RewriteRule in htaccess fails, but it could be my host - php

Question

I read that Self Sabotage is Not asking for help so here I am.
So, I have a site...it's working great using WAMP. It's working great on my current host. But I need to switch to a new host and now it's failing. I figured it's a .htaccess issue but now I'm not sure. On my current host I have no .htaccess file and it works great. On my localhost server using WAMP I had the same thing as on my new host but I just disabled the .htaccess file, renaming it to BAD.htaccess, and the site still works great. This is why I think it's a server-side problem and I need some help. On my WAMP server in vhosts I disabled +FollowSymLinks for that "domain". On my current host I had no easy way to do that so it's just whatever they gave me, but it works.
I am currently with Ionos and have switched to GreenGeeks, who use cPanel. So far I haven't found a vhosts file to edit to remove +FollowSymLinks, if that is even the problem.
Maybe it can be accomplished with .htaccess and if so here is what I need to do. First my current .htaccess:
Options +FollowSymLinks
RewriteEngine on
RewriteBase /
RewriteCond %{REQUEST_FILENAME} !-d
RewriteCond %{REQUEST_FILENAME} !-f
RewriteRule ^poems$ poems.php [R]
#RewriteRule ^poems/$ poems.php
RewriteRule ^collections$ collections/ [R]
RewriteRule ^collections/$ collections.php
RewriteRule ^poem$ poem/ [R]
RewriteRule ^poem/$ poem.php
RewriteRule ^poem/([0-9]+)/([a-zA-Z])$ poem.php?num=$1&poem=$2 [NC,L]
RewriteRule ^collection$ collection/ [R]
RewriteRule ^collection/$ collection.php
# RewriteRule ^poem/([0=9]+)$ indpoem.php?num=$1 [NC,L]
With the first two setups I can go to example.com/poems and it will redirect or rewrite to example.com/poems.php but still look like example.com/poems. Same with collections. On the new host those rewrite rules do rewrite it but the URL or URI shows example.com/poems.php, which I don't want per current SEO standards. Still, I could live with that.
However, when I get to the next level... example.com/poem/#/poem-name it fails on my new host. I do have a file called poem.php which it should rewrite to. In that file I use the following to get the # and name...
$URL = explode("/",$_SERVER['REQUEST_URI']);
So I don't have to do a _GET.
As you can see I tried to do a RewriteRule to change it from the first to example.com/poem?111&name, but that just seems silly because on WAMP I don't have to do anything. I could try rewriting it to the same URL again, but I have a feeling that won't work. And if it does it will probably be poem.php/#/name/
Any thoughts on a server config I'm missing when using cPanel. I even tried doing
Options -FollowSymLinks
in my .htaccess file with no success.
Any help would be appreciated. My WAMP and the new host have all the most recent versions of Apache and PHP.

