I have a site in chinese. I use pathauto to create the path, like this mydomain/path1_chinese_text/[node:title], the node-title is in chinese too.
In browser it is displayed correctly. But in boost cache-directory, it is displayed like this : cache/normal/domain/法兰克福/法兰克福四川楼_.html.

It has been cached, but can not be read from cache.
Has anyone idea?

thanks

Comments

Anonymous’s picture

Anonymous (not verified) commented

If you have a dedicated server/ VPS you should be able to enable the RewriteDebug log, what you are seeing on your terminal is almost certainly just a "normal" characterset translation problem, but you need to find out what your webserver is looking for (can't be turned on in a .htaccess). Everything should be configured to work in UTF8 but boost does work when this is done, as we have arabic/ japanese and successful russian installations.

It's probably worthwhile to read around threads about multi-liingual sites and boost, to confirm your .htaccess settings are correct but it's really a system's administration / configuration task.

wanneng’s picture

wanneng commented

I have a virtual server with full access. I dont know how to "enable the RewriteDebug log", could give me more tip or some help link for this topic? Is it a php config or apache config?
thanks

Anonymous’s picture

Anonymous (not verified) commented

Apache

http://httpd.apache.org/docs/2.2/mod/mod_rewrite.html#rewritelog

Drupal8’s picture

Drupal8 commented

I have the same problem with a greek url names. I using a shared hosting and a lightspeed server.
Any progress?
Anyone with successfully use of Boost can kindly send us some informations.

Anonymous’s picture

Anonymous (not verified) commented
Assigned: wanneng »
Status: Active » Closed (duplicate)

As noted here #2052049-2: Boost will not cache non-latin urls, this is a systems administration issue to do with shared libraries and boost works as designed, there is no issue in handling the filenames the browser will transmit the URI in UTFB or whatever character encoding is set, the server needs to be configured to handle the filenames produced and have the relevant character sets installed. A simple test


touch("cache/ψάρι.html");

which should create an empty file called fish.html in greek in the cache. If it's not created then it's a systems admin problem to provide the chracterset for PHP, the server and filesystem, you'll note that above requires no character set conversion, no utf8_encode etc... and should display correctly in your browser (unless your browser cannot handle UTF8) being sent through a varnish server after having been cut and pasted from google translate.

Drupal8’s picture

Drupal8 commented

After i contact the sysAdmin of the hosting Company the see no error from their part.
When i navigate from my url mysite.com/cache/normal/www.mysite.com/el/ψάρι_.html The cached version show ok.
But when i go normal mysite.com/ψάρι show uncached version...

Anonymous’s picture

Anonymous (not verified) commented

So the problem appears not to be internalisation but something with your drupal/ boost configuration.

The first steps are always to try and get a cached version of index.html as this is almost always requested with no query string.

I would suggest that you start here

https://groups.drupal.org/lighttpd

and possibly here (although it appears old)

http://www.cyberciti.biz/tips/lighttpd-drupal-htacess-seo-clean-urls.html

since the issue is almost certainly not boost but something to do with the differences between an apache and lighttpd configuration.

Drupal8’s picture

Drupal8 commented

the / end /el pages show ok with boost comment at the end...
I will probably open an issue for the configuration of server .