Ulrich Eckhardt wrote:
Unnecessary bloat. Maybe, when someone calls the code, crashes.
Bloat yes, crash no. None can call it in a way, it would crash. Since the m_sys_err cannot be non-zero. Only if a user code calls throw lock_error(...) , which user code isn't expected to, then trying to access the (undocumented internal) function message this will ever be possible.
I'm a bit frustrated, please don't take it as a personal attack.
No need to. The patch finally is applied. Some words though. 1) If you had pointed me to the old posting where I could have had seen, that the previous maintainer (Michael Glassford) already had decided to apply this patch, I would have had saved time, and just applied. 2) You, and every other have the right to expect from me, that I will not blindly apply patches without review. (And this always costs time, which I not always have.) 3) I want to thank you for your patience, and invite you to further contributions. Roland