Fix potential issue with nested @defer in non-deferrable case
#2312
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
👷 Deploy request for apollo-federation-docs pending review.Visit the deploys page to approve it
|
|
This pull request is automatically built and testable in CodeSandbox. To see build info of the built libraries, click here or the icon next to each commit SHA. |
clenfest
approved these changes
Jan 5, 2023
added 2 commits
January 6, 2023 16:35
The code handling `@defer` works roughly like this: 1. as we cross a query element (field or spread) with a `@defer`, we remember that we just "entered" a deferred section. 2. as we handle the initial elements within that `@defer`, we notice that we just entered a `@defer`, ignore in which subgraph we were, and instead attempt to "re-enter" subgraphs using a key. 3. if that work, we continue from whichever possible subgraphs we just re-entered. 4. Otherwise, if we can't handle the new element by re-entering a subgraph, then we fall back to "normal" operation and try to handle the element from the original subgraph in which we were before entering the @defer: this correspond to case where we cannot do a router-based-defer due to not having proper @key to do it. The issue however is that we were trying to re-enter subgraphs even when the next element was a fragment. This meant that we re-entered any subgraphs that had a key, but that may exclude the one subgraph in which we previously were. If we get a new `@defer` followed by a field that was only in the "original" subgraph, then we could end up not being able to find that field _even_ when falling back to step 4 above, because we had already re-entered different subgraphs due to the 1st `@defer`. The solution implemented by this commit consists in delaying the subgraph re-entry until we actually get to a field, since this is where we _have_ to re-enter.
pcmanus
force-pushed
the
defer-bug-named-fragments
branch
from
612826c
to
c791bf7
Compare
pcmanus
pushed a commit
that referenced
this pull request
Jan 9, 2023
The code handling `@defer` works roughly like this: 1. as we cross a query element (field or spread) with a `@defer`, we remember that we just "entered" a deferred section. 2. as we handle the initial elements within that `@defer`, we notice that we just entered a `@defer`, ignore in which subgraph we were, and instead attempt to "re-enter" subgraphs using a key. 3. if that work, we continue from whichever possible subgraphs we just re-entered. 4. Otherwise, if we can't handle the new element by re-entering a subgraph, then we fall back to "normal" operation and try to handle the element from the original subgraph in which we were before entering the @defer: this correspond to case where we cannot do a router-based-defer due to not having proper @key to do it. The issue however is that we were trying to re-enter subgraphs even when the next element was a fragment. This meant that we re-entered any subgraphs that had a key, but that may exclude the one subgraph in which we previously were. If we get a new `@defer` followed by a field that was only in the "original" subgraph, then we could end up not being able to find that field _even_ when falling back to step 4 above, because we had already re-entered different subgraphs due to the 1st `@defer`. The solution implemented by this commit consists in delaying the subgraph re-entry until we actually get to a field, since this is where we _have_ to re-enter.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
The code handling
@deferworks roughly like this:@defer, we remember that we just "entered" a deferred section.@defer, we notice that we just entered a@defer, ignore in which subgraph we were, and instead attempt to "re-enter" subgraphs using a key.The issue however is that we were trying to re-enter subgraphs even when the next element was a fragment. This meant that we re-entered any subgraphs that had a key, but that may exclude the one subgraph in which we previously were. If we get a new
@deferfollowed by a field that was only in the "original" subgraph, then we could end up not being able to find that field even when falling back to step 4 above, because we had already re-entered different subgraphs due to the 1st@defer.The solution implemented by this commit consists in delaying the subgraph re-entry until we actually get to a field, since this is where we have to re-enter.