Fix depth: 0, add depth: false

[idler8]

@ljharb Edit: this PR now adds tests for the current useless behavior of depth 0, and then fixes it. Fixes #261.

//当options.depth为空时,按&符进行分割
assert.deepEqual(qs.parse('ids[0]=1&ids[1]=2&ids[2]=3', { depth: 0 }), { 'ids[0]': 1, 'ids[1]': 2, 'ids[2]': 3 })

测试已经通过
我觉得这可以作为一个补充规则

[ljharb]

Google translate says the title is: "Add a rule for depth==0", and the body:

//When options.depth is empty, split by ampersand
assert.deepEqual(qs.parse('ids[0]=1&ids[1]=2&ids[2]=3', { depth: 0 }), { 'ids[0]': 1, 'ids[1]' : 2, 'ids[2]': 3 })
Test passed
I think this can be used as a supplementary rule.

[ljharb]

The dist folder is generated, so no changes should ever be made in it. Please make the changes inside lib.

Separately, please provide test cases for this new behavior. Is this a bugfix, or a new feature? I can't tell.

[idler8]

hello
fulfill your requirement.
I first used the "pull request" function, not familiar with it.

[idler8]

travis ci构建失败了？
是不是我的错误？
@ljharb

edit: (ljharb): translation added

Did the travis ci build failed?
Is it my mistake?
@ljharb

[idler8]

#246 #261
是否需要这个补充？
@ljharb

edit (ljharb): translation added

#246 #261
Do you need this supplement?
@ljharb

[idler8]

补充了上面的测试
我觉得这是一个新规则

edit (ljharb): translation added

Added the above test
I think this is a new rule.

[ljharb]

I think what we need here is to first add a bunch of tests on master with depth: 0 documenting the existing behavior - and then to rebase this PR on top of it, with your change and the resulting test changes.

[idler8]

I am sorry.
I will close this PR because my English is too bad.
I can't understand what I can do for you next.

[ljharb]

I've updated this into two commits; the first that test the current behavior, and the second that make a depth of 0 or false do something intuitive.

I'm split on the semver of this. I could consider it a bugfix, since 0 was surely never intended to act the way it does; i could consider it semver-minor, because it adds depth: false and because it conceptually allows depth to be 0+ instead of 1+; i could consider it semver-major, because someone might be relying on the super weird and useless current behavior of depth: 0.

[ljharb]

Since I'm adding depth: false, this isn't a patch.

[ljharb]

I'm going to go with this one being semver-minor.

If, in fact, someone has found a use case for the current depth: 0 behavior and this breaks them, I may revert that bugfix in the future - but depth: false will remain regardless.

[idler8]

const Req = {
	defaults: {
		baseUrl: '',
		headers: {},
		qsOptions: { arrayFormat: 'indices', encodeValuesOnly: true },
	},
	base(method, url, data, header, extend) {
		return new Promise((success, fail) => {
			wx.request(Object.assign({ method, header, url, data, success, fail }, extend));
		});
	},
	post(url, data) {
		url = this.defaults.baseUrl + url;
		data = qs.parse(qs.stringify(data, this.defaults.qsOptions), { depth: 0 });
		let header = this.defaults.headers;
		return this.base('post', url, data, header).then(res => {
			if (res.statusCode == 200) {
				if (res.data.code == 200) return res.data.data;
				return Promise.reject(res.data.message);
			}
			return Promise.reject('接口请求失败');
		});
	},
	get(url, data) {
		url = this.defaults.baseUrl + url;
		data = qs.parse(qs.stringify(data, this.defaults.qsOptions), { depth: 0 });
		let header = this.defaults.headers;
		return this.base('get', url, data, header).then(res => {
			if (res.statusCode == 200) {
				if (res.data.code == 200) return res.data.data;
				return Promise.reject(res.data.message);
			}
			return Promise.reject('接口请求失败');
		});
	},
};
Req.get('/kan/article', { ids:[1,2,3,4,5] })

$ids = $request->input('ids',[]);

I use it in WeChat applet development.
https://developers.weixin.qq.com/minigame/dev/api/network/request/wx.request.html
WeChat's data property is very bad, I need to make him and my PHPLaravel interface more suitable.

[ljharb]

@idler8 can you file a new issue with this?