+++ This bug was initially created as a clone of Bug #1054902 +++ The logic in bug 1029951 have since changed to allow inclusion of dictionaries of other "IMEngines" to be downloaded, hence, one more dir level need to be added on the CDN to add the name of the IMEngines.

Benson, would you be able to verify, merge this and update the CDN? I simply move the files we put online in bug 1054902 because it was not used, and add a few more new dictionaries. Also, the version # goes after the IMEngine ID ("latin") because the data is parsed by the IMEngine and they would all change at once if the structure changes. Thanks!

Benson, would it be possible if we could turn on CORS for the server? Accordingly to curl it currently doesn't come with that header. $ curl -I https://fxos.cdn.mozilla.net/dictionaries/1/en_us.dict HTTP/1.1 200 OK Content-Type: binary/octet-stream Content-Length: 1451390 Connection: keep-alive Date: Fri, 09 Jan 2015 08:37:31 GMT Last-Modified: Wed, 20 Aug 2014 15:20:39 GMT ETag: "d906ff451a8e747be6a726cc378f087c" Server: AmazonS3 X-Cache: Miss from cloudfront Via: 1.1 6deb65d0ad86a5cd1d4fbb8a5b3cdd2c.cloudfront.net (CloudFront) X-Amz-Cf-Id: NHL_e_U98De9zpW1Je3heBCmPasFxiRrQbOzAoke51ky9kTzgDsV_A==

I updated the files in the S3 bucket. Please verify that your changes are OK. I also added a CORS policy on the bucket. The GET requests don't include the headers but clients shouldn't have any issues fetching from the CDN and S3.

Thanks! I manually verified CORS is working as expected, on the updated URLs.