I think you're on to it, Jonathan. But I keep getting an maddening behavior in Eclipse.
I have Eclipse's Edit > Set Encoding set to UTF-8. I have response.flash='Español'. I actually see Español with my real app. I went over to the test case, which I had created with the web2py interface and edited in vim, and now I loaded it into Eclipse. I saw that the body of the code had Espa<?>ol in it, not Español as originally typed. I changed it to Español and now it prints as Español, thank heavens. In the same Eclipse session, I go back to my real app and once again type response.flash='Español' but it still comes out Español. How can that be? On Sep 4, 4:21 pm, Jonathan Lundell <[email protected]> wrote: > On Sep 4, 2011, at 12:13 PM, weheh wrote: > > > In the previous post, looks like our google groups app lost the > > improperly translated character. It should have read Espa<?>ol. The <? > > was a white question mark in a black diamond. > > That's what I saw. > > > With some further experimentation, doing something like this: > > > response.flash=('Español').decode('latin-1').encode('utf-8') > > > forces the output to display properly. Seems like a bug. > > It sounds like your editor is producing a latin-1-encoded file, and then > you're treating it as utf-8. > > If I run this file (generated in BBEdit as a utf-8 file): > > #!/usr/bin/env python > # -*- coding: utf-8 -*- > > print ('Español') > print ('Español').decode('latin-1').encode('utf8') > print ('Español').decode('utf-8').encode('utf8') > > I get this output: > > Español > Español > Español > > In case that middle line doesn't come through, the ñ has been converted to > A-tilde followed by plus-minus, while the first and third are correct.
