1.1 Read This Section First!

This section answers the most common questions about Scoop which are seen by the developers. Please read this section first.

First, a few things you should know about Scoop.

• There are no .html files for you to edit. All customization is done through the web interface.
• You must have root or root-equivalent access to the web server. If you don't and can't get it, there are a few companies that will host a Scoop site for you.
• You will need to either know a lot about Apache or be prepared to learn. If you don't and don't want to, there are a few companies that will host a Scoop site for you.
• `apachectl restart' or `apachectl graceful' won't work, despite the Apache documentation's insistence otherwise. They do not free the memory used by Scoop for its cache; if you don't do a full `apachectl stop' and wait for it to terminate completely before doing an `apachectl start', your Scoop install will still have the old cache running and you will get some very very strange effects if you change anything in the database (such as during an upgrade). Some of these very strange effects can make you think that Scoop broke during an upgrade.
• If you have Scoop running but can't create accounts, there's probably a problem with your mail server. Check that your SMTP settings are correct; if so, make sure that the SMTP server referenced allows your web server to relay through it.
• If Scoop starts but you can't stay logged in, or you can log in but get ``Permission Denied'' when you try to do something, check the cookie_host setting in the Scoop section of your Apache config file. It should match the string you are currently using as a domain name to access your Scoop site, and it must have at least two dots in it. If you are using www.example.com you may omit the `www' and use just .example.com but if you are using test.example.com do not use .test.example.com because that extra period screws things up for some browsers. If you don't have a domain name, you can use the IP address as your cookie_host and as long as you then access your Scoop site using the IP address it will work, but when you get a domain name you'll have to remember to change it.

Scoop has a lot of dependencies, and if any of them aren't set up just right, Scoop will not work, or will not work properly.

To start with, though it sounds small, you need the full expat. Scoop will run deceptively well without it, but anything RDF (both fetching headlines from other sites and producing your own headline feed) will not work. To add RDF later, you'd basically need to reinstall from scratch, as several things need expat at compile-time. See section 2.1.1 for more details.

Scoop requires Apache 1.3.x with mod_perl, as it is a mod_perl program. If Scoop does nothing but give you a directory listing and you're sure you set it up properly, you probably don't have mod_perl working. We strongly recommend that you compile this yourself as described in section 2.2, as many packaged versions don't work with Scoop properly. The mod_perl may be compiled into Apache, or loaded as a DSO (dynamic shared object).

For those running Apache 2.x, sorry; if you want to run Scoop you'll have to downgrade, as Scoop has not yet been ported to mod_perl 2.0 yet. Some work has been done to make it work with Apache 2, but it is not considered suitable for a live site.

Scoop requires MySQL 3.22 or higher (including 4.0) to provide its database. There isn't really any special configuration required for Scoop, just follow the mySQL install instructions. Packaged versions work fine. As long as you can log in and get at the database from the webserver (Scoop and the db can be on separate machines if you like) you're fine.

Next, for Scoop to run it will need perl 5.005 or newer and the perl modules listed in section 2.1.2. The install script should take care of most of them, but some are notoriously problematic and need to be installed by hand. If you run into any trouble with the perl modules, read section 2.1.2 because it describes your problem.