On Fri, Mar 9, 2012 at 1:40 PM, Joachim Breitner <m...@joachim-breitner.de> wrote: > Hi Sönke, > > Am Freitag, den 09.03.2012, 11:17 +0100 schrieb Sönke Hahn: >> Any comments welcome. > > you could elaborate the documenatation “Assumes UTF-8” – I guess this > only applies to the two ByteString variants, as String and Text _should_ > contain unicode codepoints and no encoding. Not that someone tries to > use a String where each Char corresponds to a byte in a UTF-8 encoded > string and thinks he can convert it correctly. > > I was about to suggest to merge this into the convertible package (to > fight package proliferation), but found that it seems it is already > there: > http://hackage.haskell.org/packages/archive/convertible-text/0.4.0.2/doc/html/src/Data-Convertible-Instances-Text.html > > Greetings, > Joachim > > -- > Joachim "nomeata" Breitner > m...@joachim-breitner.de | nome...@debian.org | GPG: 0x4743206C > xmpp: nome...@joachim-breitner.de | http://www.joachim-breitner.de/ > > > _______________________________________________ > Haskell-Cafe mailing list > Haskell-Cafe@haskell.org > http://www.haskell.org/mailman/listinfo/haskell-cafe >
I'm the author of convertible-text, and I consider it deprecated (it's marked as such in the synopsis). As far as string-conversions, I'm a little concerned that it's using decodeUtf8, which can throw exceptions from pure code for invalid UTF8 sequences. I would prefer decodeUtf8With lenientDecode. Actually, my *real* preference would be that a lenient decode was the default so that we aren't exposing a partial function as the default way to decode bytes. Michael _______________________________________________ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe