feat: new engine #180
Draft
feat: new engine #180
Conversation
This file contains 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
|
Currently still lacking driver code so I'm way too lazy to add any test (not that having any will motivate me enough xdd) |
|
Also, I hope this will get squashed, I want to erase that embarassing commit message... |
* chore: move to new ESLint config format ESLint is providing a new config format (typically called flat config). The old config format will be deprecated in ESLint 9.0 (currently in rc), and will be removed in ESLint 10. This commit migrate from the old config format at .eslintrc.js to the new flat config one at eslint.config.js. In addition, 3 packages were added: - @types/eslint: A dev dependency used to provide ESLint rules code suggestion. This dependency is used to provide linting through jsdoc for `styleRule` function. It has a very weak consumption (only 1 function relies on it), but it is a small types package. The old config file also relies on it (but does not install it as a npm package). - globals: A dev dependency used to configure ESLint global linting environment. It is consumed by various other package, and already present on the path. Although, some package depend on a very old version of globals (babel), and with how terrible npm symlinking is, ESLint would import version 11.x.x, with a tragic yet funny bug that will break ESLint. - typescript-eslint: A dev dependency that provide first class Typescript integration with ESLint. It provides a convenient config function fully typed to create ESLint flat config array. Additionally, provide the recommended Typescript/ESLint settings that can be conveniently extended. This commit also deleted 2 other dependencies: - @typescript-eslint/eslint-plugin - @typescript-eslint/parser Both these two packages are no longer needed to be explicitly declared as typescript-eslint already import them. And it will make better decision on which version of them it should depend on. chore: ESM support package.json is changed to declare module (ESM), practically mark the project as an ESM first class. Additionally, using ESM module to configure Playwright and ESLint. * chore: change npm lint script With the new config file format, ESLint no longer accepts --ext flag, as the new configuration file is capable of that already. * fix: __dirname on Node 18 --------- Co-authored-by: Hà Trung Kiên <[email protected]>
…lent break dmg (fribbels#248) * feat: initial break dmg calc * feat: allow upload gif * feat: adding gallagher's besotted teammate debuff * feat: replace enemy hp option with max toughness selector * feat: add break elemental type scaling * feat: replace enemyHp logic with individual toggles * feat: implement boothill talent, ruan mei break efficiency buff * fix: tests * feat: update beta timestamps, update beta 2 content * fix: change boothill sort to basic atk
* feat: undo 2.2 incorrect changes * fix: caelus imaginary weights wrong * feat: allow 0x debuffs for pioneer set, update conditional menu * feat: 2.1 light cone descriptions, fix acheron lc debuff name * feat: add warning for char/lc path mismatch, clarified teammate penacony set * fix: ui tweaks
…min stat filter precision (fribbels#254) * feat: replace conditionalstats addition with condensed substats * feat: increase block size to 150,000, switch clone method to spread operator * feat: add 0.000001 to basic stats to fix float filters * feat: save filters to 1000ths precision, add precisionRound to char preview stat hovers * fix: max main stat filter not applied to condensed stats, sort relic set display * fix: delete useless relic fields for optimizer, tweak artist label opacity
- @stylistic/indent: This rule is recommended to be disabled by ESLint author because it aren't very aware about lesser than traditional Typescript elements. In fact, it pops up in the most random places, some places hand-formatted, some conflict with formatter(s). Trivia: it conflicts with both prettier and dprint on indentation. Considering even dprint doesn't have the neccesary configuration to amend it, actually pretty(er) amazing. Alternative: disable the rule inline where it is less than amusing for the rule to kick in. Disabling a styling rule confuses the reader about unneccesary diretives, so it isn't a strong option. - @typescript-eslint/no-unused-vars: Configure the rule to ignore the _XXX pattern for unused arguments and variables. This is the TypeScript "noUnusedLocals" configuration behavior. In fact, _ is considered unnamed/throwaway in a majority of programming languages.
The front API for the new serialization (for structuredClone) is in src/lib/optimizer/new/format. It relies on user supplied serialization and deserialization method to safely serialize/deserialize instances to JSON. By default, if a class is not configured to implement Serializable<SerializedFormat, Class>, Object.create(clazz.prototype) will be used for deserialization. Revisited classes that will be serialized to integrate the new API. test/refactor: Jingliu optimization test suite - Added 2 new optimizer tests for Jingliu, one with 133 SPD limit. - Added unit tests for the new serialization API. - Grouped the test into proper test suites.
|
Personal note for the first commit in the 2 (the eslint one). Since this is just a style change, it could be removed if you don't like it. I didn't expect @stylistic/indent to be that dumb. It pops up at every single nested type conditional (type ternaries). I tried my best to satisfy this rule, but it is way to inconsistent to work with any type of formatter, except if I'm resorting to formatting by hand, or including formatting directives (// prettier-ignore next-line or // dprint-ignore). Formatting directives for indentation feels like a code smell to me. Same to styling directives for ESLint. |
|
Also PR check failed because I checked in package-lock.json on a x86 Windows, which won't contain rollup Linux specific optional package. This is indeed a npm bug. Rollup depends on "some" Rust, which requires a platform-specific build. One work around is to explicitly declare |
added 8 commits
April 4, 2024 00:37
- @stylistic/indent: This rule is recommended to be disabled by ESLint author because it aren't very aware about lesser than traditional Typescript elements. In fact, it pops up in the most random places, some places hand-formatted, some conflict with formatter(s). Trivia: it conflicts with both prettier and dprint on indentation. Considering even dprint doesn't have the neccesary configuration to amend it, actually pretty(er) amazing. Alternative: disable the rule inline where it is less than amusing for the rule to kick in. Disabling a styling rule confuses the reader about unneccesary diretives, so it isn't a strong option. - @typescript-eslint/no-unused-vars: Configure the rule to ignore the _XXX pattern for unused arguments and variables. This is the TypeScript "noUnusedLocals" configuration behavior. In fact, _ is considered unnamed/throwaway in a majority of programming languages.
The front API for the new serialization (for structuredClone) is in src/lib/optimizer/new/format. It relies on user supplied serialization and deserialization method to safely serialize/deserialize instances to JSON. By default, if a class is not configured to implement Serializable<SerializedFormat, Class>, Object.create(clazz.prototype) will be used for deserialization. Revisited classes that will be serialized to integrate the new API. test/refactor: Jingliu optimization test suite - Added 2 new optimizer tests for Jingliu, one with 133 SPD limit. - Added unit tests for the new serialization API. - Grouped the test into proper test suites.
test: move unit test into its own folder (unit-tests). feat: add the progress update feature, allowing main thread to listen for partial optimization completion event. test: added a test that for progress update.
Previously, the calculation logic (which set effects are active,...etc) is in iteration/index.ts, which make it hard to calculate a single shot damage with a specified formula and build. feat: proof of concept implementation of tingyun, dhil and jingliu 3 characters were re implemented with the new optimization system to probe for feedback.
Reference implementation for Boothill, Black Swan and Topaz. fix: stacking step for dhil
Should we rewrite it in Rust?
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.
Pull Request
Implementing the new damage engine that is more flexible and consistent, according to #173
Note
This PR is under src/lib/optimizer/new
Progress