handle not closed during nvs error paths

[mxr]

Example:

bitclock/bitclock-fw/main/libs/nvs.c

Lines 99 to 102 in c06e1b5

	err = nvs_set_u8(handle, NVS_ID_TEMP_UNIT, new_temp_unit);
	if (err != ESP_OK) {
	return err;
	}

Not sure the implications of this.

It's typical to solve this with a goto to centralize cleanup as outlined here.

[babldev]

Random, but probably a bug lol. Did you find this with AI

[mxr]

Just I

[babldev]

Not sure the implications of this.

If the write of this setting to NVS (Flash) fails the handle would be left open. In practice this probably extremely rare and not a big issue but you're proposing the right fix.