Experimental Physics and
Industrial Control System

Ralph Lange <[email protected]> · Thu, 19 Aug 2010 08:54:03 -0400

 On 18.08.2010 19:09, Jeff Hill wrote:

Not a problem with delving into the details, it's good to clarify what
the issues are.

Consider this potential scenario using the source code in revision 12099 (your version).
[...]

Aahhhh ... finally got it.

TCP threads already being in the unlocked mid-section of cac::destroyIIU

when the cac dtor starts was completely off my radar.

Thanks for the explanation and your patience,
Ralph