AO3 News

Post Header

2015-03-15 17:37:37 UTC

This release contains one bug fix and several background changes that will enable us to upgrade to newer versions of Ruby and Rails in the future.


  • Coders: james, Enigel, Sarken
  • Code reviewers: Ariana, Elz, Enigel, Sarken, Scott
  • Testers: Lady Oscar, mugenmine


  • We've upgraded to a more recent version of Rails, the framework the Archive runs on.
  • To make our downloads compatible with as many systems as possible, we were using an application called iconv to clean up their metadata (e.g. a work titled "Un Niño" would have a download file named, for example, "Un Nino.epub"). However, iconv will not be supported in newer versions of Ruby, so we have switched to a new method of cleaning the metadata.
  • The YAML parser we previously used to read our configuration files (which control things like how many works are listed on a page) is not supported in newer versions of Ruby. We've now switched to Psych, which is the default parser in Ruby 1.9.2 and up.
  • Attempting to access the Edit page for a Tag Set containing more than 20 tags of a single type was giving an error page. Now it should work again.

Known Issues

The new code for generating MOBI/EPUB/PDF/HTML file names automatically transliterates languages that use non-Latin alphabets, e.g. Japanese or Russian. We can't control how work titles are transliterated, and since it's an automated process, some results will not be correct. Previously, however, titles composed entirely of non-Latin characters were blanked, leading to identical file names like "-.epub", so this is an improvement. The transliteration will only affect the file name and not the contents of the file itself. In particular, titles will still appear correctly in your e-reader or app!

See our Known Issues page for other current issues.