Answer 1

There are a few issues here...
For the site to work without the .htaccess file at all (on your WAMP dev server and current host) then MultiViews (part of mod_negotiation) must have been enabled. It is MultiViews that "rewrites" /foo to /foo.php.
If MultiViews is enabled then your current .htaccess file is essentially overridden since the MultiViews content-negotiation occurs before your mod_rewrite directives are processed so your RewriteRule patterns fail to match.
MultiViews is disabled by default on Apache, it needs to be "explicitly" enabled. Unfortunately, some shared hosts do enable this in the server config (which causes more problems than it fixes - if you are not expecting it.)
On the new host those rewrite rules do rewrite it but the URL or URI shows example.com/poems.php.
RewriteRule ^poems$ poems.php [R]
Because MultiViews is disabled and your directives externally "redirect" /poems to /poems.php. There is no "rewrite" here. The R (redirect) flag triggers an external redirect.
RewriteRule ^poem$ poem/ [R]
RewriteRule ^poem/$ poem.php
However, for /poem you are redirecting to /poem/ (appending the trailing slash) but you have omitted the L flag so processing continues and the request is further rewritten to /poem.php. But because you have already triggered a Redirect, the request is "redirected" (not rewritten) to /poem.php, again exposing the .php.
Redirects should nearly always include the L flag. In fact, all your rules should include the L flag for optimisation and to prevent accidental conflicts.
Why are you redirecting to append the trailing slash (as if this is the preferred canonical URL)? You make no mention of this in your question text and only one of your examples includes a trailing slash on the URL as far as I can see? So, what is the preferred/canonical URL? What URL are you linking to? Incidentally, MultiViews will not append a trailing slash - so either this redirect is not required, or your site is not actually "working great" on WAMP / current host without the .htaccess file. (?) (Personally, I would not use a trailing slash.)
RewriteRule ^poem/([0-9]+)/([a-zA-Z])$ poem.php?num=$1&poem=$2 [NC,L]
:
$URL = explode("/",$_SERVER['REQUEST_URI']);
Your PHP script parses the requested URL-path (ie. $_SERVER['REQUEST_URI']), it is not referencing the query string. However, the above RewriteRule directive is rewriting to a query string - which would seem to be entirely superfluous ("silly" - as you suggest). Because you are not using the query string in your PHP script, the request still "works" when using MultiViews.
According to your PHP script, you simply need to rewrite the request to poem.php, without a query string. (Which is what MultiViews does. Although, strictly speaking, MultiViews rewrites /poem/123/name to /poem.php/123/name - passing the additional URL-path as path-info to poem.php.)
This regex only matches a single letter in the 3rd (name) path segment so it will fail to match a requested URL of the form /poem/123/name, so the request is not rewritten to poem.php and the request fails (with a 404 I suspect).
This regex also does not match a trailing slash. (So, is the trailing slash really canonical?)
RewriteCond %{REQUEST_FILENAME} !-d
RewriteCond %{REQUEST_FILENAME} !-f
These conditions are entirely superfluous where you have put them. RewriteCond directives only apply to the first RewriteRule directive that follows.
Options -FollowSymLinks
You need FollowSymLinks for mod_rewrite to work. Don't try to disable this. FollowSymLinks is actually the default Apache setting, so you only need to explicitly enable it (ie. +FollowSymLinks) in .htaccess if it has been disabled in the server config.
Solution
Personally, I would not use a trailing slash on the canonical URLs. (This is in line with most of your examples and the regex used in your rule.)
So, bringing the above points together:
Options +FollowSymLinks -MultiViews
RewriteEngine on
# Canonical redirect to remove trailing slash (from non-directories)
RewriteCond %{REQUEST_FILENAME} !-d
RewriteRule (.*)/$ /$1 [R,L]
# Internal rewrites to ".php"
RewriteRule ^poems$ poems.php [L]
RewriteRule ^collections$ collections.php [L]
RewriteRule ^poem(/\d+/[\w-]+)?$ poem.php [L]
RewriteRule ^collection$ collection.php [L]
On the other hand, if the canonical URL should include a trailing slash then change it accordingly:
Options +FollowSymLinks -MultiViews
RewriteEngine on
# Canonical redirect to append trailing slash to "all" URLs
RewriteRule !/$ ${REQUEST_URI}/ [R,L]
# Internal rewrites to ".php"
RewriteRule ^poems/$ poems.php [L]
RewriteRule ^collections/$ collections.php [L]
RewriteRule ^poem/(\d+/[\w-]+/)?$ poem.php [L]
RewriteRule ^collection/$ collection.php [L]
If you have many such URLs/pages then you can make this entirely "generic" without having to explicitly name each URL/file. ie. "poems", "collections", etc.
Alternatively, you simply enable MultiViews and let mod_negotiation rewrite the URLs. However, you will not be able to canonicalise the trailing slash or validate the request before rewriting and MultiViews applies to everything, not just your .php files, so potentially creates duplicate content. If you need to do any specific rewriting, such are rewriting the query string then MultiViews is likely to conflict.
Options +MultiViews