[vbox-dev] Patch to resolve the unicode-composite character problem of the shared-folder in Mac OS X

[Sander van Leeuwen]

Hi,

Thanks a lot for your patch! Could you please state that you either 
license this code under the MIT license
or our custom license? (see 
http://www.virtualbox.org/wiki/Contributor_information)

Thanks,

Sander

Gi youl Kim wrote:
> Mac OS X deals with filename using unicode, especially
> NFD(Normalization Form D - Canonical Decomposition). Here is link
> about the topic.
>
> http://unicode.org/reports/tr15/
>
> Mac OS X understands NFD and NFC(Normalization Form C - Canonical
> Decomposition, followed by Canonical Composition), but uses NFD as
> filename. Windows XP understands NFC but not NFD.
>
> So, when you tries to access the files in the Mac OS X (host) via
> Shared Folder from Windows XP (guest), and if the filename is Korean,
> Japanese or any language which has form in NFC & NFC, The filename
> does not appear correctly. Windows Vista understands NFD, so it won't
> be matter with Vista (guest).
>
> To resolve this problem, you should normalize the filename to NFC(from
> Mac OS X to Windows XP) or NFD(from Windows XP to Mac OS X).
>
> Following patch does that. Tested with VB 1.6.2 and revision 9784
>
> Index: src/VBox/HostServices/SharedFolders/vbsf.cpp
> ===================================================================
> --- src/VBox/HostServices/SharedFolders/vbsf.cpp	(revision 9784)
> +++ src/VBox/HostServices/SharedFolders/vbsf.cpp	(working copy)
> @@ -33,6 +33,10 @@
>  #include <iprt/string.h>
>  #include <iprt/uni.h>
>
> +#ifdef RT_OS_DARWIN
> +#include <Carbon/Carbon.h>
> +#endif
> +
>  #undef LogFlow
>  #define LogFlow Log
>
> @@ -231,6 +235,36 @@
>      }
>      else
>      {
> +#ifdef RT_OS_DARWIN
> +		SHFLSTRING *pPathParameter = pPath;
> +		size_t cbPathLength;
> +		CFMutableStringRef inStr = ::CFStringCreateMutable(NULL, 0);
> +		uint16_t ucs2Length;
> +		CFRange rangeCharacters;
> +		
> +		// Is 8 times length enough for decomposed in worst case...?
> +		cbPathLength = sizeof(SHFLSTRING) + pPathParameter->u16Length * 8 + 2;
> +		pPath = (SHFLSTRING *)RTMemAllocZ (cbPathLength);
> +		if (!pPath)
> +		{
> +			rc = VERR_NO_MEMORY;
> +			Log(("RTMemAllocZ %x failed!!\n", cbPathLength));
> +			return rc;
> +		}
> +		
> +		::CFStringAppendCharacters(inStr, (UniChar
> *)pPathParameter->String.ucs2, pPathParameter->u16Length /
> sizeof(pPathParameter->String.ucs2[0]));
> +		::CFStringNormalize(inStr, kCFStringNormalizationFormD);
> +		ucs2Length = ::CFStringGetLength(inStr);
> +		
> +		rangeCharacters.location = 0;
> +		rangeCharacters.length = ucs2Length;
> +		::CFStringGetCharacters(inStr, rangeCharacters, pPath->String.ucs2);
> +		pPath->String.ucs2[ucs2Length] = 0x0000; // NULL terminated
> +		pPath->u16Length = ucs2Length * sizeof(pPath->String.ucs2[0]);
> +		pPath->u16Size = pPath->u16Length + sizeof(pPath->String.ucs2[0]);
> +		
> +		CFRelease(inStr);
> +#endif
>          /* Client sends us UCS2, so convert it to UTF8. */
>          Log(("Root %ls path %.*ls\n", pwszRoot,
> pPath->u16Length/sizeof(pPath->String.ucs2[0]), pPath->String.ucs2));
>
> @@ -255,6 +289,10 @@
>              if (VBOX_FAILURE(rc))
>              {
>                  AssertFailed();
> +#ifdef RT_OS_DARWIN
> +				RTMemFree(pPath);
> +				pPath = pPathParameter;
> +#endif
>                  return rc;
>              }
>
> @@ -312,6 +350,10 @@
>              /* Nul terminate the string */
>              *dst = 0;
>          }
> +#ifdef RT_OS_DARWIN
> +		RTMemFree(pPath);
> +		pPath = pPathParameter;
> +#endif
>      }
>
>      if (VBOX_SUCCESS (rc))
> @@ -1340,6 +1382,28 @@
>              int rc2 = RTStrToUtf16Ex(pDirEntry->szName, RTSTR_MAX,
> &pwszString, pDirEntry->cbName+1, NULL);
>              AssertRC(rc2);
>
> +#ifdef RT_OS_DARWIN
> +			{
> +				// Convert to
> +				// Normalization Form C (composed Unicode). We need this because
> +				// Mac OS X file system uses NFD (Normalization Form D :
> decomposed Unicode)
> +				// while most other OS', server-side programs usually expect NFC.
> +				uint16_t ucs2Length;
> +				CFRange rangeCharacters;
> +				CFMutableStringRef inStr = ::CFStringCreateMutable(NULL, 0);
> +				
> +				::CFStringAppendCharacters(inStr, (UniChar *)pwszString,
> RTUtf16Len (pwszString));
> +				::CFStringNormalize(inStr, kCFStringNormalizationFormC);
> +				ucs2Length = ::CFStringGetLength(inStr);
> +				
> +				rangeCharacters.location = 0;
> +				rangeCharacters.length = ucs2Length;
> +				::CFStringGetCharacters(inStr, rangeCharacters, pwszString);
> +				pwszString[ucs2Length] = 0x0000; // NULL terminated
> +				
> +				CFRelease(inStr);
> +			}
> +#endif
>              pSFDEntry->name.u16Length = RTUtf16Len
> (pSFDEntry->name.String.ucs2) * 2;
>              pSFDEntry->name.u16Size = pSFDEntry->name.u16Length + 2;
>
> Index: src/VBox/HostServices/SharedFolders/Makefile.kmk
> ===================================================================
> --- src/VBox/HostServices/SharedFolders/Makefile.kmk	(revision 9784)
> +++ src/VBox/HostServices/SharedFolders/Makefile.kmk	(working copy)
> @@ -36,6 +36,9 @@
>  VBoxSharedFolders_INCS.win  = \
>  	$(PATH_TOOL_$(VBOX_VCC_TOOL)_ATLMFC_INC) \
>  	$(VBOX_PATH_SDK)
> +	
> +VBoxSharedFolders_LDFLAGS.darwin = \
> +	-framework Carbon
>
>  VBoxSharedFolders_SOURCES = \
>  	service.cpp \
>
> _______________________________________________
> vbox-dev mailing list
> vbox-dev at virtualbox.org
> http://vbox.innotek.de/mailman/listinfo/vbox-dev
>   


-- 
Kind regards / Mit freundlichen Gruessen / Met vriendelijke groet

--

Sun Microsystems GmbH        Sander van Leeuwen
Werkstrasse 24               Senior Staff Engineer, VirtualBox
71384 Weinstadt, Germany     mailto:Sander.Vanleeuwen at sun.com


================================================
Sitz der Gesellschaft: Sun Microsystems GmbH,
Sonnenallee 1, 85551 Kirchheim-Heimstetten
Amtsgericht Muenchen: HRB 161028
Geschaeftsfuehrer: Thomas Schroeder,
Wolfgang Engels, Dr. Roland Boehmer
Vorsitzender des Aufsichtsrates: Martin Haering
================================================