null8626 / decancer Goto Github PK

See #14 for full comments.

Some portions utilize unsafe code to optimize performance. It would be useful for developers (current and future) if the purpose and invariants of the unsafe code were laid out in a comment to prevent violations.

Not sure why this is the case. Happens on version 1.6.2 in Rust with the error attempt to subtract with overflow, error location being src/lib.rs:95:20. This also occurs with cure_char().

There is no reason for this crate to use unsafe code.

I.e. when fed unicode character “ˑ” (U+02D1) (Modifier Letter Half Triangular Colon):

fn main() {
    println!("{}", decancer::cure("ˑ").as_str());
    println!("This never prints.");
}

Hi there,

I'm experiencing an issue where decancer automatically lowercases all uppercase letters. As far as I understand, this is unintentional behavior (since it is not documented anywhere I looked). If it is intentional, could we have an option that keeps the case of non-violating characters?

Issue Details

• decancer Version: 2.0.2
• macOS Version: 14.3.1
• Chip: Apple M2 Pro
• Architecture: arm64
• Node.js Version: v20.4.0
• Behavior: The package automatically converts all uppercase letters to lowercase.
• Expected Behavior: The case of non-violating characters should be preserved.

Minimal Reproduction

import decancer from "decancer"

console.log(decancer("Test").toString()) // Expected output: "Test", actual output: "test"
console.log(decancer("TeSt").toString()) // Expected output: "TeSt", actual output: "test"

Is your feature request related to a problem?

There currently aren't any good libraries for decancering text in python.

Describe the solution you'd like

I'd like python bindings using a lib such as PyO3.

Describe alternatives you've considered

I tried reinventing the wheel but failed and would love to integrate this perfectly working library into my python projects as well.

Hi @null8626 ,
Thank you for providing this very fast and bright library !

On VSCode, I'm putting some coding rules with TS and Eslint (only on IDE, it's still a JS project), I'm having an error on decancer function

This expression is not callable.
Type 'typeof import("c:/Users/Administrator/WebstormProjects/classified-ads/node_modules/decancer/src/typings")' has no call signatures.

This is already being caught by running clippy locally, which I suggest is added to CI, under the clippy::uninit_assumed_init lint.

Ever since July 2023, i have been thinking about adding back Arabic and Hebrew support for decancer without causing issues because of their right-to-left madness. Then i've found unicode-bidi which rerenders your mixed RTL/LTR text in memory as it were to be rendered by a web browser.

The plan is to somehow implement its algorithm here - an attempt has been in the works since then, but due to school and the complexity of Unicode's bidirectional algorithm, it has been in hiatus for months.

Because of this, development on this library has (publicly) stagnated as the attention has been directed to this enhancement.

I allow users to create simple patterns with asterisks which boils down to equal/startsWith/endsWith/contains. This is currently the only missing feature from adopting this library in production.

• test -> equal
• test* -> startsWith
• *test -> endsWith
• *test* -> contains

I'd really love to see these functions being added <3. Thanks!