[TF2] Fix to prevent "es_419" from being truncated due to buffer size

[gaelcoral]

Description

I'm not sure if this is actually used anywhere, but I noticed that the rchCleanedCode char buffer was set to 6, which works fine for standard ICUs.
However, for Latin American Spanish (es_419), the string contains 6 characters + '\0', which makes 7, exceeding the buffer limit. As a result, it gets truncated to es_41.

This fix increases the buffer size to 8 to prevent truncation for es_419.

[FlaminSarge]

Looks like the max length can be 11, I know TF2 doesn't use/support any of these but for future proofing may be worth setting this to 12 or 16
https://www.localeplanet.com/icu/

[ficool2]

Fyi this function is not used in any engine/game version

[gaelcoral]

Fyi this function is not used in any engine/game version

Yeah, that's what I thought too but I believe it might be used for localization in the Steam Web API? I'm not entirely sure though.
My main intention was to fix this issue, where TF2 and Dota 2 item names appear in English on Steam when using Latin American Spanish. However, I just noticed that in Dota 2, when calling the GetStoreMetaData API and appending &language=es_41, it correctly retrieves the Latin American Spanish strings. But if you use &language=es_419, it fails and only returns English strings.

So, I suppose this PR could potentially fix that issue (at least for Dota 2).