If WordPress isn't installed at webroot export download doesn't work

[aoloe]

if wordpress is not installed at the root of the server, the export link does not respect the custon url.

in my case, the wordpress directory is not even at the root of the wordpress site...

[k1sul1]

Interesting.

The link that is shown on the notice indeed points to the root folder, even if your site doesn't exist there.
The handleDownload function also checks that the URL has /wplf-export as a guard, which is a bit problematic if your site resides on example.com/wordpress.

You should still be able to download the file and delete it manually.

Sorry for the inconvenience. This can be fixed in a few ways, the first idea I got was switching to query params completely, like checking $_GET['action'] === 'wplf-export' instead of parse_url($_SERVER["REQUEST_URI"], PHP_URL_PATH) === '/wplf-export').

I'm very busy in the coming weeks, so I don't have the time for fixing this and testing througly.

[k1sul1]

@aoloe Does this work for you? https://github.com/libreform/export/archive/allow-nonwebroot.zip

No guarantees that it works, but it just might.

[aoloe]

it works better.

but now i get a popup instead of the csv to be downloaded...

the content is:

"""
Are you sure you want to do this?
"""

the url of the popup is:

http://www.mysite.org/2018/wp/wp-admin/edit.php?action=wplf-export&wplf-export-download=%2Fhome%2Fmyuser%2Fhttpdocs%2Fwww.mysite.org%2Fhttpdocs%2F2018%2Fwp%2Fwp-content%2Fuploads%2Fwplf%2Fwplf_export_1515771766.csv

[k1sul1]

By popup, you mean the WordPress error screen? If it's a JavaScript confirm prompt it's coming from somewhere else.

I'm going with the assumption that it's the WordPress error screen. By the looks of the URL, it may be that WP doesn't like the action parameter on edit.php, but this is just a guess. By using get_site_url() in front of the URL, it should (probably) work.

5e8f54d

You can do that by editing these lines. I can't make a patch right now, but if you try this and it works, feel free to submit a pull request :)

[aoloe]

using get_site_url() in the notices hrefs does indeed solve the issue.