esp: fix I2C writes

[andreacoradi]

Before this small fix, I2C writes would incorrectly repeat the leading byte of the payload while excluding the last one.
By changing the order of the I2C commands, I was able to correctly write to my I2C device.
While consulting the TRM, I also noticed that calling self.update_config(); was unnecessary in that context.

[Grazfather]

Do you understand enough of the esp-idf to have an idea why it was wacky like this? I basically just started with a port without understand fully why it split writes into extra commands

[andreacoradi]

So not with esp-idf, but looking at the esp-hal they just do it like this:

            if cfg!(esp32) && !(start || stop) {
            // NOT US
            } else {
                add_cmd(
                    cmd_iterator,
                    Command::Write {
                        ack_exp: Ack::Ack,
                        ack_check_en: true,
                        length: write_len as u8,
                    },
                )?;
            }

And from my testing it works as intended, so I'm not really sure.
Also, TRM just mentions to create one single write command with length N+1 when starting transmission, and length M if it's in-between.
So I would trust that more.