Merge from moe writing branch from 9f986abd0f985e717dde1476aeaf058df50064ff #229

cgruber · 2015-08-29T21:38:23Z

A variety of changes imported from google's internal repository, including:

Generated sources are now formatted in google style using google-java-format
Some initial work for full compatibility with the JSR-330 spec and TCK
- Cyclic dependencies are now supported if they are broken by an indirection of injecting a Lazy<T> or Provider<T> (helping bring Dagger into compliance with JSR-330, and addressing Support dependency cycles containing Lazy or Provider #200)
- Initially posted the JSR-330 TCK sources to show how dagger is complying, but reverted, as we will be relying on the published binary artifact of the TCK.
- Other compliance work is still in google's internal repository, but is delayed pending a push of compile-testing
- Some build-time optimizations, reducing the time to produce very large graphs, or graphs with large numbers of modules
- Some correctness fixes to set bindings in subcomponents.
- Some improved error reporting on missing bindings in the context of subcomponents
- Some fixes to javadocs, particularly around Lazy<T>
- Type inference fixes that should make Dagger less brittle building in Javac 8/9+ and more generally deal with type inference incompatibilities between versions of javac.
- Miscellaneous build changes to accommodate the above
- Miscellaneous re-factorings in support of the above or forthcoming features, and to move some code into auto-common
- And more!

googlebot · 2015-08-29T21:38:29Z

We found a Contributor License Agreement for you (the sender of this pull request), but were unable to find agreements for the commit author(s). If you authored these, maybe you used a different email address in the git commits than was used to sign the CLA (login here to double check)? If these were authored by someone else, then they will need to sign a CLA as well, and confirm that they're okay with these being contributed to Google.

cgruber · 2015-08-29T22:15:26Z

clabot lies. these are googler originated changes.

cgruber · 2015-08-29T22:23:10Z

Ok, this is hella-weird. It's succeeding on oracle JDK but on OpenJDK one test is failing, when tested from the pull request branch, but both pass when tested from the branch push. I don't know if this test is flaky, or what, but it's kind of a problem, since if it fails on the merge, we won't get a snapshot.

@netdpb and @gk5885 - can you folks look at this build output and think through whether this is just a flaky test, or some environemental issue.

https://travis-ci.org/google/dagger/builds/77860604

netdpb · 2015-09-10T14:41:43Z

It's weird. The two MembersInjector fields are unused in the subcomponent,
and they appear out of order.

Why are they even there, if they're unused?
Why are they out of order? Are we using HashSet or HashMap instead of
a data structure with predictable iteration order?

On Sat, Aug 29, 2015 at 6:23 PM Christian Edward Gruber <
[email protected]> wrote:

Ok, this is hella-weird. It's succeeding on oracle JDK but on OpenJDK one
test is failing, when tested from the pull request branch, but both pass
when tested from the branch push. I don't know if this test is flaky, or
what, but it's kind of a problem, since if it fails on the merge, we won't
get a snapshot.

@netdpb https://github.com/netdpb and @gk5885
https://github.com/gk5885 - can you folks look at this build output and
think through whether this is just a flaky test, or some environemental
issue.

https://travis-ci.org/google/dagger/builds/77860604

—
Reply to this email directly or view it on GitHub
#229 (comment).

--dpb

googlebot · 2015-10-01T05:46:26Z

We found a Contributor License Agreement for you (the sender of this pull request) and all commit authors, but as best as we can tell these commits were authored by someone else. If that's the case, please add them to this pull request and have them confirm that they're okay with these commits being contributed to Google. If we're mistaken and you did author these commits, just reply here to confirm.

