mirror of
https://github.com/hasura/graphql-engine.git
synced 2024-12-17 12:31:52 +03:00
503c220840
I was trying to figure out how to pipe some information from query execution to the http log recently, and once again stumbled over the mess that is `runGQ`. Here's an attempt to break it apart a little bit. The result should by no means be considered final, but I hope it makes it somewhate easier to understand what's going on in this function. E.g. now it's once again somewhat visible how execution of queries and mutations differs. Had to stop somewhere... The PR is intended to have no functional change. It consists of individual commits which should be "obviously" such. Some thoughts and possible follow-up: - It'd be good to get rid of the ad hoc `Result` data type again eventually, but for the moment I think it's better than the tuples that used to be. - I think we're quite close to reducing the duplication with WebSocket. E.g. executeQueryStep and executeMutationStep might be reusable. - It's tempting to change the caching API slightly, so that the uncached response headers don't have to be pulled from the cache lookup result. PR-URL: https://github.com/hasura/graphql-engine-mono/pull/2669 GitOrigin-RevId: ea414d24194509ce29469d74c62fd060b750488d |
||
|---|---|---|
| .. | ||
| Backends | ||
| Base | ||
| Cache | ||
| Eventing | ||
| GraphQL | ||
| Incremental | ||
| Metadata | ||
| RQL | ||
| Server | ||
| SQL | ||
| App.hs | ||
| EncJSON.hs | ||
| GC.hs | ||
| HTTP.hs | ||
| Incremental.hs | ||
| Logging.hs | ||
| Prelude.hs | ||
| QueryTags.hs | ||
| Session.hs | ||
| Tracing.hs | ||