Truth8 java 8 extension part of the main library? #746

apflieger · 2020-08-27T14:25:57Z

Hello,

I couldn't find any discussion about that so I wanted to give it a chance.
java 8 is now 6 years old, I doubt there are many legacy project (<java 8) that introduce Truth in their code base whereas people starts new project on java 11 at least now.

What do you think?

cpovirk · 2020-08-27T16:15:37Z

Thanks. We plan to at least merge the 2 artifacts into 1 "soon" (probably months from now if it's left up to me, but this may be the kind of thing that someone could put together a PR for). We will still keep Truth and Truth8 as separate classes for the moment, but maybe we can convince ourselves that it's safe enough to merge the classes, too. (That might well break something, for Java 7 or for Android. We'd have to see. We probably won't drop compatibility with old versions of Android yet, but if Android works but Java 7 does not, we could consider dropping Java 7.)

apflieger · 2020-08-27T16:39:15Z

Yeah I mean merging Truth and Truth8 because that's what is painful to use.

cpovirk · 2020-08-27T16:58:59Z

Yeah. I forget if we've experimented to see what might break. Someday :(

…ately. This is a small step in the general direction of #746 RELNOTES=Truth is now built with `-source 8 -target 8`. This means that it no longer runs under Java 7 VMs. It continues to run under Android, even old versions, for all apps that have [enabled support for Java 8 language features](https://developer.android.com/studio/write/java8-support#supported_features). PiperOrigin-RevId: 387845958

…ately. This is a small step in the general direction of #746 While there, improve the documentation of deprecated `LiteProtoSubject.isEqualTo(MessageLite.Builder)` and `isNotEqualTo(MessageLite.Builder)`, and add a TODO about possible future steps. RELNOTES=Truth is now built with `-source 8 -target 8`. This means that it no longer runs under Java 7 VMs. It continues to run under Android, even old versions, for all apps that have [enabled support for Java 8 language features](https://developer.android.com/studio/write/java8-support#supported_features). PiperOrigin-RevId: 387845958

…ately. This is a small step in the general direction of #746 While there, improve the documentation of deprecated `LiteProtoSubject.isEqualTo(MessageLite.Builder)` and `isNotEqualTo(MessageLite.Builder)`, and add a TODO about possible future steps. RELNOTES=Truth is now built with `-source 8 -target 8`. This means that it no longer runs under Java 7 VMs. It continues to run under Android, even old versions, for all apps that have [enabled support for Java 8 language features](https://developer.android.com/studio/write/java8-support#supported_features). PiperOrigin-RevId: 388266403

bjmi · 2022-01-25T20:46:36Z

com.google.common.truth.Truth and com.google.common.truth.Truth8 classes reside in the same package but in different modules. After converting both maven artifacts to OSGI bundles I get java.lang.NoClassDefFoundError: com/google/common/truth/Subject
by calling Truth8.assertThat(Optional<?>).hasValue(expected) in a downstream test bundle.
Both bundles provide Export-Package: com.google.common.truth in MANIFEST.MF. This looks like split package problem. Truth Subject from "main" bundle is not visible in the "extension" bundle.

At the first glance, there are two possible solutions:

Move Truth8 to a subpackage like com.google.common.truth.extensions.truth8 like ProtoTruth and Re2jSubjects
Merge Truth8 extension with Truth main module (preferred)

cpovirk · 2022-01-25T20:49:32Z

All true, thanks. We intend to merge them, but it's been difficult to schedule time for Truth among our other priorities. It is frequently on my mind, but that doesn't actually help people who are hitting this issue :(

While there, make some minor changes to prepare for moving the `Truth8` methods into `Truth`. This is the first step of #746. RELNOTES=Moved the `truth-java8-extension` classes into the main `truth` artifact. There is no longer any need to depend on `truth-java8-extension`, which is now empty. (We've also removed the `Truth8` [GWT](https://www.gwtproject.org/) module.) PiperOrigin-RevId: 597038300