------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=90533797

*** Reason for rollback *** Breaks android build *** Original change description *** Replace use of deprecated Futures.fallback with Futures.catching. *** ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=90537687

…ecutes the maven build, and update the pom files with the relevant updates (not yet merged in from github). ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=95043495

------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=98345805

*** Reason for rollback *** Broke some android targets that are somehow putting the old version of Guava on the classpath for the processor. *** Original change description *** Refactor module information into ModuleDescriptor. This provides a single, consistent way to access information about modules. Note: This CL changes the order in which modules are referenced in components, but that shouldn't ever matter. *** ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=98353908

Flame graphs were taken from a build graph with 50 modules and 200 bindings per module. In each module, the provider methods have an order where each requires the object provided by the prior method. The number of samples is roughly halved, which suggests a ~50% reduction in build time for this binding graph. ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=98407197

…le Java Formatter. ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=98411635

…nce to FluentIterable.append. *** Original change description *** Automated g4 rollback of changelist 98344654. *** Reason for rollback *** Broke some android targets that are somehow putting the old version of Guava on the classpath for the processor. *** Original change description *** Refactor module information into ModuleDescriptor. This provides a single, consistent way to access information about modules. Note: This CL changes the order in which modules are referenced in components, but that shouldn't ever matter. *** *** ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=98420884

*** Reason for rollback *** Broke an internal set of tests *** Original change description *** Ensure that all sources emitted by Dagger are formatted with the Google Java Formatter. *** ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=98422425

------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=98439423

Two Produced objects compare equal if both are successful, with equal values, or both are failed, with equal exceptions. ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=98445597

…sses ComponentWriter and SubcomponentWriter so that we can use fields instead of passing arguments to methods. ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=98740715

…emberSelectSnippet(BindingKey). This lets SubcomponentWriter implement a recursive search up the component tree without combining map data structures. ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=98744070

…ava have been addressed. *** Original change description *** Automated g4 rollback of changelist 98411635. *** Reason for rollback *** Broke internal tests *** Original change description *** Ensure that all sources emitted by Dagger are formatted with the Google Java Formatter. *** ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=98745221

…w method, getMultibindingContributionSnippet(ContributionBinding). This lets SubcomponentWriter implement a recursive search up the component tree without combining map data structures. Use MemberSelect for those snippets in order to determine whether the snippet is owned by the subcomponent or a parent. ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=98753147

This adds a fake binding for the ProductionComponentMonitor so that producer factories can request them like normal provision bindings. ------------- Created by MOE: https://github.com/google/moe MOE_MIGRATED_REVID=106149671

… skipped because its input failed. ------------- Created by MOE: https://github.com/google/moe MOE_MIGRATED_REVID=106328773

This fixes TAP for domain_registry_large. ------------- Created by MOE: https://github.com/google/moe MOE_MIGRATED_REVID=106402329

…efficient, it allows us to simplify the way we store and reference fields in the future. ------------- Created by MOE: https://github.com/google/moe MOE_MIGRATED_REVID=106533383

------------- Created by MOE: https://github.com/google/moe MOE_MIGRATED_REVID=106597867

------------- Created by MOE: https://github.com/google/moe MOE_MIGRATED_REVID=106950444

…re resolved in the subgraph, not those that are inherited. ------------- Created by MOE: https://github.com/google/moe MOE_MIGRATED_REVID=106970691

------------- Created by MOE: https://github.com/google/moe MOE_MIGRATED_REVID=107049195

…tic bindings, and allow subcomponent factory methods to elide them too. ------------- Created by MOE: https://github.com/google/moe MOE_MIGRATED_REVID=107097167

--------------------- Automatic change by MOE ------------- Created by MOE: https://github.com/google/moe MOE_MIGRATED_REVID=107358545

…nce shade can't merge them. Tested: Actually ran the build shading with additional sealed bits in the classpath ------------- Created by MOE: https://github.com/google/moe MOE_MIGRATED_REVID=107439464

… of the functional tests will run, add auto-factory to support its use in the functional test, and bump guava to the top of the classpath to avoid truth/auto-factory/etc. from supplying guava-18.0. ------------- Created by MOE: https://github.com/google/moe MOE_MIGRATED_REVID=107439700

googlebot · 2015-11-12T01:13:45Z

We found a Contributor License Agreement for you (the sender of this pull request), but were unable to find agreements for the commit author(s). If you authored these, maybe you used a different email address in the git commits than was used to sign the CLA (login here to double check)? If these were authored by someone else, then they will need to sign a CLA as well, and confirm that they're okay with these being contributed to Google.

