Repository navigation
Fix the Rising Kick hit count and the Korean particle tokens in translated text - #66
Merged
Merged
Conversation
$pp:A,B$ and $pp2:A,B$ make the Korean client pick a postposition particle. English has no such thing, but 522 English entries still carried 804 of them, and the client printed them as Korean text. Every affected entry is otherwise fully English, so each token is deleted and nothing is put in its place. Five bare "$pp:" fragments with no particle pair are dropped the same way. Edited in the Weblate JSON and regenerated into Xml/string/en. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The English client prints $pp:A,B$ as Korean text, so verify now lists the English entries that still carry one. It reports instead of failing, like the other data checks, because one stray token from Weblate should not stop the nightly from publishing every other translation. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
|
Warning Review limit reachedYou've used all free OSS reviews for now. Wait for the free limit to reset to keep reviewing this public repository. Next included review available in 37 minutes. View limit details
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. |
Skill 11000031 has three single-hit attack points in all 48 level blocks, but its description says 7 hits from level 2 to 12. The typo is in the Korean source and was carried into every locale. Fixed in the Weblate JSON for en, ko, cn and jp, and directly in the de and pt XML, which have no Weblate source for this file. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
In cn, es, jp, pt and zh_Hant, 44 entries wrap the translator's own word in a Korean particle token, such as $pp2:从$ or $pp2:に$. Each token is replaced by the word it holds, and empty or bare $pp: fragments are deleted. The seven two-choice tokens get から in jp and are dropped in es and pt, where the sentence reads without them. Entries still in untranslated Korean are left for translation. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
AngeloTadeucci
changed the title
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
This fixes two string-data bugs, so they can be tested in game as one XML patch.
Rising Kick tooltip. Skill 11000031 hits 3 times at every level, but its tooltip said 7 hits from level 2 to 12. The typo is in Nexon's Korean source, and every locale copied it. Nexon fixed it later in
kr_latest, which says 3회. This used to be #65.$ppparticle tokens.$pp:A,B$makes the Korean client pick a particle for the noun before it. Outside Korean, the token is wrong:$pp:를,을$, and players saw Korean text. Every one of those strings was already fully English outside the token, so each token is deleted with nothing in its place.$pp2:从$or$pp2:に$. Deleting those would lose the word, so each token is replaced by the word it holds.Changes
WeblateConverter/Json/{en,ko,cn,jp}/stringskilldescription.json: Rising Kick levels 2 to 12, both feature variants where present, 7 changed to 3.Xml/string/{de,pt}/korskilldescription_100.xmlget the same edit directly, because those locales have no Weblate source for the file and the regen leaves them alone.WeblateConverter/Json/en/{koritemdescription,questdescription_final,scriptnpc,scriptquest,stringguide}.json: English tokens deleted. 799 were well-formed pairs. 5 were bare$pp:fragments a translator had left without the pair.WeblateConverter/Json/{cn,es,jp,pt,zh_Hant}/*.json, 11 files: each single-word token becomes its word, and empty$pp:$tokens and bare fragments are dropped. Seven tokens offer two choices: jpから、によってandから、のbecomeから. espara, porandlo, laand ptde, doare dropped, because the sentence reads without them. The two zh_Hant bingo prompts were malformed and are fixed as whole strings, to您是否要使用{0}來檢查數字{1}?and{0}請選擇要檢查的數字。.Xml/string/<locale>/: regenerated from that JSON withjson-to-xml-all.WeblateConverter/Converter/Converter.cs:verifynow prints a[DATA]line listing English keys that still contain$pp. It reports without failing, like the converter's other data checks, so one stray token from Weblate can't stop the nightly from publishing everything else.Verification
Xml/skill/11/11000031.xmlhave three attacks withdamageProperty count="1". After the change, every 11000031 line says 3 in cn, de, en, jp, kr and pt. Each changed line differs frommainby that one character. A second regen leaves the diff byte-identical, so the nightly won't revert thedeandptedits.$pp: none of the 522 changed strings has Hangul or$ppleft. The other$name:...$placeholders are unchanged, and no double spaces or spaces before punctuation were added. Each of the 455 changed XML lines equals its old line with the tokens removed.main's English JSON,verify enprints five[DATA]lines covering 455 keys and exits 0. On this branch it prints none.verify-allpasses.Not checked: any of this in a client. I don't read Chinese, Japanese or Spanish, so the reworded lines in those locales are worth a look from someone who does, especially the es and pt
lead_skillstrings. Their translations were already garbled before this change.Left alone: 479 tokens in de, pt and jp entries that are still untranslated Korean. They need translation, not stripping.
Weblate commits to these JSON files almost daily, and this changes 632 of their lines outside Weblate. Committing Weblate's pending changes before merge, then letting it pull, avoids a merge conflict on its side. Weblate may also flag the en, cn and jp Rising Kick rows as needing edits, because their ko source string changed.
Fixes MS2Community/PrivateMaple2#1570
Fixes MS2Community/PrivateMaple2#938