This commit closes #11 github issue.

Part of the github issue #11.

Now, we have the option to output our debug messages
to any FILE stream. We default it to stdout.

It makes sense to have delta-sigma minimum threshold to 0,
meaning to effectively disable this feature.

Part of the github issue #11.

Part of the github issue #11.

Now, we can specify some per-entity (dev_io, dev_mngr)
dynamic and static libraries.

Also, we had to change the libbpmclient compilation
order, as we need the static library to be compiled
before dev_io (now we depend on libbpmclient to configure
our SMIOs)

Part of the github issue #11.

Now, once this SMIO gets spawn by the DEVIO, its
config function will also be called.

Part of the github issue #11.

This is part of the github issue #11.

If the SMIO has registred a config function, then
the DEVIO, on SMIO registration will call this function
to configure the SMIO. Another step to igithub issue #11.

This is the first step towards the github issue #11.
In that we use the liblcient to configure some default
values for us.

Also, we moved some MACROS and use new halutils function.

This is just a trivial patch to better organize the
source files.

The new function can concatenate two null terminated
C strings with a character separator inbetween.

Now, we can choose whether we want libclient
to output debug messages to any FILE stream

Now, we pass the device ID for a DEVIO instance
with the option "-i". We could omit this, as the
device ID is the same as the /dev/fpga<device_number>.

This is useful if we want to specify a different
device ID than the one used in /dev/fpga<device_number>.
Also, the previous version had a hardcoded decvice ID
of 0

The halutils_stringify_* family of functions changed
its interface. So, we update them here.

Now, we have a better approch to initializing
DEVIOs abnd we don't need to emulate some PCIe
signal to dev_mngr.

This fixes #1 github issue.

This functions are used by both dev_io and dev_mngr
to convert to/from LLIO type string and LLIO type
enumeration.

This was supposed to be committed before commits
1381e8c5, c67c9829 and 006465e3.

Part of the fix of github issue #1.

Also, a few wrappers and trivial fixes were made.
Now, on dev_mngr initialization we scan the /dev
directory for our devices (/dev/fpga%d) and fill
the dev_info structures. After this, we can monitor
and spawn all of the DEVIOs that are not initialized.

This is another step in fixing the github issue #1.

This is the first step into making the dev_mngr a real
monitoring process.

This is the first patch towards the solution of
github issue #1

This is preferred as this structure is only for internal
usage

Now, instead of a "custom" error checking we use the
new verison of ASSERT macros (introduced in commit
d1cb7681) which allow us to set the "err" variable
in a more generic way.

This together with commit b487515e fixes #15 github
issue.

Now, instead of a "custom" error checking we use the
new verison of ASSERT macros (introduced in commit
d1cb7681) which allow us to set the "err" variable
in a more generic way.

A few hard limits were imposed to avoid segmentation
faults due to very large integers.

This finishes the quick fix for github issue #21.

As of now, we don't have a way of checking which
channel is tha valid one. So, we hardcode this limit
in software, for now.

This is a quick fix for github issue #21.

This avoid buffering the zmq message and hiding
our debug