Skip to content

Latest commit

 

History

History
49 lines (32 loc) · 2.72 KB

pjax-lumen.md

File metadata and controls

49 lines (32 loc) · 2.72 KB

If you follow Jeffrey Way on Laracasts you may have seen recently he added a video on how to add Pjax to your Laravel application. Well, if you're using Lumen you may have noticed that the middleware doesn't quite work the same.

If you havent already, firstly go check out the videos on Laracasts. There are two, one for setting up Pjax, the other for rewriting the Middleware so you get to know exactly whats going on.

You can also take a look at Jeffrey's Middleware on Github's Gists as I will be referencing it below.

When go through the tutorial and run this, you will find that Lumen gives you an error:

RuntimeException in Crawler.php

line 693: Unable to filter with a CSS selector as the Symfony CssSelector is not installed (you can use filterXPath instead).

Thats because Lumen is a cut down version of Laravel to make it super quick and obviously one of the bits Taylor decided to cut out was the symfony CssSelector class. No big deal, we could add it. Then we would just be adding bulk to our lean Lumen application. Instead we can adjust the Middleware.

First take a look at exactly what the error says, you can use filterXPath instead - really, its giving you the answer. Lets use filterXPath instead.

Jeffrey uses the filter function in two places, and thats whats causing your error. The first is in makeTitle() on line 58. This one is easy to change we just need to swap out filter for filterXPath and modify the selector from 'head > title' to 'descendant-or-self::head/title'. Job done.

That gives us a new makeTitle() function that looks like this,

protected function makeTitle($crawler)
{
    $pageTitle = $crawler->filterXPath('descendant-or-self::head/title')->html();

    return "<title>{$pageTitle}</title>";
}

The next occurance is in fetchContents just below. We have 2 issues here rolled into one. We need to do the same as before and switch out filter for filterXPath but we need to make sure we are still referencing the right container. In Jeffrey's Middleware $container = '#pjax-container' but we cant use that in filterXPath.

Without going into too much detail, we can chop off the # bit using ltrim and then concatenate the string using XPaths @id selector like so,

protected function fetchContents($crawler, $container)
{
    $container = ltrim($container, '#');
    $content = $crawler->filterXPath('//div[@id = "'. $container .'"]');

    if (! $content->count()) {
        abort(422);
    }

    return $content->html();
}