To fix service test that fails in CI (due to no github auth) https://github.com/badges/shields/issues/1359#issuecomment-354184074
- DRY getPhpReleases()
- Pass named options to regularUpdate
- Add json option
- php-version: Move helpers before functions, and move exports to end
- Avoid mutating the inputs
- Declare all the input and output keys
- Avoid recomputing escapeXml on the same values
- Capitalize social badge labels before measuring
Ref: #1379
This takes a naive approach to font-width computation, the most compute-intensive part of rendering badges.
1. Add the widths of the individual characters.
- These widths are measured on startup using PDFKit.
2. For each character pair, add a kerning adjustment
- The difference between the width of each character pair, and the sum of the characters' separate widths.
- These are computed for each character pair on startup using PDFKit.
3. For a string with characters outside the printable ASCII character set, fall back to PDFKit.
This branch averaged 0.041 ms in `makeBadge`, compared to 0.144 ms on master, a speedup of 73%. That was on a test of 10,000 consecutive requests (using the `benchmark-performance.sh` script, now checked in).
The speedup applies to badges containing exclusively printable ASCII characters. It wouldn't be as dramatic on non-ASCII text. Though, we could add some frequently used non-ASCII characters to the cached set.
I have a branch going to automatically generate stats on which services have tests, and make a line chart over time. I'm having a lot of fun with that so I'll keep at it.
Meanwhile, here is my subjective list of critical services, for #1358.
- Periodically log github auth information
- Tokens are hashed which reduces the security risk inherent in the logs
- A consistent hash is used so tokens can be correlated across the three data structures and across the three servers
- Add an admin endpoint for github auth information
- Tokens are returned as-is to enable troubleshooting (e.g. comparing our reqRemaining to github’s)
I rewrote the frontend in React using a module bundler. It's matched feature-for-feature with the current frontend, with only slight changes in the styling. I did not fuss about making the styling identical; the badge popup looks particularly different.
This makes the front end much easier to develop. I'm really looking forward to implementing #701, to which this paves the way.
This makes light use of Next.js, which provides webpack config and dev/build tooling. We’ll probably replace it with create-react-app or our own webpack setup because unfortunately it comes with a lot of runtime overhead (the build is 400k).
Let’s open new issues for bugs and features, and track other follow-ups here: https://github.com/badges/shields/projects/1
- Support single-server testing and a local dev server (like Next) that is on a different port from the shields server
- Refactor config schema
With this change, the suggestions work locally in #1273.
Provide greater consistency for badges related to versions. Fix#1181.
- the `version` function in `color-formatters` was previously returning the colour of the badge, but also its text with a leading _v_. It was broken down into two separate functions and the text formatting part was moved to `text-formatters`, where it really belongs.
- unit tests were added for these two functions in `color-formatters.spec` and `text-formatters.spec`, using Sazerac.
- as discussed in #1181, the leading _v_ was omitted for _xxxx-yy-zz_ date patterns. Any future exceptions can easily be added to the `ignoredVersionPatterns` pattern.
- the badge colour was previously switched to orange if a hyphen was found in the version string. This didn't seem ideal, instead pattern matching is done to find keywords such as `beta`, `alpha` or `snapshot`. Of course, this list can easily be extended.
- all badges related to versions now use the `versionText` and `versionColor` functions. There are a few rare exceptions, for instance in cases where the data returned by the service's API allows to figure things out without relying on any parsing/pattern matching (eg. `badgeData.colorscheme = prerelease ? 'orange' : 'blue';`, where `prerelease` is determined from an API's response).
This fixes a bug introduced in 076cb14, wherein we discarded tokens received
from other servers, and wherein we could save tokens with invalid
identification.
The bug was raised by Paul Melnikow.
Sazerac is a library for data-driven tests, where a series of tests asserts that the return value of a function matches the expected value. It provides nice syntax for tightening this up.
https://hackernoon.com/sazerac-data-driven-testing-for-javascript-e3408ac29d8c
This converts our tests to use it, and replaces some similar home-grown code.
I fixed one bug I encountered along the way: mikec/sazerac#12.
