On 11/21/2012 02:01 PM, David Moreau Simard wrote: > Hey there, > > Just wanted to confirm that it does seem to be a regression. > https://github.com/bestpractical/rt/commits/3.8-trunk/lib/RT/I18N.pm > https://github.com/bestpractical/rt/commit/ccafbd233e4f50bd8ef1889949e91dc6190f495b > > That eval seems to be necessary to have a "Fallback" encoding if an > e-mail is received with an encoding that doesn't exist/isn't recognized. > We've tested the eval inside the 3.8.15 I18N.pm and were able to receive > e-mails if they're a non-standard encoding again.
This behaviour is fixed and improved in 4.0. Upgrading when you can is your best bet. I don't believe 3.8 will see more work in this code, as our energy is going into 4.0 and the next major release series, 4.2. > For the time being, and I know this is not ideal, we've reverted only > I18N.pm to our previous version from 3.8.8 and everything works well again. I understand your motivation, but I'm leery of the compatibility of this. Please mention how your I18N.pm is special if you're debugging encoding issues on the list in the future. -------- We're hiring! http://bestpractical.com/jobs