Friday, February 5, 2016

Why Drupal sucks

I use "sucks" here as a terminus technicus meaning "doesn't work in the way desired or expected".

Drupal "sucks" because it gets upgraded every couple of months and you have to upgrade. Now that's fine if you are using it in a business context where there are constantly paid employees to do that. But on a free or academic site, once set up it will be only occasionally updated. And when that new release comes out it effectively advertises the flaws to all the hackers of the universe: "Hey guys! You can hack Drupal sites with this neat exploit if the Drupal version is current-version minus 1". And of course they do. So you have to upgrade, and pronto, but you can't afford to. So you get hacked. Then you have to strip down your entire server because you don't know what the hackers did. And that really sucks.

Now all this would be gone if only the Drupal people would provide an automatic or at least easy upgrade path for new releases. Instead you have to take your site offline, then back up your database (why? Is the new release buggy?), then download the new server core, delete all your old stuff except the sites directory – and don't forget to cut and paste all those arcane rewrite rules from the old to the new .htaccess file – then install the new server, copy back the old sites, run the update script, take it back online and hope and pray that all your modules still work. They might not, especially that one you can't do without, whose author didn't bother to patch either. All this takes about half an hour, and you have to do it every 2 months on every damn Drupal site you maintain. So why not automate it? Is it really so hard? And don't tell me drush does all that because it only automates a couple of the easiest steps, and then you also have to worry about whether drush worked. I notice that Drupal 8 doesn't have an automatic update feature either. So that's why I think Drupal sucks. You're better off writing your own CMS. At least that way no one will know how to hack it.

mod_rewrite: redirect to another directory with additional parameters

Mod_rewrite is one of the hardest pieces of apache to get right: the documentation is awful, there is no way to test it and the syntax is arcane. All I wanted to do was redirect URLs from one directory to another, copying the existing GET parameters and adding some more. But there are a number of tricks to get it right. So my original URLs look like this:

And I wanted to redirect it to:

An additional requirement was that I wanted to add the rule to .htaccess in the CMS-directory, not root, since that's where the rules applied. The following rule worked in the "mycms" directory:

  RewriteEngine on
  RewriteRule ^olddir$ /mycms/newdir?param2=poo&param3=roo  [L,R,QSA]
  ... (other rules)

Since I was using Drupal, the <IfModule... was already present in the existing .htaccess file. So all I needed to do was add the correct RewriteRule. To explain what I understand by this: "^" means a URL starting with "mydir" since we are already in mycms (that's where the .htaccess file is), and ending ($) there. So only parameters follow. Now redirect to /mycms/newdir. If you don't start with "/" it will prepend the entire server path. Next, add your desired parameters, and finally append the existing ones via the QSA flag, which means "query string append". The "R" flag is needed because this is a redirect. The "L" flag is needed because otherwise other rules might be applied and this is the "last" rule we need here. I'm not sure all this is correct, but it is simple and it works. So I'm sticking with it.