feat(sdk/java): add @Internal annotation for fields #3
+3,743
−1,704
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
eunomie
force-pushed
the
sdk-java-internal-field
branch
from
8a1e9be to
7376df7
Compare
eunomie
force-pushed
the
sdk-java-dag
branch
3 times, most recently
from
fd71751 to
771e168
Compare
Not a simple revert commit, since we've started using new features :( Signed-off-by: Justin Chadwell <[email protected]>
…agger#9774) * cache result wrapping and context cancel management in dagql cache This updates dagql's cache to use the same "singleflight" implementation added for the local sync rework. The idea of deduping equivalent calls is the same as the previous cachemap implementation, but with two important additional features: * Context cancellation is handled better. * Previously when calls were deduped, the "first" context was always used, which meant that if it was canceled (i.e. first client disconnected for any reason), all other clients waiting on the same call would get canceled too. * Now, the call is only canceled if *all* caller contexts are canceled. * Results returned from the cache are wrapped in a type that are ref-counted and releasable. When all refs for the result are released, they are removed from the cache. * This is not used in dagql quite yet, but will be in follow-up PRs. It still makes sense to include in this PR since it comes along for the ride with re-using the singleflight implementation from the local sync code. The main goals here are to: * Test the context cancelation improvements in isolation; they are a subtle change so worth verifying by themselves * Setup the follow-ups that actually use the ref counted results to support engine-wide dagql cache Signed-off-by: Erik Sipsma <[email protected]> * supplement unit tests w/ integ test Signed-off-by: Erik Sipsma <[email protected]> --------- Signed-off-by: Erik Sipsma <[email protected]>
Signed-off-by: Solomon Hykes <[email protected]>
Signed-off-by: Jeremy Adams <[email protected]>
* chore(sdk/java): global Dagger client on Dagger.dag() AbstractModule is not required anymore and is just a (kind of useless) wrapper on top of Dagger.dag() The default template is not using it anymore. The best way to benefit from this global client is to import static io.dagger.client.Dagger.dag and to use `dag()` instead the previous `dag` field * remove dag variables and client in constructor * remove dag and simplify Json serialize/deserialize * update integration tests to reflect new changes * remove AbstractModule Java SDK is experimental, this is a breaking change, but we will move forward with it * update docs accordingly Signed-off-by: Yves Brissaud <[email protected]>
eunomie
force-pushed
the
sdk-java-internal-field
branch
2 times, most recently
from
d881a3a to
0f70a8f
Compare
* fix: don't set SSH_ASKPASS to echo
I don't really have context for why this was added, but it's definitely
not right, resulting in *always* setting a completely incorrect auth
token.
See the following:
$ GIT_TERMINAL_PROMPT=0 SSH_ASKPASS=echo git credential fill <<EOF
protocol=https
host=github.com
EOF
protocol=https
host=github.com
username=Username for 'https://github.com':
password=Password for 'https://Username%20for%20%27https%3A%2F%2Fgithub.com%27%3A%[email protected]':
This outputs the prompts into the username/password fields, instead of
*correctly* erroring out, and not attaching this as an authentication
token.
I'm not actually sure if this was causing user-facing bugs, but I hit
this while working on a total refactor of the git implementation.
Signed-off-by: Justin Chadwell <[email protected]>
* only set SSH_ASKPASS when no http/https; handle empty password
Signed-off-by: Erik Sipsma <[email protected]>
---------
Signed-off-by: Justin Chadwell <[email protected]>
Signed-off-by: Erik Sipsma <[email protected]>
Co-authored-by: Erik Sipsma <[email protected]>
Update `generateClient` to only set fields in the configuration so now the generation logic happens in `generatedContextDirectory`. That way, it works with `dagger develop`. Also add persistency in `dagger.json` for generated clients. Correctly use contextDirectory in codegen instead of only the source directory. Signed-off-by: Tom Chauveau <[email protected]>
…ger#9800) Wolfi made a change that causes glibc triggers to require busybox triggers having already run: wolfi-dev/os@8229c03 This isn't obviously wrong, but our current alpine module likely runs triggers in a slightly different order and we thus end up with a glibc trigger running, trying to execute /usr/bin/sh and then failing because the busybox symlink (created in its trigger) hasn't been installed yet. The fix here is a quick hack to unblock. Will look at something more robust to follow up with. Signed-off-by: Erik Sipsma <[email protected]>
