Skip to content
This repository

HTTPS clone URL

Subversion checkout URL

You can clone with HTTPS or Subversion.

Download ZIP

The awesome module that will run your parrot programs from within apache!

branch: master

Fetching latest commit…

Octocat-spinner-32-eaf2f5

Cannot retrieve the latest commit at this time

Octocat-spinner-32 loader
Octocat-spinner-32 module
Octocat-spinner-32 pudding
Octocat-spinner-32 .gitignore
Octocat-spinner-32 Configure.pl
Octocat-spinner-32 DESIGN.md
Octocat-spinner-32 ISSUES.md
Octocat-spinner-32 LICENSE
Octocat-spinner-32 Makefile
Octocat-spinner-32 README.md
Octocat-spinner-32 TESTING.md
Octocat-spinner-32 hinxed.winxed
README.md

mod_parrot

The awesome module that will run your parrot programs from within apache!

Prerequisites

  • Apache HTTPD 2.2 or newer
  • Apache APXS (included in apache-dev packages, or whatever they are called)
  • pkg-config
  • parrot > 4.4.0 including parrot_config
  • libicu if parrot was build with that

Configuration

An installation helper is not yet done, if you must install it on a 'live' system, simply make and copy the contents of ./build to the apache module directory like so:

cp -v ./build/* `apxs -q LIBEXECDIR`

Debian famously renames apxs to apxs2, so you will have to use that instead. After that, edit your local httpd.conf to load and activate the module:

LoadModule modules/mod_parrot.so

Quite probably, if you are on windows and you did get this far, you will have to rename .so to .dll. (If you did you should contact me, and I will celebrate the thing with you). The last thing you must edit are the languages on which mod_parrot will respond. It currently uses a rather simplistic mapping from filenames to languages, so you will have to add

ParrotLanguage winxed .winxed .wxd .wnxd

to make mod_parrot respond to requests ending in .winxed, .wxd, or .wnxd. In essence, this should work for perl6 and other supported languages, but unhappily does not for nqp (the perl6 bootstrapping language).

Note that I'd like to move this over to a mime-based approach, but that is not yet done nor totally figured out.

Loaders

mod_parrot uses scripts known as 'loaders' to determine how your program is invoked. The default loader is the CGI loader, which emulates a CGI enviroment inside the apache process. This loader can only ever be used in single-threaded MPM's such as prefork. (If you do not know what this is and are using unix, you are probably safe, however on windows this is very much untrue). A much wiser choice is the PSGI loader. Read about the PSGI loader on its website, which is also multi-thread compatible because it does not have to manipulate the enviroment or change directories. As you might notice if you are a ruby or python programmer, PSGI is nearly exactly equal to WSGI / Rack. This is intentional, because it hopefully allows many different language frameworks to be ported to mod_parrot with minimal effort.

The parrot PSGI implementation calls the main subroutine from the requested file, CGI style. Silly, right, I know. mod_parrot should really have some way to specify the PSGI entry-point, but it isn't quite there yet.

Whats next and coming

  • Run scripts in a child-interpreter. The trick here is rethrowing exceptions.
  • Respond to requests based on mime type (via mod_mime). The actual language should be determined by the contents of the file.
  • Store a persistent (pool of) interpreter(s) for swiftly handling responses.
  • Run NQP as a language, which is troubling, and perl6, which should not be so, but which I can't test right now.
  • Add a way to run a script upon errors (ParrotError option?) and preferably, upon startup (ParrotConfigure) and receiving requests (ParrotRoute or ParrotReceive)
  • And as mentioned earlier, a way to specify the entrypoint for PSGI applications (ParrotEntry? That sounds pretty much alright.)
  • Cache PackfileViews obtained by compilation, so scripts can run faster.
  • So much more!
Something went wrong with that request. Please try again.