object that responds to the
389 "each" method and then yields the buffer, so
390 body.each { |buf| request.write(buf) }
392 would just magically work whether body is an Array, FileIO object or an
394 object that implements iterators. Perl doesn't have such a beautiful
395 thing in the language unless autobox is loaded. PSGI should not make
396 autobox as a requirement, so we only support a simple array ref or file
397 handle.
398 Writing an IO::Handle-like object is pretty easy since it's only
400 "getline" and "close". You can also use PerlIO to write an object that
401 behaves like a filehandle, though it might be considered a little
402 unstable.
403 See also IO::Handle::Util to turn anything iterators-like into
405 IO::Handle-like.
406 How should server determine to switch to sendfile(2) based serving?
408 First of all, an application SHOULD always set a IO::Handle-like object
410 (or an array of chunks) that responds to "getline" and "close" as a
411 body. That is guaranteed to work with any servers.
412 Optionally, if the server is written in perl or can tell a file
414 descriptor number to the C-land to serve the file, then the server MAY
415 check if the body is a real filehandle (possibly using Plack::Util's
416 "is_real_fh" function), then get a file descriptor with "fileno" and
417 call sendfile(2) or equivalent zero-copy data transfer using that.
418 Otherwise, if the server can't send a file using the file descriptor
420 but needs a local file path (like mod_perl or nginx), the application
421 can return an IO::Handle-like object that also responds to "path"
422 method. This type of IO-like object can easily be created using
423 IO::File::WithPath, IO::Handle::Util or Plack::Util's "set_io_path"
424 function.
425 Middlewares can also look to see if the body has "path" method and does
427 something interesting with it, like setting "X-Sendfile" headers.
428 To summarize:
430 • When to serve static files, applications should always return a
432 real filehandle or IO::Handle object. That should work everywhere,
433 and can be optimized in some environments.
434 • Applications can also set IO::Handle like object with an additional
436 "path" method, then it should work everywhere again, and can be
437 optimized in even more environments.
438 What if I want to stream content or do a long-poll Comet?
440 The most straightforward way to implement server push is for your
442 application to return a IO::Handle-like object as a content body that
443 implements "getline" to return pushed content. This is guaranteed to
444 work everywhere, but it's more like pull than push, and it's hard to do
445 non-blocking I/O unless you use Coro.
446 If you want to do server push, where your application runs in an event
448 loop and push content body to the client as it's ready, you should
449 return a callback to delay the response.
450 # long-poll comet like a chat application
452 my $app = sub {
453 my $env = shift;
454 unless ($env->{'psgi.streaming'}) {
455 die "This application needs psgi.streaming support";
456 }
457 return sub {
458 my $respond = shift;
459 wait_for_new_message(sub {
460 my $message = shift;
461 my $body = [ $message->to_json ];
462 $respond->([200, ['Content-Type', 'application/json'], $body]);
463 });
464 };
465 };
466 "wait_for_new_message" can be blocking or non-blocking: it's up to you.
468 Most of the case you want to run it non-blockingly and should use event
469 loops like AnyEvent. You may also check "psgi.nonblocking" value to see
470 that it's possible and fallback to a blocking call otherwise.
471 Also, to stream the content body (like streaming messages over the
473 Flash socket or multipart XMLHTTPRequest):
474 my $app = sub {
476 my $env = shift;
477 unless ($env->{'psgi.streaming'}) {
478 die "This application needs psgi.streaming support";
479 }
480 return sub {
481 my $respond = shift;
482 my $writer = $respond->([200, ['Content-Type', 'text/plain']]);
483 wait_for_new_message(sub {
484 my $message = shift;
485 if ($message) {
486 $writer->write($message->to_json);
487 } else {
488 $writer->close;
489 }
490 });
491 };
492 };
493 Which framework should I use to do streaming though?
495 We have servers that support non-blocking (where "psgi.nonblocking" is
497 set to true), but the problem is that framework side doesn't
498 necessarily support asynchronous event loop. For instance Catalyst has
499 "write" method on the response object:
500 while ($cond) {
502 $c->res->write($some_stuff);
503 }
504 This should work with all servers with "psgi.streaming" support even if
506 they are blocking, and it should be fine if they're running in multiple
507 processes ("psgi.multiprocess" is true).
