ping [PATCH] Fix a deadlock bug in static variable initialization in libsupc++

[Paolo Carlini]

Benjamin Kosnik wrote:

>Let's get this in folks.
>  
>
I agree, 100%. That said, only to better understand the issue, I'd like 
to see a minimum of discussion about the binary compatibility issues: is 
it guaranteed that it will be possible to link an object build with the 
current headers to another built with the new headers and obtain an 
overall behavior no worse than the current one? In the remote case that 
this is not true, I'm ok anyway with the patch, but that should be clear 
to all the interested people and in case a note added to the 
documentation about that.

Thanks,
Paolo.