Fix UnicodeDecodeError on malformed UTF-8 radio responses

[007hacky007]

Summary

Some Frontier Silicon radios truncate text metadata (e.g. netRemote.play.info.text) at a fixed byte boundary, which can split a multi-byte UTF-8 character in half. This causes a UnicodeDecodeError in __call() at:

doc = ET.fromstring(await result.text(encoding="utf-8"))

This crashes the entire API call, making the device entity unavailable in downstream consumers like Home Assistant.

Root cause

The radio firmware returns text fields with a fixed-size buffer. When the text contains multi-byte UTF-8 characters (common in languages with diacritics - Czech, German, French, etc.), the firmware may cut the buffer mid-character.

For example, a real response from a Hama DIT2010 tuned to a Czech station (Expres FM):

00000120: 2074 7261 706e c3bd 206d 6f64 6572 c3a1   trapn.. moder..
00000130: 746f 7273 6bc3 bd20 7673 7475 7079 2061  torsk.. vstupy a
00000140: 2070 6f64 6f62 6ec3 3c2f 6338 5f61 7272   podobn.</c8_arr

The byte 0xc3 at position 0x14e is the start of a 2-byte UTF-8 sequence (e.g. c3 a9 = e), but the continuation byte is missing - the text was truncated right before the closing </c8_array> XML tag.

This produces:

UnicodeDecodeError: 'utf-8' codec can't decode byte 0xc3 in position 311: invalid continuation byte

Full traceback from Home Assistant:

File "homeassistant/components/frontier_silicon/media_player.py", line 157, in async_update
    info_text = await afsapi.get_play_text()
File "afsapi/api.py", line 414, in get_play_text
    return await self.handle_text(API["text"])
File "afsapi/api.py", line 274, in handle_text
    return unpack_xml(await self.handle_get(item), "value/c8_array")
File "afsapi/api.py", line 255, in handle_get
    return await self.__call(f"GET/{item}")
File "afsapi/api.py", line 220, in __call
    doc = ET.fromstring(await result.text(encoding="utf-8"))
UnicodeDecodeError: 'utf-8' codec can't decode byte 0xc3 in position 311: invalid continuation byte

Fix

Pass errors="replace" to aiohttp's response.text() so that malformed bytes are replaced with the Unicode replacement character (U+FFFD) instead of raising an exception. The text podobn\xc3 becomes podobn\ufffd - a single character is lost but the rest of the response (including the XML structure) is preserved.

Test plan

• Tested against a live Hama DIT2010 radio at a station broadcasting Czech text metadata with diacritics that triggers truncation
• Before fix: get_play_text() raises UnicodeDecodeError, device entity becomes unavailable
• After fix: get_play_text() returns the text with a replacement character, device stays functional