Trinitycore: [3.3.5]Problems with declining names of pets and characters.

[Infernales]

Description:
Problems with declining names of pets and characters with Russian client.
Current behaviour:
When creating a character with a name on Russian or when changing the name of a pet on Russian, a window appears with a choice of declination of the name.
https://i.imgur.com/PGRb4yT.png
When we choose the first option, a window appears with an error declination name change.
https://i.imgur.com/tLSzOFA.png
If you choose the second option, then everything is fine.
https://i.imgur.com/8jzD1tF.png
Expected behaviour:
The first option should work without errors.
Branch(es):
3.3.5
TC rev. hash/commit:
8ecc45c
Operating system:
Linux

[jackpoz]

this sounds like an issue that only contributors with a russian client can fix (just like I ended up fixing utf-8 issues on a sv-se OS)

[illfated]

I am wondering if it is an automatic refusal of using the word "Test" (Russian: "Tect") as a name in the first place. I presume that is why you get suggestions of variations with 1 more letter instead of simply "Test".

[Infernales]

@illfated The problem will be with absolutely any name on Russian. I chose the word "тест" as more understandable. The fact that there is added a single letter is normal, because we are inclined words depending on the context. This is a feature of the Russian language and similar languages, for example Ukrainian. For example, take the name John(Джон):
English: His name is John.
Russian: Его имя Джон.
English: I'm not happy with John.
Russian: Я не доволен Джоном.
English: John has a holiday today.
Russian: У Джона сегодня праздник.
@jackpoz If you need a Russian client, I can provide a torrent file.

[Infernales]

This problem occurred after 309851e. @Treeston, when will you stop breaking what already worked? xD

[Treeston]

I'll stop trying to defuse land mines the moment we run out of land mines waiting to blow up.

[Treeston]

And yeah, that would've started occurring after 309851e until 6a33021, a bunch of things broke in that commit because I fucked up name normalization by accident (no relation to cyrillic specifically).

Can you confirm the issue also exists on 6a33021?

[Infernales]

@Treeston The issue is relevant at commit 537c4d1
There is an assumption that the error is somewhere in the calculation of the number of letters in the declined name.

[Infernales]

@Treeston These functions did not need to be adapted to your changes?
https://github.com/TrinityCore/TrinityCore/blob/3.3.5/src/server/game/Globals/ObjectMgr.cpp#L8728
https://github.com/TrinityCore/TrinityCore/blob/3.3.5/src/server/game/Globals/ObjectMgr.cpp#L8562
https://github.com/TrinityCore/TrinityCore/blob/3.3.5/src/server/game/Globals/ObjectMgr.cpp#L8519

[Infernales]

@Treeston Where after your changes to the bool normalizePlayerName(std::string& name) calculates the length of a name with declined names?
https://github.com/TrinityCore/TrinityCore/blob/3.3.5/src/server/game/Globals/ObjectMgr.h#L888
If I correctly understood the comment on the variable MAX_INTERNAL_PLAYER_NAME, then it was calculated here:
https://github.com/TrinityCore/TrinityCore/commit/309851ea2442457c87f791bb44abfa89e04af1af#diff-1596ecd5424665946209f97259f1dba8L152

[Treeston]

MAX_INTERNAL_PLAYER_NAME was used to determine the size of wstr_buf (in GetMainPartOfName) and nowhere else. It is no longer used after the changes.

[Infernales]

@Treeston Well then, I ran out of ideas about this bug. If you suddenly have solutions to this bug, then I will be happy to test them.

[github-actions[bot]]

eaa22b8a1722fd362bf5ef2c6944b5c50b121afd