1pa | 1Password Command Line Interface written in Go for Linux | Command Line Interface library

It is well know that you shouldn't use a regex to handle xhtml.

I would use jsoup to remove the whole tag and later add it empty.

But if you want to use a regex, then you can use a regex like this:

The 60 ms you are seeing for a write/read/write/read sequence corresponds pretty well to typical defaults for scheduled transfers. Here's the port configuration for an FTDI (very popular chipset for USB serial ports)

If the serial port initialization code doesn't explicitly set timeouts, what you get is the scheduled transfers every 16ms.

If you instead call SetCommTimeouts, you can arrange for the USB device to forward the received data buffer every time there's a gap on the serial RX wire. To do this, set the ReadIntervalTimeout to the transfer time for just a couple bytes. At 921 kbaud, each byte takes 10-11 microseconds, so the lowest possible timeout of 1 millisecond for ReadIntervalTimeout corresponds to a gap of about 92 bytes.

However, due to the different quality-of-implementation of drivers that come with various USB devices, it's possible to run into devices that don't have hardware support for inter-character timeout. In such cases, it is probably best to disable ReadIntervalTimeout and ReadTotalTimeoutMultiplier, and just use ReadTotalTimeoutConstant.

Using either of these timeout configurations, the data transfer across USB will occur more timely and your code will make progress.

However, the USB latency will still be on the order of 3 milliseconds per receive (1 while the RX line is busy, 1 ms idle to trigger the timeout, and another 1 ms to wait for the next USB timeslot). To do better than 1 message every 3 milliseconds, you need to pipeline so that multiple messages are in-flight, by replacing the stop-and-wait protocol with some sliding window scheme.

Your 220 ohm resistor is probably only specified as accurate to 5%. And all of your analog readings are relative to the Arduino's 5V power supply, which itself is probably only accurate to 5%. So your 8.8% result (not sure how you got 0.85%) is within the expected range, just considering these two possible sources of error.

You are making at least one calculation error: that 0.0049 is just an approximation, the actual voltage conversion factor is 5/1024. That's 3.5% off right there.

It's correct, that format is called hexadecimal for doubles.

The man page says:

For a conversion, the double argument is converted to hexadecimal notation (using the letters abcdef) in the style [-]0xh.hhhp[+-]d [...] the exponent consists of a positive or negative sign followed by a decimal number representing an exponent of 2.

So it's correct that while the mantissa is in hex, the exponent is still decimal.