This prevents linking problems.
Signed-off-by: Guilherme Maciel Ferreira <guilherme.maciel.ferreira@gmail.com>

Using modifiable char pointers to access read-only strings may
lead to invalid memory access.
Signed-off-by: Guilherme Maciel Ferreira <guilherme.maciel.ferreira@gmail.com>

Use enum LOG_LEVELS instead of int for log and trace levels.

According to ISO/IEC 9899:2011, section 6.7.2.2
(Enumeration specifiers), paragraph 4: "Each enumerated type
shall be compatible with char, a signed integer type, or an
unsigned integer type. The choice of type is implementation-
defined, but shall be capable of representing the values of
all the members". Section J.3.9 also exposes the implementation
defined nature of enumeration signess.

A negative value (-1) is assigned to objects of type enum
LOG_LEVELS. However, the enumeration type might be implemented
as unsigned integer, because there are only positive enumerators
(enumeration members). Thus, this may cause signess mismatches on
some architectures.
Signed-off-by: Guilherme Maciel Ferreira <guilherme.maciel.ferreira@gmail.com>

The functions Log_destToFile() and Log_compareEntries() are used only
by Log_dumpTrace(). The latter is conditionally compiled. Thus, those
functions should also be conditionally compiled.
Signed-off-by: Guilherme Maciel Ferreira <guilherme.maciel.ferreira@gmail.com>

The -fno-common flag indicates the Log_levels is defined in many
places.

First, it's not required to have a global Log_levels, just the
enum type LOG_LEVELS. The problem of define a global variable in
many modules is that the linker may choose the wrong one. This
causes very hard-to-find linker bugs.

Second, the Log_levels is not extern and waste binary space.
Signed-off-by: Guilherme Maciel Ferreira <guilherme.maciel.ferreira@gmail.com>

This change shuts the "suspicious semicolon" warning.
Signed-off-by: Guilherme Maciel Ferreira <guilherme.maciel.ferreira@gmail.com>

Signed-off-by: Guilherme Maciel Ferreira <guilherme.maciel.ferreira@gmail.com>

The one_iteration() function accepts no parameters.
Signed-off-by: Guilherme Maciel Ferreira <guilherme.maciel.ferreira@gmail.com>

The Broker_recordFFDC() function definition is conditionally
compiled. So must be its declaration/prototype.
Signed-off-by: Guilherme Maciel Ferreira <guilherme.maciel.ferreira@gmail.com>

These variables are not used in these modules. Thus, there is no
need for the linker to link these variables here.

The s variable is unused since commit eab1a71d.
And the s variable disapeared in commit 00bdf989.
Signed-off-by: Guilherme Maciel Ferreira <guilherme.maciel.ferreira@gmail.com>

According to ISO/IEC 9899:2011, section 7.22.4.4 (The exit function),
"the constant EXIT_FAILURE is the implementation independent way to
issue failures to exit(3)." Whereas EXIT_SUCCESS is the portable way
to return success.
Signed-off-by: Guilherme Maciel Ferreira <guilherme.maciel.ferreira@gmail.com>

Signed-off-by: Guilherme Maciel Ferreira <guilherme.maciel.ferreira@gmail.com>

According to ISO/IEC 9899:2011, section 6.7.6.3 (Function declarators
including prototypes): "The special case of an unnamed parameter of
type void as the only item in the list specifies that the function
has no parameters."
Signed-off-by: Guilherme Maciel Ferreira <guilherme.maciel.ferreira@gmail.com>

Link with OpenSSL crypto library

Signed-off-by: Otavio Rodolfo Piske <opiske@redhat.com>

Do use the OpenSSL crypto library otherwise linking of SSL versions of the libraries fail due to missing symbols
Signed-off-by: Otavio Rodolfo Piske <opiske@redhat.com>

Proposed fix for OpenSSL compilation issue on OS X

CMake fixes on Microsoft Windows

add missing OPENSSL define to CMakeLists.txt

Signed-off-by: Otavio Rodolfo Piske <opiske@redhat.com>

Signed-off-by: Otavio Rodolfo Piske <opiske@redhat.com>

Switch to cmake build on Travis CI so that it can, more easily, bypass lack of OpenSSL in newer versions of OS X
Signed-off-by: Otavio Rodolfo Piske <opiske@redhat.com>