Thanks alot Alan and Andrew for your help!!
However, it seems one patch has slipped. It is about "int8_t". PROTON-1324: Interpretation of "int8_t" on Solaris using SunStudio is different from GCC one https://issues.apache.org/jira/browse/PROTON-1324 Regards, Adel ________________________________ From: Alan Conway <acon...@redhat.com> Sent: Thursday, October 20, 2016 2:27:10 PM To: Adel Boutros; Andrew Stitcher; dev Cc: rabih.prom...@gmail.com Subject: Re: [Proton-c][Solaris][SunStudio-2.1] Compiling Qpid Proton-c and C++ bindings on Solaris On Thu, 2016-10-20 at 07:27 +0000, Adel Boutros wrote: > > >>> Symbol hiding: PROTON-1316 removes symbol hiding for gcc > completely, so that change can't go in as is. I'm sure that wasn't > intentional, just a side effect of doing what was necessary for > Solaris without checking properly again on Linux. > > This is not true, the patch is protected by "if" which will only > execute the code for SunStudio compiler and will not affect GCC in > any way (elseif (${CMAKE_CXX_COMPILER_ID} STREQUAL "SunPro" in > CMakeLists.txt and elif defined(__SUNPRO_C) || defined(__SUNPRO_CC) > in code). > The original patch did delete the HIDE lines in the GCC section - I'm sure this was an accident, I restored them as intended. > >>> This change doesn't quite address my concerns: For SunPro the > ENABLE_HIDE_UNEXPORTED_SYMBOLS toggle is not honoured. > > I apologise for the error in the CMakeLists.txt. We missed the "if > (ENABLE_HIDE_UNEXPORTED_SYMBOLS) " for the code submitted. > @Alan > Will you handle updating the patch here? Done. I made the minimal change to respect the ENABLE_HIDE... flag for solaris, I haven't re-factored the flag-setting section. I've committed the patches, any further improvements can be made directly on master. Cheers, Alan.
