When the error handler is called because a couchbase server is unreachable, our application calls lcb_destroy() to clean up resources. However, this seems to trigger a segmentation fault. Is this a known issue or are we doing something wrong?
I had expected that the C client does not rely on the lcb_t instance to exist or to even point to allocated memory after any of the callbacks is called, because the callbacks may do anything - like calling lcb_destroy or freeing memory.
You can’t destroy the lcb_t from any of the callbacks.