508 Catalyst::Engine::PSGI also supports setting an IO::Handle-like object
510 that supports "getline", so using IO::Handle::Util
511 my $io = io_from_getline sub {
513 return $data; # or undef when done()
514 };
515 $c->res->body($io);
516 And that works fine to do streaming, but it's blocking (pull) rather
518 than asynchronous server push, so again you should be careful not to
519 run this application on non-blocking (and non-multiprocess) server
520 environments.
521 We expect that more web frameworks will appear that is focused on, or
523 existent frameworks will add support for, asynchronous and non-blocking
524 streaming interface.
525 Is psgi.streaming interface a requirement for the servers?
527 It is specified as SHOULD, so unless there is a strong reason not to
529 implement the interface, all servers are encouraged to implement this
530 interface.
531 However, if you implement a PSGI server using an Perl XS interface for
533 the ultimate performance or integration with web servers like Apache or
534 nginx, or implement a sandbox like environment (like Google AppEngine
535 or Heroku) or distributed platform using tools like Gearman, you might
536 not want to implement this interface.
537 That's fine, and in that case applications relying on the streaming
539 interface can still use Plack::Middleware::BufferedStreaming to
540 fallback to the buffered write on unsupported servers.
541 Why CGI-style environment variables instead of HTTP headers as a hash?
543 Most existing web application frameworks already have code or a handler
545 to run under the CGI environment. Using CGI-style hash keys instead of
546 HTTP headers makes it trivial for the framework developers to implement
547 an adapter to support PSGI. For instance, Catalyst::Engine::PSGI is
548 only a few dozens lines different from Catalyst::Engine::CGI and was
549 written in less than an hour.
550 Why is PATH_INFO URI decoded?
552 To be compatible with CGI spec (RFC 3875) and most web servers'
554 implementations (like Apache and lighttpd).
555 I understand it could be inconvenient that you can't distinguish
557 "foo%2fbar" from "foo/bar" in the trailing path, but the CGI spec
558 clearly says "PATH_INFO" should be decoded by servers, and that web
559 servers can deny such requests containing %2f (since such requests
560 would lose information in PATH_INFO). Leaving those reserved characters
561 undecoded (partial decoding) would make things worse, since then you
562 can't tell "foo%2fbar" from "foo%252fbar" and could be a security hole
563 with double encoding or decoding.
564 For web application frameworks that need more control over the actual
566 raw URI (such as Catalyst), we made the "REQUEST_URI" environment hash
567 key REQUIRED. The servers should set the undecoded (unparsed) original
568 URI (containing the query string) to this key. Note that "REQUEST_URI"
569 is completely raw even if the encoded entities are URI-safe.
570 For comparison, WSGI (PEP-333) defines both "SCRIPT_NAME" and
572 "PATH_INFO" be decoded and Rack leaves it implementation dependent,
573 while fixing most of PATH_INFO left encoded in Ruby web server
574 implementations.
575 <http://www.python.org/dev/peps/pep-0333/#url-reconstruction>
577 <http://groups.google.com/group/rack-devel/browse_thread/thread/ddf4622e69bea53f>
578
580 WSGI's FAQ clearly answers lots of questions about how some API design
581 decisions were made, some of which can directly apply to PSGI.
582 <http://www.python.org/dev/peps/pep-0333/#questions-and-answers>
584
586 If you have a question that is not answered here, or things you totally
587 disagree with, come join the IRC channel #plack on irc.perl.org or
588 mailing list <http://groups.google.com/group/psgi-plack>. Be sure you
589 clarify which hat you're wearing: application developers, server
590 implementors or middleware developers. And don't criticize the spec
591 just to criticize it: show your exact code that doesn't work or get too
592 messy because of spec restrictions etc. We'll ignore all nitpicks and
593 bikeshed discussion.
594
596 Tatsuhiko Miyagawa <miyagawa@bulknews.net>
597
599 Copyright Tatsuhiko Miyagawa, 2009-2010.
600 This document is licensed under the Creative Commons license by-sa.
602perl v5.34.0 2021-07-22 PSGI::FAQ(3)