While there, make some minor changes to prepare for moving the `Truth8` methods into `Truth`. This is the first step of #746. RELNOTES=Moved the `truth-java8-extension` classes into the main `truth` artifact. There is no longer any need to depend on `truth-java8-extension`, which is now empty. (We've also removed the `Truth8` [GWT](https://www.gwtproject.org/) module.) PiperOrigin-RevId: 597243015

At that point, some calls to `assertThat(Object)` will become calls to the new overload. That would be a problem if: - they call `isEqualTo` or `isNotEqualTo`, which we've turned into throwing `@DoNotCall` methods - they have already collected the `Stream` _or_ they want to operate on the `Stream` afterward This CL makes `isEqualTo` and `isNotEqualTo` "work" (though it leaves them deprecated, since they're still a bad idea), and it avoids collecting the `Stream` until necessary. The `isEqualTo` change requires some weird plumbing because I made its failure message retain our warning about `Stream.equals`. That requires creating a new `StreamSubject` with a different `FailureMetadata` instance, which isn't the kind of thing we normally do. (We've done something similar in `ThrowableSubject`, but it is simpler because (a) `ThrowableSubject` uses the "normal" (i.e.., non-no-arg) `check` and (b) `ThrowableSubject` doesn't have to worry about avoiding re-collecting a `Stream`.) We may want to clamp back down on `isEqualTo` in the future. I've set myself a calendar reminder for mid-year. For now, I just want to make `assertThat(Stream)`, which will already be mildly disruptive, from being even more disruptive. (progress toward #746) RELNOTES=Made `StreamSubject` avoid collecting the `Stream` until necessary, and made its `isEqualTo` and `isNotEqualTo` methods no longer always throw. PiperOrigin-RevId: 598241143

At that point, some calls to `assertThat(Object)` will become calls to the new overload. That would be a problem if: - they call `isEqualTo` or `isNotEqualTo`, which we've turned into throwing `@DoNotCall` methods - they have already collected the `Stream` _or_ they want to operate on the `Stream` afterward This CL makes `isEqualTo` and `isNotEqualTo` "work" (though it leaves them deprecated, since they're still a bad idea), and it avoids collecting the `Stream` until necessary. The `isEqualTo` change requires some weird plumbing because I made its failure message retain our warning about `Stream.equals`. That requires creating a new `StreamSubject` with a different `FailureMetadata` instance, which isn't the kind of thing we normally do. (We've done something similar in `ThrowableSubject`, but it is simpler because (a) `ThrowableSubject` uses the "normal" (i.e.., non-no-arg) `check` and (b) `ThrowableSubject` doesn't have to worry about avoiding re-collecting a `Stream`.) We may want to clamp back down on `isEqualTo` in the future. I've set myself a calendar reminder for mid-year. For now, I just want to make `assertThat(Stream)`, which will already be mildly disruptive, from being even more disruptive. (progress toward #746) RELNOTES=Made `StreamSubject` avoid collecting the `Stream` until necessary, and made its `isEqualTo` and `isNotEqualTo` methods no longer always throw. PiperOrigin-RevId: 598607794

…tWithMessage(...).that(optional).isPresent()`, etc., including for `Stream` as well as `Optional`. That is, you no longer need to use `about(optionals())` (or `about(streams())`). (Later, we'll do the same thing for lesser-used APIs like `optionalInts().) This CL does _not_ make it possible to write `Truth.assertThat(optional).isPresent()`: For that, you still need to use `Truth8`. A future CL will eliminate the need to use `Truth8`. This continues our work on #746. RELNOTES=Added `that` overloads to make it possible to write type-specific assertions when using `expect.that(optional)` and `expect.that(stream)`. PiperOrigin-RevId: 597240108

