java.lang.AssertionError: assertion failed: Modified names for XXX is empty #4185
Labels
Comments
|
I've not been able to replicate this using a simple project. |
|
@jebbench : Interesting. It might be worth reporting this issue upstream at https://github.com/sbt/zinc . According to https://github.com/pantsbuild/pants/blob/release_1.3.0.dev6/3rdparty/jvm/org/scala-sbt/BUILD |
|
Interesting, this is coming from sbt/sbt#2490. Will fix soon. I think Grzgorz didn't think about Java APIs while enforcing the contract between name hashes and API changes. |
|
Relates to #4477 |
|
...and thus to #4729. |
stuhood
pushed a commit
that referenced
this issue
May 11, 2018
### Problem (this change depends on #4728) As described on #4728, we currently need internal knowledge of zinc analysis in order to make it portable, and to extract products that are used by other tasks. Since the zinc analysis format is changing significantly (moving to protobuf, gaining/losing fields, etc), we need to stop parsing it. ### Solution Incorporates the version of the zinc wrapper published in #4728, and uses it in a `analysis.zinc` task to generate the `product_deps_by_src` and `classes_by_source` products from a `zinc_analysis` product published by zinc. Expands the extracted `Zinc` subsystem from #4720 to encapsulate compile dependency calculation, and expose the `FingerprintStrategy` and `DependencyContext` that `compile.zinc` was using. ### Result Zinc analysis is now a black box from the perspective of the pants python code. Also, the `compile.jvm-dep-check` task moved to the `lint` goal, and gained the unused dependency checks that used to be inlined in `compile.zinc` (and which would have prevented separation of the analysis extraction task). Fixes #4477, fixes #4513, fixes #4185, fixes #5444, and hopefully fixes a few other odd incremental compilation bugs.
stuhood
pushed a commit
that referenced
this issue
May 11, 2018
### Problem (this change depends on #4728) As described on #4728, we currently need internal knowledge of zinc analysis in order to make it portable, and to extract products that are used by other tasks. Since the zinc analysis format is changing significantly (moving to protobuf, gaining/losing fields, etc), we need to stop parsing it. ### Solution Incorporates the version of the zinc wrapper published in #4728, and uses it in a `analysis.zinc` task to generate the `product_deps_by_src` and `classes_by_source` products from a `zinc_analysis` product published by zinc. Expands the extracted `Zinc` subsystem from #4720 to encapsulate compile dependency calculation, and expose the `FingerprintStrategy` and `DependencyContext` that `compile.zinc` was using. ### Result Zinc analysis is now a black box from the perspective of the pants python code. Also, the `compile.jvm-dep-check` task moved to the `lint` goal, and gained the unused dependency checks that used to be inlined in `compile.zinc` (and which would have prevented separation of the analysis extraction task). Fixes #4477, fixes #4513, fixes #4185, fixes #5444, and hopefully fixes a few other odd incremental compilation bugs.
stuhood
pushed a commit
that referenced
this issue
May 11, 2018
### Problem (this change depends on #4728) As described on #4728, we currently need internal knowledge of zinc analysis in order to make it portable, and to extract products that are used by other tasks. Since the zinc analysis format is changing significantly (moving to protobuf, gaining/losing fields, etc), we need to stop parsing it. ### Solution Incorporates the version of the zinc wrapper published in #4728, and uses it in a `analysis.zinc` task to generate the `product_deps_by_src` and `classes_by_source` products from a `zinc_analysis` product published by zinc. Expands the extracted `Zinc` subsystem from #4720 to encapsulate compile dependency calculation, and expose the `FingerprintStrategy` and `DependencyContext` that `compile.zinc` was using. ### Result Zinc analysis is now a black box from the perspective of the pants python code. Also, the `compile.jvm-dep-check` task moved to the `lint` goal, and gained the unused dependency checks that used to be inlined in `compile.zinc` (and which would have prevented separation of the analysis extraction task). Fixes #4477, fixes #4513, fixes #4185, fixes #5444, and hopefully fixes a few other odd incremental compilation bugs.
|
I still get this with pants 1.7.0rc0. Others here have reported this as well. If I compile again the problem goes away. I don't know how to reproduce it. |
|
Pants is most likely not using the latest Zinc version since this change was merged and released some time ago, with the removal of that assertion. |
|
@youngderekm : |
|
(see the linked PR) |
|
Zinc version 1.1.7 should have the issue here reported fixed, let me know if that doesn’t work! |
|
Twitter is currently working on incorporating the 1.8.x series, so I'll probably be cutting a |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Pants 1.3.0.dev3
When compiling Java classes after adding hashcode and equals methods we get the error:
To reproduce:
The text was updated successfully, but these errors were encountered: