Shields has long had an in-memory cache with some complex logic for determining when the cached values are used and when they are flushed. At the time this was implemented, squeezing cache performance was helpful since there was no downstream cache. For years now we've used Cloudflare as a CDN, so trying to cache onboard is less useful than before. Furthermore, since the cache is very small and only used in fairly restrictive ways, it probably has very little impact on performance.
There is an internal `makeBadge()` function which is called from a few places in the server and from the public `makeBadge()` function which is a light wrapper. (Eventually we want to dogfood the public API: that's the work of #4950, and this helps with it by aligning the interfaces.)
Related to that is #3370, which is about aligning the `serviceData` schema (i.e. the result of `handle()`) with the public `makeBadge()` function.
A legacy quirk of the _private_ `makeBadge()` function is accepting a `text: ['label', 'message']` array instead of separate `{ label, message }` props like the rest of the codebase. `coalesceBadge()` has to translate from `{ label, message }` to `text: ['label', message']`. This removes that bit of indirection.
It also rewrites most of the tests of `coalesceBadge()` to use `.includes()`, providing IMO a slight improvement in readability.
* Validate input to BadgeFactory.create() (#3875)
* validate input to create()
* remove deprecated properties (#3881)
* remove BadgeFactory class (#3884)
* Template literal templates (#4459)
- Remove use of the doT template library and move to generating SVG output using javascript template literals.
- Drop SVGO and mostly manually implement the optimisations.
- Add a bunch more tests
Co-authored-by: Paul Melnikow <github@paulmelnikow.com>
* drop raster support in package CLI (#4523)
* drop raster support in package CLI
* update docs
* rename gh-badges package to badge-maker
* rename gh-badges dir to badge-maker
* update relative imports and other refs to in parent dir
'gh-badges' --> 'badge-maker'
* update snyk service tests
This change is only tangentially related
We've used the shields repo as an example for these tests so
moving files around in our repo has a knock-on effect on them
* add missing type hints to dev style page
* write the changelog/migration guide for v3
* use extension in README CLI example
* update CLI help
whoops - missed this in #4523
* bump version
* update for self-hosting users
* README updates
* drop .format param from CLI, always output SVG
* Change text[] to label and message, Remove JSON output
- Change text[] to label and message
- Fix message only badge
- Remove JSON output format
- Update the docs
* update package-lock
* rename 'template' to 'style'
* handle invalid styles in coalesceBadge
* ensure makeBadge is passed a string for template in coalesceBadge()
issue #4925
* fix (logo/no label text/label color specified) case
issue #4926
* add example of (logo/no label text/label color specified) to style debug page
* update type defs
* padding fix for FTB style
Co-authored-by: Paul Melnikow <github@paulmelnikow.com>
Use of this feature [has been discouraged for a long time](https://nodejs.org/api/domain.html).
Since most of our code is now bubbling through async, we aren't really getting these "vendor errors" anymore.
Errors that _do_ bubble up through the services have been reported to Sentry since #3706, though they seem to be missing a bunch of their stack traces. Sentry also seems to be combining unrelated internal errors. (https://github.com/badges/shields/issues/3709#issuecomment-514299441) Maybe this will help.
In #2698 we decided to put legacy helper functions in `core/legacy`. I think that’s a fine idea, though if we’re going to have a bunch of badge helper functions in there, it seems like it is probably better to keep these two important but esoteric helper functions with the core code to which they are most coupled. So I added `legacy-` to the name, and put them in `core/base-service`.
