Document private functions with @internal tag #2205
Conversation
craicoverflow
force-pushed
the
style/annotate-internal-functions
branch
from
74da857 to
27b8641
Compare
|
I can confirm the same values locally. I just did not add internal flags to |
src/__tests__/starWarsData.js
Outdated
| @@ -122,6 +122,8 @@ function getCharacter(id) { | |||
|
|
|||
| /** | |||
| * Allows us to query for a character's friends. | |||
| * | |||
| * @internal | |||
There was a problem hiding this comment.
@craicoverflow Thanks a lot for PR 👍
One I didn't mention is that __tests__ & __fitures__ are not part of NPM package so all function exported from these folders are private by definition.
So can you please revert all changes under __tests__ folder?
src/jsutils/Path.js
Outdated
| @@ -7,6 +7,8 @@ export type Path = {| | |||
|
|
|||
| /** | |||
| * Given a Path and a key, return a new Path containing the new key. | |||
| * | |||
| * @internal | |||
There was a problem hiding this comment.
Same goes for jsutils folder:
https://github.com/graphql/graphql-js/blob/master/src/jsutils/README.md
These functions are not part of the module interface and are subject to change.
There was a problem hiding this comment.
Sure, I can change that 👍
It's maybe worth noting that in this same file, pathToArray is exported as public in src/index.js (aliased as responsePathAsArray).
Line 287 in 83d5011
There was a problem hiding this comment.
@craicoverflow Good catch 👍 I need to think about it 🤔 💭
craicoverflow
force-pushed
the
style/annotate-internal-functions
branch
from
20c91a7 to
1f835bb
Compare
craicoverflow
force-pushed
the
style/annotate-internal-functions
branch
from
1f835bb to
88b4aa8
Compare
IvanGoncharov
added
the
PR: polish 💅
IvanGoncharov
changed the title
Closes #2183
Manually added an
@internalJSDoc tag to exported functions that are not part of the public API.I began by by making an ESLint rule which lives in the
resourcesdirectory.Although this helped me to identify the exported functions which are not exported in
src/index.js, the rule is not yet stable enough to be included and I may look to follow this up with a separate PR, dealing exclusively with the ESLint rule.