seg fault on all PG products, OS X trial


I’ve been using the Mac OS X fortran preview (7.0-4) to assess a landscape evolution code compiled as a 64-bit target, and am pretty happy with it so far. When I came back to attempt to debug an issue I’ve been having I got segmentation faults with the executables, then when I tried pgdbg, then pgf95, then pgcc I also immediately got segmentation faults.

Before panicking about some updates that have been applied to my machine, I’d like to find out if this is a licensing issue. An earlier message from Pat Brooks mentioned that this preview would go through the 4th of July, and this started happening the 4th of June. Coincidence? or should I indeed panic that something has gone seriously wrong because of other software?


Hi kc,

Don’t panic, but I don’t think it’s the compilers. Any binary you create is not time bombed so if it worked once, it should work again. Also, I asked Pat who confirmed that the demo license is good until July 4th.

Besides the compiler and your application, do you have any other 64-bit binaries that you can try? Do the 32-bit compilers work? What OSX update did you apply?


Thanks Mat,

I’m running 10.4.9, with the recent security updates as well. However, the difference between Friday of last week and now is a recent update of a particular virtualization package beta which loads kernel extensions, if you know what I mean and I think you do. I’ve just gotten to a point in an instance of that package where I can shut down my windows VM and close the app, which should unload the extensions. I’ll let you know if that works; it would be a very interesting datapoint for that beta program as well.

I can run the same landscape evolution code compiled with g95 (32-bit), and your reply reminded me to try the 32-bit version of pgf95 which works fine, so that would have answered my initial query; so it goes.



It is apparently the other package interfering with the 64-bit version. Everything worked OK once I killed that app, then the segfaults returned when I started it again. Thanks for your quick reply which eliminated that variable (of course this would happen on the day that could have triggered a bug from a typo or the like), even though if I’d thought of 32-bit pgf95 I wouldn’t have had to bug you.