(Edited out!)
I guess catastrophic corruption does occur after all, from an asynchronous crash...
I am curious about the library's resilience against database corruption, such as...
Excellent, thank you! Just to give a little context, I have benchmarked scenarios...
Excessive syscalls with tiny read() calls
strncmp() should be replaced by memcmp()
dbh_set_data() not updating data when growing record size