…tWithMessage(...).that(optional).isPresent()`, etc., including for `Stream` as well as `Optional`. That is, you no longer need to use `about(optionals())` (or `about(streams())`). (Later, we'll do the same thing for lesser-used APIs like `optionalInts().) This CL does _not_ make it possible to write `Truth.assertThat(optional).isPresent()`: For that, you still need to use `Truth8`. A future CL will eliminate the need to use `Truth8`. This continues our work on #746. RELNOTES=Added `that` overloads to make it possible to write type-specific assertions when using `expect.that(optional)` and `expect.that(stream)`. PiperOrigin-RevId: 598637400

… main `Truth` class. (and prepare AutoValue tests for the new overloads, since the Eclipse compiler considers the (temporary!) signatures of `assertThat(Optional)` to be ambiguous) We'll post some migration suggestions as part of the release notes. This is the biggest part of #746. RELNOTES=Added `assertThat` overloads for `Optional` and `Stream` to the main `Truth` class. PiperOrigin-RevId: 596422911

…he main `Truth` class. (split off from cl/603361985 because I was waiting on fixes to a couple projects) We'll post some migration suggestions as part of the release notes. This is one of the remaining loose ends of #746. RELNOTES=Added `Truth.assertThat(Path)` and `(OptionalLong)` overloads (copied from `Truth8`). PiperOrigin-RevId: 603474032

…he main `Truth` class. (split off from cl/603361985 because I was waiting on fixes to a couple projects) We'll post some migration suggestions as part of the release notes. This is one of the remaining loose ends of #746. RELNOTES=Added `Truth.assertThat(Path)` and `(OptionalLong)` overloads (copied from `Truth8`). PiperOrigin-RevId: 603713061

This prepares for some migrations off `Truth8` as part of google/truth#746. Notably, after this commit, any Gerrit code that compiles in Google's monorepo should also compile in the Gerrit Git repo. (The reverse is not yet universally true, but I think it is true as long as Gerrit continues to avoid static imports from `Truth8`.) Change-Id: Iefd3f4b2bdc233aeaa75b66f88d529dad712e0db

This will break most users who static import both `Truth.assertThat` and `Truth8.assertThat`. The fix is usually as simple as replacing every reference to `Truth8` with a reference to `Truth`. But we'll post some additional migration information as part of the release notes, as we've already done for [1.3.0](https://github.com/google/truth/releases/tag/v1.3.0) and [1.4.0](https://github.com/google/truth/releases/tag/v1.4.0). (The type parameters existed to avoid that static import conflict. However, the type parameters also _cause other static import conflicts_, so we don't want them in place in the long term.) This is one of the remaining loose ends of #746. RELNOTES=Removed temporary type parameters from `Truth.assertThat(Stream)` and `Truth.assertThat(Optional)`. This can create build errors, which you can fix by replacing all your references to `Truth8` with references to `Truth`. PiperOrigin-RevId: 602078126

This will break most users who static import both `Truth.assertThat` and `Truth8.assertThat`. The fix is usually as simple as replacing every reference to `Truth8` with a reference to `Truth`. But we'll post some additional migration information as part of the release notes, as we've already done for [1.3.0](https://github.com/google/truth/releases/tag/v1.3.0) and [1.4.0](https://github.com/google/truth/releases/tag/v1.4.0). (The type parameters existed to avoid that static import conflict. However, the type parameters also _cause other static import conflicts_, so we don't want them in place in the long term.) This is one of the remaining loose ends of #746. RELNOTES=Removed temporary type parameters from `Truth.assertThat(Stream)` and `Truth.assertThat(Optional)`. This can create build errors, which you can fix by replacing all your references to `Truth8` with references to `Truth`. PiperOrigin-RevId: 604754613

… main `Truth` class. This includes deprecating the old class. This continues work on #746. RELNOTES=Deprecated `Truth8`. All its functionality is now supported through the main `Truth` API. PiperOrigin-RevId: 603756994

