Re: [GHC] #9114: Invalid UTF8 not round-tripped correctly

15 May 2014

      #9114: Invalid UTF8 not round-tripped correctly
-------------------------------------+------------------------------------
        Reporter:  nomeata           |            Owner:
            Type:  bug               |           Status:  new
        Priority:  normal            |        Milestone:
       Component:  libraries/base    |          Version:  7.6.3
      Resolution:                    |         Keywords:
Operating System:  Unknown/Multiple  |     Architecture:  Unknown/Multiple
 Type of failure:  None/Unknown      |       Difficulty:  Unknown
       Test Case:                    |       Blocked By:
        Blocking:                    |  Related Tickets:
-------------------------------------+------------------------------------

Comment (by nomeata):

 No, I believe the bug is in `GHC.IO.Encoding` or `GHC.IO.Encoding.UTF8`. I
 tried to give an example using that interface, but failed to work with
 these buffers; the following code always prints `'\NUL'`:

 {{{
 import GHC.IO.Encoding
 import GHC.IO.Buffer
 import GHC.IO.Encoding.Types

 main = do
     te <- getFileSystemEncoding
     case te of
       TextEncoding _ decIO _ -> do
         BufferCodec encode recover close getState setState <- decIO
         bb <- newByteBuffer 2 ReadBuffer
         writeWord8Buf (bufRaw bb) 0 0xC0
         writeWord8Buf (bufRaw bb) 1 0xB7
         checkBuffer bb
         cb <- newCharBuffer 1 WriteBuffer
         (InputUnderflow,_,cb') <- encode bb cb
         close
         c <- peekCharBuf (bufRaw cb') 0
         print c
 }}}

--
Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/9114#comment:4
GHC http://www.haskell.org/ghc/
The Glasgow Haskell Compiler