#include <vbento.h>

Inheritance diagram for VBentoChar:

Public Member Functions
	VBentoChar ()
	Constructs with uninitialized name and a space char.
	VBentoChar (VBinaryIOStream &stream)
	Constructs by reading from stream.
	VBentoChar (const VString &name, const VCodePoint &c)
	Constructs from supplied name and value.
virtual	~VBentoChar ()
	Destructor.
virtual VBentoAttribute *	clone () const
VBentoChar &	operator= (const VBentoChar &rhs)
virtual void	getValueAsXMLText (VString &s) const
	Returns a string suitable for an XML attribute value, including escaping via _escapeXMLValue() if needed.
virtual void	getValueAsString (VString &s) const
	Returns a printable form of the attribute value.
virtual void	getValueAsBentoTextString (VString &s) const
	Returns a Bento Text form of the attribute value.
const VCodePoint &	getValue () const
	Returns the attribute's value.
void	setValue (const VCodePoint &c)
	Sets the attribute's value.
Static Public Member Functions
static const VString &	LEGACY_DATA_TYPE_ID ()
	The data type name / class ID string.
static const VString &	DATA_TYPE_ID ()
	The data type name / class ID string.
static VBentoChar *	newFromLegacyCharStream (VBinaryIOStream &stream)
	Constructs by reading 1 byte and using it as a Unicode code point value.
Protected Member Functions
virtual Vs64	getDataLength () const
	Returns the length of this object's raw data only.
virtual void	writeDataToBinaryStream (VBinaryIOStream &stream) const
	Writes the object's raw data only to a binary stream.

Detailed Description

In Vault versions up to 4.0:

The type ID is 'char'.
The internal storage is a VChar object, which is a thin wrapper for a C char.
In binary form, the stream data was always 1 byte corresponding to the C char.
In text form, the text data was the char surrounded by single quotes, formatted using c. In Vault version 4.1:
The type ID is 'u8ch'.
The internal storage is a VCodePoint object, which stores a Unicode code point.
In binary form, the data is 1 to 4 bytes of UTF-8 data. This is more efficient than always using a full 32-bit value, and avoids endianness concerns.
In text form, the data is the same 1 to 4 bytes of UTF-8 data.
When reading, if we encounter the legacy type ID ('char'), we read the single byte as before, and form the VCodePoint from the byte value. (This is actually a way to ensure old data containing high-bit non-ASCII values are properly migrated to proper Unicode.)
When writing, we always write out the new type. (Is it worth considering a compile-time switch to write out the old data type?) (A runtime switch is trickier because it might need to be per-object; a global runtime switch could have threading issues.) (A possible approach would be to write out the legacy type for ASCII values, or write both types.)

Definition at line 1216 of file vbento.h.

Constructor & Destructor Documentation

VBentoChar::VBentoChar ( VBinaryIOStream & stream ) [inline]

Constructs by reading from stream.

Parameters:

stream the stream to read

Definition at line 1225 of file vbento.h.

const VCodePoint& VBentoChar::getValue ( ) const [inline]

Returns the attribute's value.

Definition at line 1236 of file vbento.h.

void VBentoChar::setValue ( const VCodePoint & c ) [inline]

Sets the attribute's value.

Parameters:

c	the attribute value

Definition at line 1237 of file vbento.h.

virtual Vs64 VBentoChar::getDataLength ( ) const [inline, protected, virtual]

Returns the length of this object's raw data only.

Definition at line 1241 of file vbento.h.

virtual void VBentoChar::writeDataToBinaryStream ( VBinaryIOStream & stream ) const [inline, protected, virtual]

Writes the object's raw data only to a binary stream.

Parameters:

stream the stream to write to

Definition at line 1242 of file vbento.h.

The documentation for this class was generated from the following files: