Skip to content

verified-on.md

Latest commit

 

History

History
81 lines (70 loc) · 3.95 KB

verified-on.md

File metadata and controls

81 lines (70 loc) · 3.95 KB
Raw

Verified on real hardware — 3.0.0

The commands below were run over SSH on the maintainer's Raspberry Pi on 2026-09-18 against develop at commit bdc3ce3, which is tag 3.0.0 plus documentation and one extra line in examples/loopback.php (the loopback OK verdict); no library code changed. Output is pasted verbatim from journalctl of the systemd-run unit that executed /tmp/pi-run.sh as the unprivileged user femus.
Board Raspberry Pi, aarch64, kernel 6.18.34+rpt-rpi-v8
OS Debian GNU/Linux 13 (trixie)
PHP PHP 8.4.24 (cli) (built: Jul 31 2026 05:11:11) (NTS)
Port /dev/serial0 -> ttyS0 (enable_uart=1, serial console disabled)
Wiring jumper between GPIO14 (TXD) and GPIO15 (RXD), pins 8 and 10: everything written comes back
User femus, member of dialout; no sudo needed for the port
Power vcgencmd get_throttled = 0x50005 (under-voltage present during the run; the run is a few seconds long)
Clock the Pi has no NTP sync; the timestamps in the transcript are its local clock, not the real date
Test suite on the Pi OK, but some tests were skipped! Tests: 55, Assertions: 104, Skipped: 5 (the five SocatTest cases skip because socat is not installed on the Pi; they run in CI on Linux and macOS)

Loopback at 9600 and 115200

== 2026-09-17T10:31:22+01:00 commit bdc3ce3
== Linux 6.18.34+rpt-rpi-v8 aarch64; PHP 8.4.24 (cli) (built: Jul 31 2026 05:11:11) (NTS)
== throttled: throttled=0x50005
== port: /dev/serial0 -> ttyS0; user femus groups: femus adm dialout cdrom sudo audio video plugdev games users netdev gpio i2c spi render input
== unit tests
OK, but some tests were skipped!
Tests: 55, Assertions: 104, Skipped: 5.
== php examples/loopback.php /dev/serial0 9600
Opened /dev/serial0 at 9600
readLine: 'ping'
loopback OK: 'ping' went out on TX and came back on RX
timeout: No "\n" received within 0.500 s
readAvailable: ''
[exit 0]
== php examples/loopback.php /dev/serial0 115200
Opened /dev/serial0 at 115200
readLine: 'ping'
loopback OK: 'ping' went out on TX and came back on RX
timeout: No "\n" received within 0.500 s
readAvailable: ''
[exit 0]
== stty -F /dev/serial0 -a (after the last run)
speed 115200 baud; rows 0; columns 0; line = 0;
== at-command against the loopback (echo of AT, then timeout)
< AT
No answer within 2 s
[exit 0]
== done 2026-09-17T10:31:27+01:00

What this shows:

  • new SerialPort('/dev/serial0', 9600) and 115200 open and configure the UART without sudo; stty -a afterwards reports the requested speed.
  • write("ping\n") followed by readLine(timeout: 1.0) returns 'ping' at both speeds.
  • readLine(timeout: 0.5) with nothing arriving throws TimeoutException after the timeout.
  • readAvailable() on an idle port returns '' without waiting.
  • examples/at-command.php against the loopback reads back its own AT line and then reports the 2 s timeout instead of hanging.

Two earlier runs at commit b3c9f6f (tag 3.0.0), one as root and one as femus, produced the same output minus the loopback OK line, which did not exist yet.

Continuous integration

GitHub Actions run 35309138350 on commit b3c9f6f (tag 3.0.0): PHP 8.2, 8.3, 8.4, 8.5 and --prefer-lowest on ubuntu-latest with socat, plus PHP 8.4 on macos-latest with socat; every job green, the five SocatTest pty cases included.

Not yet verified

  • Windows: the mode command string is unit-tested and the device-name normalisation (COM3, COM3:, \\.\COM3) is unit-tested; opening a real COM port, fread behaviour and the write path have not been run on a Windows machine. Planned for 3.0.1 on the maintainer's Windows PC.
  • A real device at the other end (Arduino Nano over USB, ESP8266 AT firmware): pending; the loopback proves the transport, not a conversation.
  • Parity, stop-bit and flow-control settings other than the default 8N1, no flow control: the stty and mode strings are unit-tested, no hardware run.