Take the 2-minute tour ×
Code Review Stack Exchange is a question and answer site for peer programmer code reviews. It's 100% free, no registration required.

I want to "canonicalize" URLs for my static (files and folders) website.

The 'Aims' describes what I want to accomplish. The 'Code' gives my current .htaccess.

Everything works right now, but I wonder if the code could be improved:

  • any changes for better performance?
  • can some rules be removed?
  • can some rules be merged?
  • can something be shortened?
  • are some explaining comments wrong?

In my .htaccess there shouldn't be anything else as what is described in 'Aims' and my examples. So if there should be something that has nothing to do with it, it is probably unneeded (if I don't miss an important part right now).

Aims

  • Strip the file ending .html (but keep all other endings).
  • if the file is index.html, strip "index", too, and don't keep a (folder) trailing slash
  • no trailing slashs for files or folders
    • if someone adds a trailing slash, redirect to variant without

Example 1

  • Physical file: example.com/foo/bar.html
  • Desired URL: example.com/foo/bar
  • URLs that should redirect (301) to desired URL:
    • example.com/foo/bar.html
    • example.com/foo/bar/

Example 2 (if index.html)

  • Physical file: example.com/foo/index.html
  • Desired URL: example.com/foo
  • URLs that should redirect (301) to desired URL:
    • example.com/foo/index.html
    • example.com/foo/index
    • example.com/foo/
    • example.com/foo.html

Example 3 (if non-HTML file)

  • Physical file: example.com/foo/bar.png
  • Desired URL: example.com/foo/bar.png (= same as physical)
  • URLs that should redirect (301) to desired URL:
    • none

Code (.htaccess)

# Turn MultiViews off. (MultiViews on causes /abc to go to /abc.ext.) 
Options +FollowSymLinks -MultiViews

# It stops DirectorySlash from being processed if mod_rewrite isn't. 
<IfModule mod_rewrite.c>

  # Disable mod_dir adding missing trailing slashes to directory requests.
  DirectorySlash Off

  RewriteEngine On

  # If it's a request to index(.html) 
  RewriteCond %{THE_REQUEST} \ /(.+/)?index(\.html)?(\?.*)?\  [NC]
  # Remove it. 
  RewriteRule ^(.+/)?index(\.html)?$ /%1 [R=301,L]

  # if request has a trailing slash
  RewriteCond %{REQUEST_URI} ^/(.*)/$
  # but it isn't a directory
  RewriteCond %{DOCUMENT_ROOT}/%1 !-d
  # and if the trailing slash is removed and a .html appended to the end, it IS a file
  RewriteCond %{DOCUMENT_ROOT}/%1.html -f
  # redirect without trailing slash
  RewriteRule ^ /%1 [L,R=301]

  # Add missing trailing slashes to directories if a matching .html does not exist. 
  # If it's a request to a directory. 
  RewriteCond %{REQUEST_FILENAME}/ -d
  # And a HTML file does not (!) exist.
  RewriteCond %{REQUEST_FILENAME}/index.html !-f
  # And there is not trailing slash redirect to add it. 
  RewriteRule [^/]$ %{REQUEST_URI}/ [R=301,L] 

  RewriteCond %{REQUEST_FILENAME} -d
  # And a HTML file exists.
  RewriteCond %{REQUEST_FILENAME}/index.html -f
  # And there is a trailing slash redirect to remove it. 
  RewriteRule ^(.*?)/$ /$1 [R=301,L]

  RewriteCond %{REQUEST_FILENAME} -d
  # And a HTML file exists.
  RewriteCond %{REQUEST_FILENAME}/index.html -f
  # And there is no trailing slash show the index.html. 
  RewriteRule [^/]$ %{REQUEST_URI}/index.html [L]

  # Remove HTML extensions. 
  # If it's a request from a browser, not an internal request by Apache/mod_rewrite. 
  RewriteCond %{ENV:REDIRECT_STATUS} ^$
  # And the request has a HTML extension. Redirect to remove it. 
  RewriteRule ^(.+)\.html$ /$1 [R=301,L]

  # If the request exists with a .html extension. 
  RewriteCond %{SCRIPT_FILENAME}.html -f
  # And there is no trailing slash, rewrite to add the .html extension. 
  RewriteRule [^/]$ %{REQUEST_URI}.html [QSA,L]

</IfModule>

(some parts of this file are by Jon Lin over at Stack Overflow)

share|improve this question
add comment

1 Answer

First, I'll start with an observation that applying these rewriting rules — in particular, removing trailing slashes in the URL — can break pages that reference relative URLs. I assume you know that and want to proceed anyway.

In general, you should realize that every RewriteRule is conditional. If the path does not match the pattern, the RewriteRule and all of its RewriteConds are skipped. You should take advantage of that fact to eliminate a few superfluous RewriteConds.

You have seven RewriteRules:

  1. Rule 1

    # If it's a request to index(.html) 
    RewriteCond %{THE_REQUEST} \ /(.+/)?index(\.html)?(\?.*)?\  [NC]
    # Remove it. 
    RewriteRule ^(.+/)?index(\.html)?$ /%1 [R=301,L]
    

    It is highly unorthodox to use %{THE_REQUEST}, which works at the raw HTTP level (e.g., GET /index.html HTTP/1.1). It contains the raw URL, whereas you usually care about the decoded path. The fix is to remove the RewriteCond altogether, since it is redundant anyway.

    I recommend incorporating Rule 6 into this rule as a simplification.

  2. Rule 2

    # if request has a trailing slash
    RewriteCond %{REQUEST_URI} ^/(.*)/$
    # but it isn't a directory
    RewriteCond %{DOCUMENT_ROOT}/%1 !-d
    # and if the trailing slash is removed and a .html appended to the end, it IS a file
    RewriteCond %{DOCUMENT_ROOT}/%1.html -f
    # redirect without trailing slash
    RewriteRule ^ /%1 [L,R=301]
    

    Again, the first RewriteCond should be incorporated into the RewriteRule instead. There's no need to use %{REQUEST_URI}, since RewriteRule naturally works on paths.

    Avoid hard-coding the assumption that the resources are relative to the document root. They may have been remapped to another portion of the filesystem.

  3. Rule 3

    # Add missing trailing slashes to directories if a matching .html does not exist. 
    # If it's a request to a directory. 
    RewriteCond %{REQUEST_FILENAME}/ -d
    # And a HTML file does not (!) exist.
    RewriteCond %{REQUEST_FILENAME}/index.html !-f
    # And there is not trailing slash redirect to add it. 
    RewriteRule [^/]$ %{REQUEST_URI}/ [R=301,L] 
    

    Avoid referencing %{REQUEST_URI} in the RewriteRule. Instead, use regular expression capturing:

    RewriteRule (.*[^/])$ $1/ [R=301,L]
    
  4. Rule 4

    RewriteCond %{REQUEST_FILENAME} -d
    # And a HTML file exists.
    RewriteCond %{REQUEST_FILENAME}/index.html -f
    # And there is a trailing slash redirect to remove it. 
    RewriteRule ^(.*?)/$ /$1 [R=301,L]
    

    The directory test is superfluous. Also, since Rule 2 also strips trailing slashes, I would swap this with Rule 3 so that the rules to strip trailing slashes are placed together.

  5. Rule 5

    RewriteCond %{REQUEST_FILENAME} -d
    # And a HTML file exists.
    RewriteCond %{REQUEST_FILENAME}/index.html -f
    # And there is no trailing slash show the index.html. 
    RewriteRule [^/]$ %{REQUEST_URI}/index.html [L]
    

    Again, the directory test is superfluous, and you can avoid referencing %{REQUEST_URI} by using regular expression capturing.

  6. Rule 6

    # Remove HTML extensions. 
    # If it's a request from a browser, not an internal request by Apache/mod_rewrite. 
    RewriteCond %{ENV:REDIRECT_STATUS} ^$
    # And the request has a HTML extension. Redirect to remove it. 
    RewriteRule ^(.+)\.html$ /$1 [R=301,L]
    

    As previously mentioned, this can be incorporated into Rule 1.

  7. Rule 7

    # If the request exists with a .html extension. 
    RewriteCond %{SCRIPT_FILENAME}.html -f
    # And there is no trailing slash, rewrite to add the .html extension. 
    RewriteRule [^/]$ %{REQUEST_URI}.html [QSA,L]
    

    Use regular expression capturing. Also, use of %{SCRIPT_FILENAME} is weird and inconsistent with the rest of the rules.

I found that the ruleset was prone to infinite redirects. As you probably found, when the non-redirecting rules rewrite the URL, they trigger an internal subrequest in Apache, causing the entire ruleset to be evaluated again. You probably encountered this problem and put in the check for %{ENV:REDIRECT_STATUS} as a workaround. I would generalize that check by making it the very first rule.

Options +FollowSymLinks -MultiViews

<IfModule mod_rewrite.c>
  # Disable mod_dir adding missing trailing slashes to directory requests.
  DirectorySlash Off

  RewriteEngine On

  ######################################################################
  # Canonicalizing redirects
  ######################################################################

  # Skip all rewrites of internal subrequests (see below).
  RewriteCond %{ENV:REDIRECT_STATUS} .
  RewriteRule . - [L]

  # Strip .html or /index.html
  RewriteRule ^(.*?)(/index)?(\.html)$ $1 [NC,R=301,L]

  # Strip trailing slash if...
  # It isn't a directory, and if the trailing slash is removed and a .html
  # appended to the end, it IS a file.
  RewriteCond %{REQUEST_FILENAME} !-d
  RewriteCond %{REQUEST_FILENAME} (.*?)/?$
  RewriteCond %1.html -f [OR]
  RewriteRule (.*)/$ $1 [R=301,L]

  # Strip trailing slash if...
  # It is a directory that contains an index.html file.
  RewriteCond %{REQUEST_FILENAME} (.*?)/?$
  RewriteCond %1/index.html -f
  RewriteRule (.*)/$ $1 [R=301,L]

  # Add trailing slash if...
  # If is a directory and index.html does not exist.
  RewriteCond %{REQUEST_FILENAME} -d
  RewriteCond %{REQUEST_FILENAME}/index.html !-f
  RewriteRule (.*[^/])$ $1/ [R=301,L]

  ######################################################################
  # URL-to-filesystem mapping
  #
  # Even with the [L] flag, these mapping rules will trigger an internal
  # subrequest, causing the mod_rewrite ruleset to be re-evaluated.
  # Therefore, to prevent infinite recursion, the canonicalizing
  # redirect rules above need to skipped if %{ENV:REDIRECT_STATUS} is
  # set.
  ######################################################################

  # If there is no trailing slash, try appending /index.html
  RewriteCond %{REQUEST_FILENAME}/index.html -f
  RewriteRule (.*[^/])$ $1/index.html [L]

  # If there is no trailing slash, try appending .html
  RewriteCond %{REQUEST_FILENAME}.html -f
  RewriteRule (.*[^/])$ $1.html [QSA,L]
</IfModule>
share|improve this answer
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.