Thread: Future Expires help

So adding a dot makes it possible to modify the url, sort of?

Er...I'm not sure what you mean there.

It's just a way of keeping things organised. There are some small technical differences between subdomains and simple folders -- such as how cookies can be set, and whether you get another DNS lookup -- but basically you're just using them for convenience.

In particular, a subdomain is cheaper than running a completely separate domain (which would incur a registrar fee).

The "multisite" setup is especially "tidy", because you can keep all the settings separate. It's like having a separate account.

What about site navigation, how do you avoid having to rename the links?

What about site navigation, how do you avoid having to rename the links?

Originally Posted by Jacob

Wherever practical, make all your links relative. For example:

• Good: "/articles/jam/raspberry"
• Bad: "dev.domain.com/articles/jam/raspberry"

(If you're seriously into performance, you can even insert links using a function that automatically creates the shortest possible URL!)

This works well within your HTML code, but sometimes in your application code (e.g. PHP) you might need full URLs. You can construct these with a PHP function.

For example, on my site, a full URL in PHP could be set like this:

Code:

$URL = $site->domain() . "/articles/jam/raspberry";

$site is an object that I've already instantiated. It's domain() method returns the domain part of the URL (e.g. "http://www.domain.com/"). I can supply complete code if you like.

(If you're seriously into performance, you can even insert links using a function that automatically creates the shortest possible URL!)

How many kilobits do you think you have saved?

/articles/jam/raspberry

That is what I thought, then you could browse your site as if it were live yeah?

How do you set up your localhost enviroment, do you use Wamp?

I can supply complete code if you like.

Yeah I would be interested to see how it works and if I could apply myself

Thanks

How many kilobits do you think you have saved?

Originally Posted by Jacob

Not many. It's the sort of thing that Yahoo would do. It would mainly be useful on pages with long lists of links.

That is what I thought, then you could browse your site as if it were live yeah?

Yes, exactly.

How do you set up your localhost enviroment, do you use Wamp?

Yes. I have four copies of my site:

• Localhost, for whatever I'm currently working on.
• A development site on the server.
• A staging site on the server (I don't use this much).
• The live site.

I use Mercurial, a version control system, to manage the files. In practice, I almost always treat Localhost as the "central" repository, and push changes out to the remote repositories on dev, stage, or live.

Yeah I would be interested to see how it works and if I could apply myself

Thanks

Bear in mind this is just what works for me. I imagine there are better implementations from, say, a PHP framework like CakePHP.

I've redacted some directory names.

Code:

class Site {    // Resolves between dev, stage, and live sites

    function branch() {
        switch (htmlentities(dirname(__FILE__), ENT_QUOTES)) {
            case "/home/live_user/public_html":
            $branch = "live";
            break;

            case "/home/stage_user/public_html":
            $branch = "stage";
            break;

            case "/home/dev_user/public_html":
            $branch = "dev";
            break;

            case "C:\\local\\path_to_website\\public_html":
            $branch = "local";
            break;

            default:
            die();
        }
        return $branch;
    }


    function domain() {
        switch ($this->branch()) {
            case "dev": case "stage":
            $root = "http://www.".$this->branch().".badmintonbible.com/";
            break;

            case "live":
            $root = "http://www.badmintonbible.com/";
            break;

            case "local":
            $root = "http://localhost/";
            break;

            default:
            die();
        }
        return $root;
    }

    function web_root() {
        switch ($this->branch()) {
            case "dev":
            $root = "/home/dev_user/public_html/";
            break;

            case "stage":
            $root = "/home/stage_user/public_html/";
            break;

            case "live":
            $root = "/home/live_user/public_html/";
            break;

            case "local":
            $root = "C:\\local\\path_to_website\\public_html"";
            break;

            default:
            die();
        }
        return $root;
    }

    function bin_root() {
        switch ($this->branch()) {
            case "local":
            $root = "C:\\local\\path_to_website\\bin_directory"";
            break;

            case "dev":
            $root = "/home/dev_user/bin_directory/";
            break;

            case "stage":
            $root = "/home/stage_user/bin_directory/";
            break;

            case "live":
            $root = "/home/live_user/bin_directory/";
            break;

            default:
            die();
        }
        return $root;
    }

    function current_page() {
        $path = htmlentities($_SERVER['PHP_SELF'], ENT_QUOTES);
        return $path;
    }

    function current_page_full() {
        $host = htmlentities($_SERVER['HTTP_HOST']);
        $path = "http://".$host.$this->current_page();
        return $path;
    }
}

$site = new Site();

Back tracking here:

The development and staging sites should probably be password-protected.

Why is this?

Thanks

You don't particularly want visitors and search engines discovering your development environment. It's possible that security won't be as thorough amongst other things.

Bear in mind this is just what works for me. I imagine there are better implementations from, say, a PHP framework like CakePHP.

Originally Posted by Mike Hopley

This is the perfect place for an abstract factory pattern implementation Bit complexer, but a handy design pattern to know and a lovely place to practice.

This is the perfect place for an abstract factory pattern implementation Bit complexer, but a handy design pattern to know and a lovely place to practice.

Originally Posted by janvt

Sounds interesting.

What benefit would the abstract factory pattern give, compared to my naive code?

In this case: none, seeing as you'd end up with more code than you have now. Was just pointing out that this a text book example. Actually you're almost using a factory implemetation, you're just using switch blocks in the class that defines your interface instead of implementing the logic in child classes. The factory pattern would start making sense if you had, say, differences in the implementation of the "current_page" function.