* chore(sdk/java): set version of java sdk Update all the pom.xml files when calling `sdk | java | bump` This is in preparation to release the dependencies to maven central. Signed-off-by: Yves Brissaud <[email protected]> * chore(sdk/java): fix version in user module Build the module dependencies using the version of the engine and use that version in the user module pom.xml. Signed-off-by: Yves Brissaud <[email protected]> * fix(sdk/java): fix replacement of placeholder the `dagger-module` placeholder was in conflict with a real `dagger-module` entry (generated sources) so make it more clear one is a placeholder to be sure to edit the right one Signed-off-by: Yves Brissaud <[email protected]> * update integration tests Signed-off-by: Yves Brissaud <[email protected]> --------- Signed-off-by: Yves Brissaud <[email protected]>
* docs: Add recording steps Signed-off-by: Vikram Vaswani <[email protected]> * Updated code Signed-off-by: Vikram Vaswani <[email protected]> * Removed old img module Signed-off-by: Vikram Vaswani <[email protected]> * Updated instructions Signed-off-by: Vikram Vaswani <[email protected]> * Updated code Signed-off-by: Vikram Vaswani <[email protected]> * Updated code Signed-off-by: Vikram Vaswani <[email protected]> * Fix linter Signed-off-by: Vikram Vaswani <[email protected]> * Updated code Signed-off-by: Vikram Vaswani <[email protected]> * Updated code Signed-off-by: Vikram Vaswani <[email protected]> --------- Signed-off-by: Vikram Vaswani <[email protected]>
eunomie
force-pushed
the
sdk-java-internal-field
branch
from
0f70a8f to
76e5d75
Compare
Did a bit more research and realized the hack to ensure busybox symlinks exist in time for glibc's trigger actually may be necessary to get compatibility w/ apko's behavior. The comment added in the code explains more. Signed-off-by: Erik Sipsma <[email protected]>
* docs: update list of features Signed-off-by: Jeremy Adams <[email protected]> * chore: update readme features Signed-off-by: Jeremy Adams <[email protected]> --------- Signed-off-by: Jeremy Adams <[email protected]> Co-authored-by: Greg Kogan <[email protected]>
* notify discord on benchmark failure Signed-off-by: Connor Braa <[email protected]> do not merge: testing with secrets provider Signed-off-by: Connor Braa <[email protected]> disable cancel-in-progress? Signed-off-by: Connor Braa <[email protected]> pass dagger cloud token explicitly Signed-off-by: Connor Braa <[email protected]> try to appease linter Signed-off-by: Connor Braa <[email protected]> add op token wiring Signed-off-by: Connor Braa <[email protected]> add discord webhook Signed-off-by: Connor Braa <[email protected]> secret passing Signed-off-by: Connor Braa <[email protected]> remove test error Signed-off-by: Connor Braa <[email protected]> * add conditional to only send notifications for non-PR runs Signed-off-by: Connor Braa <[email protected]> * in the op secretprovider, handle OP_SERVICE_ACCOUNT_TOKEN= as though it were unset Signed-off-by: Connor Braa <[email protected]> --------- Signed-off-by: Connor Braa <[email protected]>
* docs for llm azure openai service Signed-off-by: kpenfound <[email protected]> * missed word Signed-off-by: kpenfound <[email protected]> * add var examples Signed-off-by: kpenfound <[email protected]> * Update docs/current_docs/agents.mdx Co-authored-by: vikram-dagger <[email protected]> Signed-off-by: Jeremy Adams <[email protected]> --------- Signed-off-by: kpenfound <[email protected]> Signed-off-by: Jeremy Adams <[email protected]> Co-authored-by: Jeremy Adams <[email protected]> Co-authored-by: vikram-dagger <[email protected]>
Fields can now be private. No need to have them public anymore.
All non transient/static fields will be serialized by default and
registered as a field to the Dagger API.
If you want to have a field serialized but not exposed it to the
API (so a user can retrieve the value) then it needs to be annotated with
@internal
Signed-off-by: Yves Brissaud <[email protected]>
variable was not used Signed-off-by: Yves Brissaud <[email protected]>
- public serialized fields are exposed by default - private serialized fields needs a `@Function` annotation to be exposed By exposed this means a default function with the name of the field will be created by the API so that the value can be exposed to the user Signed-off-by: Yves Brissaud <[email protected]>
Signed-off-by: Yves Brissaud <[email protected]>
eunomie
force-pushed
the
sdk-java-internal-field
branch
from
76e5d75 to
63350c3
Compare
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.
Fields can now be private. No need to have them public anymore.
All non transient/static fields will be serialized by default and registered as a field to the Dagger API.
If you want to have a field serialized but not exposed it to the API (so a user can retrieve the value) then it needs to be annotated with
This PR also removed a useless Jsonbuilder creation. This is a left over for a very long time.