This doesn't mean that we should have naming sequences in those patterns, but our types that derive names from inputted types should not jump through hoops to have a style-compliant name. Addresses Github issue #173 ------------- Created by MOE: https://github.com/google/moe MOE_MIGRATED_REVID=107509623

------------- Created by MOE: https://github.com/google/moe MOE_MIGRATED_REVID=107703506

------------- Created by MOE: https://github.com/google/moe MOE_MIGRATED_REVID=107708335

Tested: pushed a trial branch with this change to travis (https://travis-ci.org/google/dagger/jobs/90810414) ------------- Created by MOE: https://github.com/google/moe MOE_MIGRATED_REVID=107711893

cgruber · 2015-11-13T03:02:02Z

Huzzah! Builds clean!

gk5885 · 2015-11-13T08:16:52Z

Lgtm

cgruber · 2015-11-13T08:18:48Z

After a long journey, this branch cometh home. (For the record, I got an LGTM from @gk5885 outside of this issue).

cgruber · 2015-11-13T08:19:04Z

Oh, and hten also in this issue, but github lags. :)

…f985e717dde1476aeaf058df50064ff Merge from moe writing branch from 9f986ab

googlebot added the cla: no label Aug 29, 2015

cgruber force-pushed the moe_writing_branch_from_9f986abd0f985e717dde1476aeaf058df50064ff branch from 707a8df to e65ed9d Compare August 29, 2015 22:13

cgruber added cla: yes and removed cla: no labels Aug 29, 2015

googlebot added cla: no and removed cla: yes labels Aug 29, 2015

cgruber force-pushed the moe_writing_branch_from_9f986abd0f985e717dde1476aeaf058df50064ff branch from e65ed9d to 3960cb0 Compare October 1, 2015 05:46

cgruber force-pushed the moe_writing_branch_from_9f986abd0f985e717dde1476aeaf058df50064ff branch from 3960cb0 to b50f06c Compare October 21, 2015 00:31

cgruber and others added 10 commits October 20, 2015 17:32

Un-secret jesse beder's username

1e87aff

Replace use of deprecated Futures.fallback with Futures.catching.

096293d

------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=90533797

Create an internal test that generates the open-source project and ex…

487124f

…ecutes the maven build, and update the pom files with the relevant updates (not yet merged in from github). ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=95043495

Add documentation for multibindings.

4692153

------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=98345805

Ensure that all sources emitted by Dagger are formatted with the Goog…

ff589dd

…le Java Formatter. ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=98411635

cgruber force-pushed the moe_writing_branch_from_9f986abd0f985e717dde1476aeaf058df50064ff branch from b50f06c to 2d42077 Compare October 21, 2015 00:55

gk5885 and others added 6 commits October 20, 2015 18:06

Remove usage of an internal API not released in Guava.

9e4d56c

------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=98439423

Add value semantics to Produced.

9230291

Two Produced objects compare equal if both are successful, with equal values, or both are failed, with equal exceptions. ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=98445597

First step at refactoring ComponentGenerator: Move methods to new cla…

de9434d

…sses ComponentWriter and SubcomponentWriter so that we can use fields instead of passing arguments to methods. ------------- Created by MOE: http://code.google.com/p/moe-java MOE_MIGRATED_REVID=98740715

jbeder and others added 12 commits November 11, 2015 16:33

Hook up the ProductionComponentMonitor.

0e6f248

This adds a fake binding for the ProductionComponentMonitor so that producer factories can request them like normal provision bindings. ------------- Created by MOE: https://github.com/google/moe MOE_MIGRATED_REVID=106149671

Clarify the javadoc of ProducerMonitor in the case when a producer is…

6767ed5

… skipped because its input failed. ------------- Created by MOE: https://github.com/google/moe MOE_MIGRATED_REVID=106328773

Make InjectProcessingStep defer incomplete types

1505feb

This fixes TAP for domain_registry_large. ------------- Created by MOE: https://github.com/google/moe MOE_MIGRATED_REVID=106402329

Use raw types rather than package proxies. In addition to being more …

28032ee

…efficient, it allows us to simplify the way we store and reference fields in the future. ------------- Created by MOE: https://github.com/google/moe MOE_MIGRATED_REVID=106533383

Remove ProxyClassAndField as it is no longer used.

191de01

------------- Created by MOE: https://github.com/google/moe MOE_MIGRATED_REVID=106597867

Small refactor to simplify Resolver.lookupBindings.

