"Curl: Download Error" with relative path in anchor tag

[MartijnThomas]

Great package and extremely easy API, I bumped into one problem though. When I click on a link that has a relative path, I get a curl error:

Curl: Download Error: <url> malformed, status 0 on url /some-link

This is a node that failed for example:

<a href="/some-link">The title of the link</a>

The code used (with the default driver):

$page = new \Openbuildings\Spiderling\Page();

$page->visit('https://example.com/');

$urls = $page->all('.LinkItem h3 > a');

foreach($urls as $url)
{
    $url->click();
}

[halfer]

Just a guess here, @MartijnThomas : if you click a link, won't the browser follow it, and thus clicking other links in the same page doesn't make any sense? They presumably no longer exist.

This would be fine with JavaScript links, or anchor links, I would expect.

[MartijnThomas]

Not sure what you mean. Letting the crawler click on a link is kind of the idea of a crawler, and yes I'm sure the page existed, other crawlers crawled it fine.

[halfer]

You seem to have a list of links in $urls. I would guess that as soon as the first one is clicked, the PHP representation of the page is dropped, since following a link removes the page from memory and loads the new one.

I've not used \Openbuildings\Spiderling\Page in itself, but certainly in Web Driver with Selenium, or tools like Goutte, when you click on any link, any other links will no longer be valid, and cannot be clicked (with minor exceptions, see my previous note).

In your example, how many links are there? How many fail? My guess is the second one fails, for the reason given here.