Is the unused gearman_log() function still needed?

[p-alik]

While reviewing `libgearman/log.cc` for #476, noticed that the bare `gearman_log()` function has no call sites anywhere in the repo:

• Defined: `libgearman/log.cc:63`
• Declared: `libgearman/log.hpp:62`

The rest of the file's logging entry points (`gearman_log_info`, `gearman_log_debug`, `gearman_log_error`) are all actively used by callers in `error.cc`, `universal.cc`, `add.cc`, `connection.cc`, and `worker.cc`, but `gearman_log()` itself is only referenced by its own definition/declaration and by doc-comments on the other two functions that point back to it for "argument details."

It's not part of the public `libgearman-1.0` API headers (no `gearman_client_set_log_fn`/`gearman_worker_set_log_fn`-style entry depends on it), so it looks like internal-only dead code — but it's possible it's kept intentionally for symmetry/future use, or used by out-of-tree consumers.

Question: should `gearman_log()` be removed, or is there a reason to keep it around?

[esabol]

Yeah, I don't know. It seems safer to me to leave it alone in case something somewhere uses it, but if it's not documented in the API, it's probably OK to remove. I could go either way on it. Let's hope @SpamapS knows the history of it.