These functions provide bounds checked access to a region of memory where data is being read or written. They are based on, and similar to, the isc_buffer_ functions in the ISC library.
A buffer is a region of memory, together with a set of related subregions. The used region and the available region are disjoint, and their union is the buffers region. The used region extends from the beginning of the buffer region to the last used byte. The available region extends from one byte greater than the last used byte to the end of the buffers region. The size of the used region can be changed using various buffer commands. Initially, the used region is empty.
The used region is further subdivided into two disjoint regions: the consumed region and the remaining region. The union of these two regions is the used region. The consumed region extends from the beginning of the used region to the byte before the current offset (if any). The remaining region the current pointer to the end of the used region. The size of the consumed region can be changed using various buffer commands. Initially, the consumed region is empty.
The active region is an (optional) subregion of the remaining region. It extends from the current offset to an offset in the remaining region. Initially, the active region is empty. If the current offset advances beyond the chosen offset, the active region will also be empty.
/----- used region -----\\/-- available --\\
| consumed | remaining | |
a bcd e
a == base of buffer.
b == current pointer. Can be anywhere between a and d.
c == active pointer. Meaningful between b and d.
d == used pointer.
e == length of buffer.
a-e == entire length of buffer.
a-d == used region.
a-b == consumed region.
b-d == remaining region.
b-c == optional active region.
lwres_buffer_init() initializes the lwres_buffer_t *b and assocates it with the memory region of size length bytes starting at location base.
lwres_buffer_invalidate() marks the buffer *b as invalid. Invalidating a buffer after use is not required, but makes it possible to catch its possible accidental use.
The functions lwres_buffer_add() and lwres_buffer_subtract() respectively increase and decrease the used space in buffer *b by n bytes. lwres_buffer_add() checks for buffer overflow and lwres_buffer_subtract() checks for underflow. These functions do not allocate or deallocate memory. They just change the value of used.
A buffer is re-initialised by lwres_buffer_clear(). The function sets used , current and active to zero.
lwres_buffer_first makes the consumed region of buffer *p empty by setting current to zero (the start of the buffer).
lwres_buffer_forward() increases the consumed region of buffer *b by n bytes, checking for overflow. Similarly, lwres_buffer_back() decreases buffer bs consumed region by n bytes and checks for underflow.
lwres_buffer_getuint8() reads an unsigned 8-bit integer from *b and returns it. lwres_buffer_putuint8() writes the unsigned 8-bit integer val to buffer *b.
lwres_buffer_getuint16() and lwres_buffer_getuint32() are identical to lwres_buffer_putuint8() except that they respectively read an unsigned 16-bit or 32-bit integer in network byte order from b. Similarly, lwres_buffer_putuint16() and lwres_buffer_putuint32() writes the unsigned 16-bit or 32-bit integer val to buffer b, in network byte order.
Arbitrary amounts of data are read or written from a lightweight resolver buffer with lwres_buffer_getmem() and lwres_buffer_putmem() respectively. lwres_buffer_putmem() copies length bytes of memory at base to b. Conversely, lwres_buffer_getmem() copies length bytes of memory from b to base.