page size 100

[zicjin]

No description provided.

[zicjin]

1 All 6 ACCESS_TOKEN are invalid and used in the wrong way? document

2

// star-history.service.ts
// count step (to get indexes)
const step = NUMBER_OF_SAMPLES / samplesForPage;

It's very confusing. It took me a lot of time to understand the original meaning of this code. It just so happens that the default pagesize is 30, so you use the NUMBER_OF_SAMPLES constant?
In addition, I learned from here that pagesize can be changed, if changed to a larger value, does it mean that ACCESS_TOKEN usage can be reduced?

3 I got this compilation error of grpc in the windows. Unless you downgrade to node v10 or update firebase to the latest version. So I updated all the packages to the latest version using npm-check by the way.

[pnowy]

Can we split this dependencies upgrade? I mean as a separate pull request?

[pnowy]

Ad.1 Yes - the Github deprecated the approach (https://developer.github.com/changes/2019-11-05-deprecated-passwords-and-authorizations-api/#authenticating-using-query-parameters) but I think these tokens should be removed at all (those tokens are not active anymore).

I think the right approach would be to add some login screen to user Github auth in order to retrieve the token (without scopes - they are not needed in this case) or some additional modal to provide access token in order to increase rate limit - otherwise there will be the problem.

Ad.2 The constant NUMBER_OF_SAMPLES was the number of samples for chart (for popular repository there could be a lot of pages/stars so it was enough to get only 30 samples from different pages to get correct chart). But yeah - I can agree that this code can be refactored because is quite confusing and not easy to read.

I would split this changes int smaller separte changes.

Ad. 3. Lets upgrade it as separate PR. Will you do that or should I?

[zicjin]

I built a new pull request.

In fact, I prefer the way auth token works in this project. There are many star-history projects that are authorized by the user OAuth elsewhere. But I think the current way of using Auth token is advantageous and convenient enough for the product. OAuth is always a threshold. If we can collect 20 token, we can have 100,000 requests/per hour , it's enough. And over time, the cache hit ratio of popular repo will become higher and higher.