Add a Text instance for ToSV and FromSV (see #15).

Not yet tested.

cbits/p5embed.c

@@ -326,6 +326,19 @@ perl5_init ( int argc, char **argv )
     return my_perl;
 }
 
+// switch on Utf8 flag
+void
+perl5_SvUTF8_on(SV *sv)
+{
+  SvUTF8_on(sv);
+}
+
+// convert SV to a Utf8 string
+char*
+perl5_sv_2pvutf8(SV* sv, STRLEN* lp) {
+  return sv_2pvutf8(sv, lp);
+}
+
 SV *
 perl5_get_sv(const char *name)
 {

cbits/p5embed.h

@@ -31,7 +31,8 @@ SV * perl5_newSVnv ( double iv );
 bool perl5_SvTRUE ( SV * sv );
 SV * perl5_get_sv ( const char *name );
 SV * perl5_get_cv ( const char *name );
-
+void perl5_SvUTF8_on(SV *sv);
+char* perl5_sv_2pvutf8(SV* sv, STRLEN* lp);
 
 #include <HsFFI.h>
 extern SV ** hsPerl5Apply ( HsStablePtr *sub, SV **args, int cxt );

src/Language/Perl5.hs

@@ -62,9 +62,13 @@ module Language.Perl5
 import Control.Concurrent
 import Control.Exception (bracket, throwIO, Exception(..))
 import Control.Monad
-import Data.Dynamic (toDyn)
-import Data.Int
-import Data.List ( intercalate)
+
+import qualified Data.ByteString as BS
+import           Data.Dynamic (toDyn)
+import           Data.Int
+import           Data.List ( intercalate)
+import           Data.Text (Text)
+import qualified Data.Text.Encoding as T
 
 import Foreign hiding (void)
 import Foreign.C.Types
@@ -127,6 +131,21 @@ instance FromSV String where
         cstr <- perl5_SvPV sv
         peekCString cstr
 
+instance ToSV Text where
+  toSV txt = BS.useAsCStringLen (T.encodeUtf8 txt) $ \(cstr, len) -> do
+               sv <- perl5_newSVpvn cstr (toEnum len)
+               perl5_SvUTF8_on sv
+               return sv
+
+instance FromSV Text where
+    fromSV sv =
+          alloca $ \(lenPtr :: Ptr CSize) -> do
+            cStr <- perl5_sv_2pvutf8 sv lenPtr
+            len  <- peek lenPtr
+            -- to do: unsafePackCStringLen is presumably faster
+            T.decodeUtf8  <$> BS.packCStringLen (cStr, fromIntegral len)
+
+
 -- | For convenience, a 'ToSV' instance is provided for 'Int's.
 -- However, it's lossy: actually, a Perl 'SV' will only fit
 -- an 'Int32'.

src/Language/Perl5/Internal.hs

@@ -236,6 +236,27 @@ foreign import ccall "perl5_SvTRUE"
 foreign import ccall "perl5_get_sv"
     perl5_get_sv :: CString -> IO SV
 
+-- |
+-- Turn on the UTF-8 status of an 'SV' (the data is not changed, just the
+-- flag). Do not use frivolously.
+--
+-- source: <https://perldoc.perl.org/perlapi#SvUTF8_on>.
+foreign import ccall "perl5_SvUTF8_on"
+    perl5_SvUTF8_on :: SV -> IO ()
+
+-- |
+-- @
+-- perl5_sv_2pvutf8 sv lp
+-- @
+--
+-- Return a pointer to the UTF-8-encoded representation of the SV, and set
+-- @*lp@ to its length. May cause the SV to be upgraded to UTF-8 as a
+-- side-effect.
+--
+-- source: <https://perldoc.perl.org/perlapi#sv_2pvutf8>
+foreign import ccall "perl5_sv_2pvutf8"
+    perl5_sv_2pvutf8 :: SV -> Ptr CSize -> IO (Ptr CChar)
+
 -- |
 -- @perl5_get_cv name flags@: Uses strlen to get the length of name, then calls
 -- get_cvn_flags.