I'm using ActiveMQ-CPP library version 3.8.2.
in activemq-cpp-3.8.2/activemq/util/PrimitiveValueNode.h, the symbol
"MAP_TYPE" is used as an enumeration labels.
This symbol is also used in a #define macro in bits/mman-linux.h. This is
one of the header files that is packaged with glibc 2.19-5 (and possibly
older versions as well).
This results in compile errors in some circumstances. If the glibc include
happens to come first, then the #define changes MAP_TYPE into a numerical
literal, causing compile errors when you get to the ActiveMQ-CPP headers.
If it isn't first, then compilation happens, but the MAP_TYPE symbol is
still being clobbered by the macro.
Should I submit this as a bug? It isn't really that ActiveMQ-CPP is doing
anything wrong, but it seems to be a conflict with a ubiquitous library.
in activemq-cpp-3.8.2/activemq/util/PrimitiveValueNode.h, the symbol
"MAP_TYPE" is used as an enumeration labels.
This symbol is also used in a #define macro in bits/mman-linux.h. This is
one of the header files that is packaged with glibc 2.19-5 (and possibly
older versions as well).
This results in compile errors in some circumstances. If the glibc include
happens to come first, then the #define changes MAP_TYPE into a numerical
literal, causing compile errors when you get to the ActiveMQ-CPP headers.
If it isn't first, then compilation happens, but the MAP_TYPE symbol is
still being clobbered by the macro.
Should I submit this as a bug? It isn't really that ActiveMQ-CPP is doing
anything wrong, but it seems to be a conflict with a ubiquitous library.