… main `Truth` class. This includes deprecating the old class. This continues work on #746. RELNOTES=Deprecated `Truth8`. All its functionality is now supported through the main `Truth` API. PiperOrigin-RevId: 604762265

This prepares for some migrations off `Truth8` as part of google/truth#746. Notably, after this commit, any Gerrit code that compiles in Google's monorepo should also compile in the Gerrit Git repo. (The reverse is not yet universally true, but I think it is true as long as Gerrit continues to avoid static imports from `Truth8`.) Change-Id: Iefd3f4b2bdc233aeaa75b66f88d529dad712e0db Release-Notes: skip

This continues our work on #746. RELNOTES=n/a PiperOrigin-RevId: 607454835

This continues our work on #746. RELNOTES=n/a PiperOrigin-RevId: 607468197

… main `Truth` class. This includes deprecating the old class. This continues work on #746. RELNOTES=Deprecated `Truth8`. All its functionality is now supported through the main `Truth` API. PiperOrigin-RevId: 604762265

This continues our work on #746. RELNOTES=n/a PiperOrigin-RevId: 607468197

cpovirk · 2024-02-29T21:59:41Z

With the release of Truth 1.4.2, this is essentially done: You no longer need the truth-java8-extension artifact, and you no longer need the Truth8 class.

We could still do some more internal cleanup (like to move the tests of the Truth8 functionality into the core directory). But from the perspective of users, this is finally done.

This is an internal loose end in #746. RELNOTES=n/a PiperOrigin-RevId: 611587332

This is an internal loose end in #746. RELNOTES=n/a PiperOrigin-RevId: 611600975

This upgrade makes it possible to use static imports for `assertThat(Optional)` methods again, so I'm also reverting: - 454e167 - 177450d In order to do that, I had to make a change to your `OptionalSubject`. If you'd prefer, I can keep your `OptionalSubject` as it is, and you can continue to use the qualified calls (`OptionalSubject.assertThat(...)`) that I introduced in the earlier commits. I'm also finishing up some changes that were started in a few files in e463a3f: I am changing call sites from `Truth8.assertThat` to `assertThat` (static imported from `Truth`). (There is one exception: In `ApprovalCopierIT`, I have to use a qualified call site because another `assertThat` method is declared in scope, so it would take priority. I at least changed it from the deprecated `Truth8.assertThat` to `Truth.assertThat`.) This ties up one of the loose ends left after google/truth#746. My hope is that the only remaining loose ends are in plugins/code_owners: - cbed17a9a891822d617f3ed1ce59fd7599fe2c8b - 3f2a11b5a25774b30b7094108343578746929256 Release-Notes: skip Change-Id: Ic6c941adaf9ae3007a9dd52ecba76de15798cf40

As of Truth 1.4.2, we can use static imports again. This largely reverts commits cbed17a and 3f2a11b. This is the last Gerrit-related loose end that I know of from google/truth#746, aside from the deprecated automerger plugin: https://gerrit-review.git.corp.google.com/c/plugins/automerger/+/406837 Change-Id: I9dda0e4f8825e1e635f07c786dba0951414d9eb4

cpovirk added P3 not scheduled type=enhancement Make an existing feature better labels Aug 27, 2020

cpovirk mentioned this issue May 29, 2021

Provide Java Module name #605

Open

copybara-service bot mentioned this issue Aug 1, 2021

Build Truth with -source 8 -target 8, updating annotations appropriately. #896

Merged

hendrikebbers mentioned this issue Dec 1, 2022

Using module-info.java in Build hashgraph/hedera-services#4395

Merged

cpovirk mentioned this issue Dec 15, 2022

Wanted: Truth8Junit to support Java 8 types with assume() #1054

Closed

cpovirk self-assigned this Jan 10, 2024

cpovirk added P2 has an ETA and removed P3 not scheduled labels Jan 10, 2024

copybara-service bot mentioned this issue Jan 10, 2024

Move truth-java8-extension classes into the main truth artifact. #1217

Merged

copybara-service bot mentioned this issue Jan 15, 2024

Prepare StreamSubject for adding Truth.assertThat(Stream). #1220

Closed

copybara-service bot mentioned this issue Jan 15, 2024

Make it possible to write expect.that(optional).isPresent(), assertWithMessage(...).that(optional).isPresent(), etc., including for Stream as well as Optional. #1221

