net.client: fix tab completion on macOS (libedit parse_and_bind)

`unpythonic.net.client` issued

    readline.parse_and_bind("tab: complete")

unconditionally.  That's the GNU readline dialect.  macOS ships
`readline` backed by `libedit` (not GNU readline), which speaks a
different `parse_and_bind` dialect — the GNU form is silently
ignored there, so macOS users of the REPL client have had no
working tab completion for some time.

Add a `platform.system() == "Darwin"` branch that issues
`readline.parse_and_bind("bind ^I rl_complete")` on macOS instead
— the libedit dialect that actually wires up tab completion.
Mirrors the pattern long-used in `raven.librarian.minichat`, and
newly added in the same session to `mcpyrate.repl.macropython`.

The `unpythonic.net` REPL subsystem remains documented as
POSIX-only (see 2.0.0 CHANGELOG — `ptyproxy` depends on
`termios`, unavailable on Windows).  This commit does *not* make
the client Windows-capable; it just brings `net.client`'s
`parse_and_bind` into consistent shape with the rest of the
fleet, so that when we eventually do the Windows-compat work
for `unpythonic.net`, this one line is already correct.  Also
promotes the previous `# TODO: do we need to call this, PyPy
doesn't support it?` comment to the same definitive form
`# PyPy ignores this, but not needed there.` used in mcpyrate.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Author: Technologicat

CHANGELOG.md

@@ -4,6 +4,7 @@
 
 **Fixed**:
 
+- `unpythonic.net.client`: tab completion now works on macOS. The REPL client issued `readline.parse_and_bind("tab: complete")` unconditionally, but macOS ships `readline` backed by `libedit` (not GNU readline), which speaks a different `parse_and_bind` dialect: the GNU form is silently ignored and tab completion does nothing. Now detects `platform.system() == "Darwin"` and issues `readline.parse_and_bind("bind ^I rl_complete")` on macOS instead. This fix is a prerequisite for eventually supporting `unpythonic.net.client/server` on MS Windows too — the `unpythonic.net` REPL subsystem is still documented as POSIX-only (see 2.0.0), but the `parse_and_bind` branch is now in the right shape for when the rest of the Windows-compat work happens.
 - `unpythonic.misc.timer` and `unpythonic.timeutil.ETAEstimator`: switched the underlying clock from `time.monotonic()` to `time.perf_counter()`. Both are monotonic (guaranteed since Python 3.3), but `perf_counter` is documented as *"a clock with the highest available resolution to measure a short duration"*, whereas `monotonic`'s resolution is implementation-defined. On Windows specifically, `time.monotonic()` is backed by a low-resolution (~16 ms) tick counter, so a `with timer() as t: ...` block that ran in microseconds — such as a PyPy-JIT'd `for _ in range(int(1e6)): pass` — recorded `t.dt` as **exactly 0.0**, silently producing wrong results and, in downstream code that divided by it, a `ZeroDivisionError`. On POSIX the two clocks are usually backed by the same high-resolution source, so this was a latent Windows-only bug. `ETAEstimator` is not affected as a correctness bug at its typical per-task scale (seconds to minutes), but was switched for consistency. We give up `monotonic`'s "comparable across processes" guarantee, which neither class needs — both measure a dynamic extent in wall-clock time within a single process.
 - `unpythonic.test.runner`: module discovery crashed on MS Windows with `re.error: bad escape (end of pattern) at position 0`. The runner used `re.sub(os.path.sep, ...)` to convert a relative path into a dotted module name, which worked by accident on POSIX (where `os.path.sep` is `/`, not a regex metacharacter), but on Windows `os.path.sep` is a lone backslash — an incomplete escape as a regex pattern. Fixed by using `str.replace` instead, which treats both arguments as literal strings. Affects any project that reuses `unpythonic.test.runner` for its own macro-enabled tests on Windows.
 

unpythonic/net/client.py

@@ -32,6 +32,7 @@
 for a remote tab completer, and a separate client-side `input()` loop.)
 """
 
+import platform
 import readline  # noqa: F401, input() uses the readline module if it has been loaded.
 import socket
 import select
@@ -171,7 +172,14 @@ class SessionExit(Exception):
             # Set up remote tab completion, using a custom completer for readline.
             # https://stackoverflow.com/questions/35115208/is-there-any-way-to-combine-readline-rlcompleter-and-interactiveconsole-in-pytho
             readline.set_completer(controller.complete)
-            readline.parse_and_bind("tab: complete")  # TODO: do we need to call this, PyPy doesn't support it?
+            # macOS ships `readline` backed by `libedit`, which speaks a
+            # different `parse_and_bind` dialect than GNU readline.  Detect
+            # by platform to keep tab completion working on Macs.  See
+            # https://stackoverflow.com/questions/7116038/python-repl-tab-completion-on-macos
+            if platform.system() == "Darwin":  # macOS
+                readline.parse_and_bind("bind ^I rl_complete")
+            else:  # Linux, Windows (pyreadline3)
+                readline.parse_and_bind("tab: complete")  # PyPy ignores this, but not needed there.
 
             with socket.socket(socket.AF_INET, socket.SOCK_STREAM) as sock:  # remote REPL session
                 sock.connect((host, repl_port))  # TODO: IPv6 support