gcc48: simple program crashes with Segmentation fault: 11

[trac-1948@…]

This worked fine with 2.3.1 on Mavericks.

#/tmp/ramdisk> cat t.cc 

#include <sstream>

int main()
{
   std::ostringstream oss;
   oss << " location ";
   return 0;
}

#/tmp/ramdisk> g++-mp-4.8 t.cc
#/tmp/ramdisk> ./a.out 

Segmentation fault: 11

#/tmp/ramdisk> g++ t.cc
#/tmp/ramdisk> ./a.out 
#/tmp/ramdisk> which g++-mp-4.8

/opt/local/bin/g++-mp-4.8

#/tmp/ramdisk> which g++

/usr/bin/g++

[ryandesign (Ryan Carsten Schmidt)]

Are other versions of MacPorts gcc affected as well, or just gcc48?

[trac-1948@…]

Haven't checked with other versions of GCC. I just noticed that this issue does not occur on my other computer. Which is strange, I've got two Macs, both recently installed from scratch with Yosemite and MacPorts 2.3.2, I'll try to find a difference...

[trac-1948@…]

On one machine Xcode 6.1 was installed from the App Store, on the other it was downloaded from the developer page (but still 6.1 final), can't quite imagine how that would make a difference though.

[trac-1948@…]

Is there a possibility to retro-actively determine in which order the ports were installed? This would not be the first time that it made an important difference.

[Veence (Vincent)]

Does not happen here on my MBA with Yosemite and Xcode 6.1

Air > cat > t.cc
#include <sstream>

int main()
{
   std::ostringstream oss;
   oss << " location ";
   return 0;
}
Air > 
Air > g++-mp-4.8 t.cc
Air > ./a.out 
Air > 

[larryv (Lawrence Velázquez)]

I can’t reproduce this, either. Could you attach a crash log?

[larryv (Lawrence Velázquez)]

Also, what does port installed gcc48 say?

[trac-1948@…]

Ok, it seems to be gcc 4.8.3_0 vs. 4.8.3_1, I'll double check though.

[trac-1948@…]

Yepp, 4.8.3_1 fixed it. Must have been updated in the few hours between installing the first and the second computer... Sorry for the noise.

[larryv (Lawrence Velázquez)]

No worries.