clintwebb / devplus Goto Github PK

The DpSocket object is not really designed for good blocking when waiting
for data.  In some instances, if there is no data, we should wait for some
data.  To do this, it needs to do a 'select' on the socket handle to
effectively block until data has arrived.

It would also be good if a derivative of DpSocket can have virtual
functions that fire when data arrives, and if all the data has not been
retrieved from the socket, fire it again.  If no data is pulled off the
socket then dont fire again until more data arrives.

Modify the DpDataQueue object to actually store things in chunks so that
when it is pulling data out of the queue when there is a LOT of data in it,
it can be at least somewhat efficient.  The rotating queue that expands
would probably work.  A linked list of queue chunks could be used.  If a
smaller peice is needed than what is in the first chunk, then grab that
peice of data and resize the first chunk.  When inserting data,  make a
copy of the chunk of data being added, and then add that pointer to  the
end of the list.  This should hopefully improve performance a bit when
using the queue for high-speed data communications.

The idea would be to totally reduce the need to resize chunks of memory, as
resizing is more expensive than creating multiple chunks and then using them.

To make it easy to profide information about DevPlus we will provide a wiki
as a homepage for it.  

DevPlus is taken. We are a group of developers / testers of digital
television platforms. You can visit us on http://devplus-wiki.info
We'll really apreciate you change the name of your group.
Thank you,
Digital TB ([email protected])

There is hardly any documentation for developers who wish to use this
library.  

originally the DevPlus tools were in seperate files, but they were merged
into a single pair of header and source files.  Now that the toolchain has
expanded and we now have to specify a larger number of _EXCLUDE defines to
remove the objects we dont want (or cant compile) in each project.  

To make the code a bit more simple, we would split each logical component
into a seperate header and source file.  This will make it a bit more
complicated for compilation scripts because each source file would then
have to be compiled as well.  

To enhance this further, each component could be compiled into an object
file and put in the object archive and imported into a project with the -l
parameter to gcc.