Closed

copybara-service bot mentioned this issue Jan 15, 2024

Copy Truth8.assertThat overloads for Optional and Stream to the main Truth class. #1222

Closed

copybara-service bot mentioned this issue Feb 2, 2024

Copy Truth8.assertThat overloads for Path and OptionalLong to the main Truth class. #1236

Closed

copybara-service bot mentioned this issue Feb 6, 2024

Remove temporary type parameters. #1239

Merged

copybara-service bot mentioned this issue Feb 6, 2024

Update docs to reflect that the Java 8 assertions have "moved" to the main Truth class. #1240

Closed

copybara-service bot mentioned this issue Feb 6, 2024

Update docs to reflect that the Java 8 assertions have "moved" to the main Truth class. #1241

Merged

copybara-service bot pushed a commit that referenced this issue Feb 15, 2024

Document more about how and why to migrate off Truth8.

ce685bc

This continues our work on #746. RELNOTES=n/a PiperOrigin-RevId: 607454835

copybara-service bot mentioned this issue Feb 15, 2024

Document more about how and why to migrate off Truth8. #1242

Merged

copybara-service bot pushed a commit that referenced this issue Feb 15, 2024

Document more about how and why to migrate off Truth8.

4d01301

This continues our work on #746. RELNOTES=n/a PiperOrigin-RevId: 607454835

copybara-service bot pushed a commit that referenced this issue Feb 15, 2024

Document more about how and why to migrate off Truth8.

4dd34fe

This continues our work on #746. RELNOTES=n/a PiperOrigin-RevId: 607468197

copybara-service bot pushed a commit that referenced this issue Feb 15, 2024

Document more about how and why to migrate off Truth8.

a6d312e

This continues our work on #746. RELNOTES=n/a PiperOrigin-RevId: 607468197

cpovirk added a commit that referenced this issue Feb 16, 2024

Document more about how and why to migrate off Truth8.

3406074

This continues our work on #746. RELNOTES=n/a PiperOrigin-RevId: 607468197

cpovirk closed this as completed Feb 29, 2024

copybara-service bot pushed a commit that referenced this issue Feb 29, 2024

Migrate tests off Truth8.

0f6348a

This is an internal loose end in #746. RELNOTES=n/a PiperOrigin-RevId: 611587332

copybara-service bot mentioned this issue Feb 29, 2024

Migrate tests off Truth8. #1255

Merged

copybara-service bot pushed a commit that referenced this issue Feb 29, 2024

Migrate tests off Truth8.

213a156

This is an internal loose end in #746. RELNOTES=n/a PiperOrigin-RevId: 611600975

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Truth8 java 8 extension part of the main library? #746

Truth8 java 8 extension part of the main library? #746

apflieger commented Aug 27, 2020

cpovirk commented Aug 27, 2020

apflieger commented Aug 27, 2020

cpovirk commented Aug 27, 2020

bjmi commented Jan 25, 2022

cpovirk commented Jan 25, 2022

cpovirk commented Feb 29, 2024

Truth8 java 8 extension part of the main library? #746

Truth8 java 8 extension part of the main library? #746

Comments

apflieger commented Aug 27, 2020

cpovirk commented Aug 27, 2020

apflieger commented Aug 27, 2020

cpovirk commented Aug 27, 2020

bjmi commented Jan 25, 2022

cpovirk commented Jan 25, 2022

cpovirk commented Feb 29, 2024