CBufBase Class Reference

Destructor

Gets a pointer descriptor to represent data from just before the specified data byte backward to the beginning of the contiguous region containing that byte.

Derived classes provide the implementation.

See also: CBufFlat::BackPtr CBufSeg::BackPtr

Compresses the buffer so as to occupy minimal space.

Normally, you would call this when a buffer has reached its final size, or when you know it will not expand again for a while, or when an out-of-memory error has occurred and your program is taking measures to save space. Compression in these circumstances releases memory for other programs to use, but has no adverse effect on performance.

Derived classes provide the implementation.

See also: CBufFlat::Compress CBufSeg::Compress

Deletes data from the buffer.

Derived classes provide the implementation.

See also: CBufFlat::Delete CBufSeg::Delete

Inserts an uninitialised region into the buffer.

Data at and beyond the insertion position is moved to make way for the inserted region. Data before the insertion position remains in place.

Note:

1. The inserted region is not initialised. After using ExpandL(), you should then use a series of Write()s to fill this region with data.

2. Use ExpandL() followed by a series of Write()s when you know the amount of data to be inserted, in advance. It is more efficient than a series of InsertL()s. In addition, once the result of the ExpandL() has been checked, it is guaranteed that the Write()s do not leave, which can sometimes be useful.

Inserts data into the buffer.

Data at and beyond the insertion position is moved to make way for the inserted data. Data before the insertion position remains in place.

Notes:

1. Insertion may require more buffer space to be allocated.

2. In the case of flat buffers, the buffer is extended by a ReAllocL() of the buffer's heap cell, to the smallest multiple of the granularity that will contain the data required. If this reallocation fails, the insertion is impossible and a leave occurs.

3. In the case of segmented buffers, a reallocation is performed if the segment containing the insertion position has insufficient space, and immediately-neighbouring segments cannot be used to contain the new data. As many new segments as are necessary to contain the inserted data are allocated. Each new segment's length is the buffer's granularity. If extension or new allocation fails, a leave occurs.

4. Insertion may also require data to be shuffled. In the case of flat buffers, data beyond the insertion point is shuffled up to create a gap; the new data is then inserted into this gap. In the case of segmented buffers, shuffling is minimised by inserting the new data into newly-allocated buffers, and shuffling only immediately-neighbouring buffers if possible. This may result in some wastage of space, but is much more time-efficient for large amounts of data.

Inserts the specified number of bytes of data from a descriptor into the buffer.

aLength bytes of data from aDes are inserted into the buffer at aPos. Data at and beyond the insertion position is moved to make way for the inserted data. Data before the insertion position remains in place.

Notes:

1. Insertion may require more buffer space to be allocated.

2. In the case of flat buffers, the buffer is extended by a ReAllocL() of the buffer's heap cell, to the smallest multiple of the granularity that will contain the data required. If this reallocation fails, the insertion is impossible and a leave occurs.

3. In the case of segmented buffers, a reallocation is performed if the segment containing the insertion position has insufficient space, and immediately-neighbouring segments cannot be used to contain the new data. As many new segments as are necessary to contain the inserted data are allocated. Each new segment's length is the buffer's granularity. If extension or new allocation fails, a leave occurs.

4. Insertion may also require data to be shuffled. In the case of flat buffers, data beyond the insertion point is shuffled up to create a gap: the new data is then inserted into this gap. In the case of segmented buffers, shuffling is minimised by inserting the new data into newly-allocated buffers, and shuffling only immediately-neighbouring buffers if possible. This may result in some wastage of space, but is much more time-efficient for large amounts of data.

Inserts bytes of data from the specified address into the buffer.

Inserts aLength bytes of data found at address aPtr into the buffer at aPos. Data at and beyond the insertion position is moved to make way for the inserted data. Data before the insertion position remains in place.

Notes:

1. Insertion may require more buffer space to be allocated.

2. In the case of flat buffers, the buffer is extended by a ReAllocL() of the buffer's heap cell, to the smallest multiple of the granularity that will contain the data required. If this reallocation fails, the insertion is impossible and a leave occurs.

2. In the case of segmented buffers, a reallocation is performed if the segment containing the insertion position has insufficient space, and immediately-neighbouring segments cannot be used to contain the new data. As many new segments as are necessary to contain the inserted data are allocated. Each new segment's length is the buffer's granularity. If extension or new allocation fails, a leave occurs.

4. Insertion may also require data to be shuffled. In the case of flat buffers, data beyond the insertion point is shuffled up to create a gap: the new data is then inserted into this gap. In the case of segmented buffers, shuffling is minimised by inserting the new data into newly-allocated buffers, and shuffling only immediately-neighbouring buffers if possible. This may result in some wastage of space, but is much more time-efficient for large amounts of data.

Gets a pointer descriptor to represent the data from the specified position to the end of the contiguous region containing that byte.

Derived classes provide the implementation.

See also: CBufFlat::Ptr CBufSeg::Ptr

Reads data from the buffer into a descriptor.

Data, starting at the specified buffer position is written to the descriptor, filling the descriptor.

Reads the specified number of bytes of data from the buffer into a descriptor.

Reads the specified number of bytes of data from the buffer into a specified address.

Deletes all data in the buffer.

Its behaviour is the same as calling Delete(0,Size()). The buffer is compressed before the function returns.

Re-sizes the buffer to the specified size.

The new size can be larger or smaller than the existing size.

If the new size is larger than the existing size, the buffer is expanded by adding uninitialised data to the end of it.

If the new size is smaller than the existing size, the buffer is reduced; any data at the end of the buffer is lost.

Notes:

1. If the new size is larger than the existing size, the function is equivalent to Delete(aSize,Size()-aSize).

2. If the new size is smaller than the existing size, the function is equivalent to ExpandL((Size(),aSize-Size()).

3. The motivations for using ResizeL() are the same as those for using Delete() and ExpandL().

Gets the number of data bytes in the buffer.

Note that the number of heap bytes used by the buffer may be greater than its size, because there is typically extra room to allow for expansion. Use the Compress() function to reduce the extra allocation as much as possible.

Writes data from a descriptor to the buffer.

The data in the descriptor overwrites the data in the buffer from the insertion point onwards.

No new space is allocated; this function cannot fail (provided the parameters are specified within the bounds of the buffer and descriptor).

No shuffling occurs; new data is written to the memory locations occupied by the data it overwrites.

Writes the specified number of bytes of data from a descriptor to the buffer.

The data in the descriptor overwrites the data in the buffer from the insertion point onwards.

No new space is allocated; this function cannot fail (provided the parameters are specified within the bounds of the buffer and descriptor).

No shuffling occurs; new data is written to the memory locations occupied by the data it overwrites.

Writes the specified number of bytes of data from the specified address to the buffer.

The data in the buffer is overwritten from the insertion point onwards.

No new space is allocated; this function cannot fail (provided the parameters are specified within the bounds of the buffer and descriptor).

No shuffling occurs: new data is written to the memory locations occupied by the data it overwrites.