How to make Wordpress permalinks ignore custom url rewrite - wordpress

Question

i'm currently working on a wordpress site that needs to have the option to be offered in french. i've found a way to make the theme work with the fr_FR po and mo files when i add a querystring variable l. i.e.
site.tld will yield the vanilla english site, while site.tld/?l=fr will activate the following code in my functions.php to serve the french translation:
<?php
// http://codex.wordpress.org/Plugin_API/Filter_Reference/locale
add_filter( 'locale', 'set_my_locale' );
function set_my_locale( $lang ) {
if ("fr" == substr(strtolower(trim(strip_tags(stripslashes($_GET['l'])))), 0, 2)) {
// set language to french
return "fr_FR";
} else {
// return original locale
return $lang;
}
}
?>
this setup is already working. my question is: how do i rewrite the url so instead of site.tld/?l=fr i can just prepend the folder structure with fr i.e. site.tld/fr/?
so like if there's a page site.tld/portoflio/autumn/ in english, site.tld/fr/portfolio/autumn/ will spit out the french one. i got this idea from the apple website, where the language is always prepended to the folder structure.
i can already make this happen with an external redirect in htaccess:
RewriteBase /
RewriteCond ^[a-z]{2}/
RewriteRule ^([a-z]{2})/(.*)$ /$2?l=$1 [R,L]
this works, but once i remove the R flag it serves the french-translated 404 not found isntead. i'm guessing what i did is messing-up with wordpress' rewrite rules because i need to use pretty permalinks. right now i'm set to use Month and name (/%year%/%monthnum%/%postname%/) in the admin general options.
my question is, how do i make wordpress ignore the /fr/ part and still serve the correct page/post/etc?
is this even possible? am i on the right track here? i need your help especially doing what's WISE and not just what's NICE. i tried this http://pmg.co/a-mostly-complete-guide-to-the-wordpress-rewrite-api but for the life of me i can't make it work. :/
UPDATE: ok, so here's some progress taking cue from #relu below:
i made sure my rules were canned from .htaccess
i then added the rules from #relu but in the following way, because Relu's always sent me to the same address without the /fr:
<?
// http://pmg.co/custom-wordpress-shortlinks
add_action( 'init', 'pmgtut_add_rewrites' );
function pmgtut_add_rewrites() {
add_rewrite_rule( '^([a-z]{2})/?$', 'index.php?l=$matches[1]', 'top' );
add_rewrite_rule( '^([a-z]{2})/(.*)$', 'index.php?l=$matches[1]&q=$matches[2]', 'top' );
}
add_filter( 'query_vars', 'pmgtut_query_vars', 10, 1 );
function pmgtut_query_vars( $vars ) {
$vars[] = 'l';
return $vars;
}
?>
now the /fr stays in the address bar, so that's a good thing, but two problems persist:
wordpress is serving me the main index page. seems like it's not using the &q=$matches[2] part of the rule; and
the locale still doesn't get set properly. i checked if the variable l is getting fed, so i added echo 'l: $l'; after $l = get_query_var('l'); and interestingly i get two echoes: one l: and another l: fr right after that. is the locale filter being run twice? seems like the first time it doesn't see the value of the queryvar, and then there seems to be a second time that it outputs l: with fr already in there.. at the end of the day the locale still doesn't get changed.
aaaah help.
FINAL UPDATE: in my last breath of frustration i searched again, and found this plugin qtranslate. does what i need. thanks y'all, esp to #relu's epic effort and stayin pow'r.

Answer 1

Add this to your functions.php
add_filter( 'rewrite_rules_array','my_insert_rewrite_rules' );
add_filter( 'query_vars','my_insert_query_vars' );
add_action( 'wp_loaded','my_flush_rules' );
// flush_rules() if our rules are not yet included
function my_flush_rules(){
$rules = get_option( 'rewrite_rules' );
if ( ! isset( $rules['^([a-z]{2})/(.*)$'] ) ) {
global $wp_rewrite;
$wp_rewrite->flush_rules();
}
}
// Adding a new rule
function my_insert_rewrite_rules( $rules )
{
$newrules = array();
$newrules['^([a-z]{2})/(.*)$'] = 'index.php?l=$matches[1]&q=$matches[2]';
return $newrules + $rules;
}
// Adding the l var so that WP recognizes it
function my_insert_query_vars( $vars )
{
array_push($vars, 'l');
return $vars;
}
Haven't tested this, it may need some tweaking, but this is the way to do it.
I've basically adapted the version found here.
It also registers the "l" query var so you can get it's value by calling: get_query_var
add_filter( 'locale', 'set_my_locale' );
function set_my_locale( $lang ) {
$l = get_query_var('l');
if ("fr" == $l) {
// set language to french
return "fr_FR";
} else {
// return original locale
return $lang;
}
}
UDATE: Ignore the above!
Remove the my_insert_query_vars() filter and replace it with:
function register_rewrite_tag() {
add_rewrite_tag('%l%', '([a-z]{2})');
}
add_action('init', 'register_rewrite_tag');
Here are also the updated rewrites for pages and posts, I've looked at WordPress' $wp_rewrite->rules and adapted them to fit this special case:
function pmgtut_add_rewrites() {
add_rewrite_rule( '^([a-z]{2})/?$', 'index.php?l=$matches[1]', 'top' );
add_rewrite_rule( '^([a-z]{2})/(.?.+?)(/[0-9]+)?/?$', 'index.php?l=$matches[1]&pagename=$matches[2]&page=$matches[3]', 'top' ); // Pages
add_rewrite_rule( '^([a-z]{2})/([^/]+)(/[0-9]+)?/?$', 'index.php?l=$matches[1]&name=$matches[2]&page=$matches[3]', 'top' ); // Posts
}

Answer 2

maybe out of topic but here's another rewrite rule for languages of the form; en_us/ or jp/
add_rewrite_rule( '^([a-z]{2}_[A-Z]{2}|^[a-z]{2})/?$', 'index.php?l=$matches[1]', 'top' );