npx taze major and npx taze minor generate different minor outputs when major version is 0

[runofthemill]

Describe the bug

It looks like taze minor and taze major interpret minor changes differently when the major version is 0; I think this is unexpected behavior, but let me know if it is meant to be this way!

In the repro:

❯ npx taze minor
success Install finished in 5.57s

dependencies are already up-to-date

~/projects/node-ugjrw2 7s
❯ npx taze major
success Install finished in 5.601s


node-starter - 1 minor

  vitest  dev  ~2mo  ^0.23.4  →  ^0.25.3  ~6d  


Add -w to write package.json

Reproduction

https://stackblitz.com/edit/node-ugjrw2?file=package.json&view=editor

System Info

❯ npx envinfo --system --binaries --browsers
success Install finished in 6.376s


  System:
    OS: Linux 5.0 undefined
    CPU: (8) x64 Intel(R) Core(TM) i9-9880H CPU @ 2.30GHz
    Memory: 0 Bytes / 0 Bytes
    Shell: 1.0 - /bin/jsh
  Binaries:
    Node: 16.14.2 - /usr/local/bin/node
    Yarn: 1.22.19 - /usr/local/bin/yarn
    npm: 7.17.0 - /usr/local/bin/npm

Used Package Manager

npm

Validations

• Follow our Code of Conduct
• Read the Contributing Guide.
• Check that there isn't already an issue that reports the same bug to avoid creating a duplicate.
• Check that this is a concrete bug. For Q&A, please open a GitHub Discussion instead.
• The provided reproduction is a minimal reproducible of the bug.

[ntnyq]

Major version equal to zero is an exception

[runofthemill]

@ntnyq yep so perhaps the behavior is expected, but I would expect this:

❯ npx taze minor
success Install finished in 5.57


node-starter - 1 minor

  vitest  dev  ~2mo  ^0.23.4  →  ^0.25.3  ~6d  


Add -w to write package.json

~/projects/node-ugjrw2 7s
❯ npx taze major
success Install finished in 5.601s


node-starter - 1 minor

  vitest  dev  ~2mo  ^0.23.4  →  ^0.25.3  ~6d  


Add -w to write package.json

or this:

❯ npx taze minor
success Install finished in 5.57s

dependencies are already up-to-date

~/projects/node-ugjrw2 7s
❯ npx taze major
success Install finished in 5.601s

dependencies are already up-to-date

But instead there's nothing in taze minor and a minor update visible in taze major, which appears (to me) to be inconsistent, major version 0 or not.

[antfu]

The behavior is expected. Only the output for counters needs some improvements. PR welcome

[runofthemill]

@antfu which outputs need improvement? I might be able to work on that :)

could you help me understand why the behavior is expected? it seems counterintuitive that a minor update would show up for taze major but not taze minor. unless vitest ^0.23.4 → ^0.27.1 is being misclassified as minor?

[antfu]

❯ npx taze major

- node-starter - 1 minor
+ node-starter - 1 major

  vitest  dev  ~2mo  ^0.23.4  →  ^0.25.3  ~6d  

The counter is wrong. But it only shows up with taze major so I consider it expected.

[runofthemill]

So seems like I'm entering a rabbit hole 😅

A minimal reproduction classifies 0.23.4 → 0.25.3 as a minor change:

const semver = require('semver'); // "semver": "^7.3.8" in package.json

console.log(`${semver.diff('0.23.4', '0.25.3')}`); // outputs "minor"

What do you think about a minimal change like this?

// src/utils/versions.ts
export function changeVersionRange(version: string, mode: Exclude<RangeMode, 'latest' | 'newest'>) {
  if (!semver.validRange(version))
    return null

  if (mode === 'default')
    return version

  const min = semver.minVersion(version)
  if (!min)
    return null

+  if (semver.major(min) === 0 && mode === 'minor')
+   return `>=${min} <1.x`

  return {
    major: '>=',
    minor: '^',
    patch: '~',
  }[mode] + min
}

[antfu]

Isn't your change making it the opposite? 0.x change should be considered as major. I feel it's something wrong with semver.diff, maybe we should patch it instead?