5dcbd96

------------- Created by MOE: https://github.com/google/moe MOE_MIGRATED_REVID=106950444

Only write and initialize fields in subcomponents for bindings that a…

e424d92

…re resolved in the subgraph, not those that are inherited. ------------- Created by MOE: https://github.com/google/moe MOE_MIGRATED_REVID=106970691

Fix Dagger maven build for android examples

a6ac28b

------------- Created by MOE: https://github.com/google/moe MOE_MIGRATED_REVID=107049195

Make subcomponents not have instance fields for modules with only sta…

fa92875

…tic bindings, and allow subcomponent factory methods to elide them too. ------------- Created by MOE: https://github.com/google/moe MOE_MIGRATED_REVID=107097167

Update changelog with dagger 2 information, and reformat.

6a2237a

--------------------- Automatic change by MOE ------------- Created by MOE: https://github.com/google/moe MOE_MIGRATED_REVID=107358545

Remove package sealing certs from the jars when they're shaded up, si…

93c9abf

…nce shade can't merge them. Tested: Actually ran the build shading with additional sealed bits in the classpath ------------- Created by MOE: https://github.com/google/moe MOE_MIGRATED_REVID=107439464

cgruber force-pushed the moe_writing_branch_from_9f986abd0f985e717dde1476aeaf058df50064ff branch from 4ce8303 to a012aa2 Compare November 12, 2015 01:13

googlebot added cla: no and removed cla: yes labels Nov 12, 2015

cgruber force-pushed the moe_writing_branch_from_9f986abd0f985e717dde1476aeaf058df50064ff branch from d3de6fb to a012aa2 Compare November 12, 2015 20:51

ronshapiro and others added 4 commits November 12, 2015 18:52

Make it possible to depend on subcomponent builders.

99d6408

------------- Created by MOE: https://github.com/google/moe MOE_MIGRATED_REVID=107703506

Use a flag formulation that is compatible with standard javac.

2049b2c

------------- Created by MOE: https://github.com/google/moe MOE_MIGRATED_REVID=107708335

Bump the android API/tools version to 23.

91f7d8b

Tested: pushed a trial branch with this change to travis (https://travis-ci.org/google/dagger/jobs/90810414) ------------- Created by MOE: https://github.com/google/moe MOE_MIGRATED_REVID=107711893

cgruber assigned gk5885 Nov 13, 2015

cgruber added a commit that referenced this pull request Nov 13, 2015

Merge pull request #229 from google/moe_writing_branch_from_9f986abd0…

9293bcf

…f985e717dde1476aeaf058df50064ff Merge from moe writing branch from 9f986ab

cgruber merged commit 9293bcf into master Nov 13, 2015

cgruber deleted the moe_writing_branch_from_9f986abd0f985e717dde1476aeaf058df50064ff branch November 13, 2015 08:19

jan-stoltman-bolt mentioned this pull request May 6, 2024

Detect dependency cycles containing Lazy or Provider during the compilation time #4298

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Merge from moe writing branch from 9f986abd0f985e717dde1476aeaf058df50064ff #229

Merge from moe writing branch from 9f986abd0f985e717dde1476aeaf058df50064ff #229

cgruber commented Aug 29, 2015

googlebot commented Aug 29, 2015

cgruber commented Aug 29, 2015

cgruber commented Aug 29, 2015

netdpb commented Sep 10, 2015

googlebot commented Oct 1, 2015

googlebot commented Nov 12, 2015

cgruber commented Nov 13, 2015

gk5885 commented Nov 13, 2015

cgruber commented Nov 13, 2015

cgruber commented Nov 13, 2015

Merge from moe writing branch from 9f986abd0f985e717dde1476aeaf058df50064ff #229

Merge from moe writing branch from 9f986abd0f985e717dde1476aeaf058df50064ff #229

Conversation

cgruber commented Aug 29, 2015

googlebot commented Aug 29, 2015

cgruber commented Aug 29, 2015

cgruber commented Aug 29, 2015

netdpb commented Sep 10, 2015

googlebot commented Oct 1, 2015

googlebot commented Nov 12, 2015

cgruber commented Nov 13, 2015

gk5885 commented Nov 13, 2015

cgruber commented Nov 13, 2015

cgruber commented Nov 13, 2015