99designs / gqlgen / 28546026211
0%
master: 0%

LAST BUILD BRANCH: gocomputed
DEFAULT BRANCH: master
Ran
Jobs 1
Files 0
Run time 1min
Badge
Embed ▾
README BADGES
x

If you need to use a raster PNG badge, change the '.svg' to '.png' in the link

Markdown

Textile

RDoc

HTML

Rst
coverage: 0.0%. Remained the same
28546026211

Pull #4242

github

StevenACoffman 
feat(federation): per-entity @requires strategies with preloaded batch delivery

Improve federation entity-resolver @requires handling and batch
(@entityResolver(multi: true)) resolution. Backward compatible: with no new
directive arguments or options, existing schemas regenerate byte-for-byte
unchanged.

Per-entity @requires strategy. How an entity resolver receives @requires data is
now selectable per entity via @entityResolver(requires: "..."), overriding a
package default and mirroring how `multi` already resolves per entity. The
directive selects among the three strategies that describe how data reaches the
entity resolver:

  - default:   unmarshaled onto the returned entity, after the resolver runs
  - explicit:  a user Populate<Entity>Requires hook, after the resolver
  - preloaded: unmarshaled onto the resolver INPUT, before the resolver runs

computed stays package-level (computed_requires): it routes @requires to
standalone field resolvers, off the axis the directive selects on, so it is not a
requires: value. A requires: "computed" is rejected with a message pointing at
the package option. (A per-field @goComputed directive is a planned follow-up.)

preloaded strategy. A multi resolver receives its input with @key and @requires
fields populated before it runs, so the whole batch's @requires data is visible
in one scope — enabling a single batched computation (e.g. one ML-inference call)
instead of per-entity work. Implemented via modelgen ExtraFields on the
<Entity>By<Keys>sInput type. Flat scalar/enum @requires only (gqlgen has no
unmarshaler for output object types); such entities use computed.

Per-index errors for batch resolvers. A multi resolver can return a
graphql.BatchErrorList to fail individual entities while the rest resolve; the
runtime nulls each failed entity and reports its error against _entities[index].
Reuses the existing BatchErrors mechanism (fedruntime.SplitEntityBatchErrors), so
the resolver signature is unchanged.

... (continued)
Pull Request #4242: fed batch

0 of 0 relevant lines covered (NaN%)

0.0 hits per line

Jobs
ID Job ID Ran Files Coverage
1 28546026211.1 0
0.0
 GitHub Action Run
Source Files on build 28546026211
Detailed source file information is not available for this build.