Automatic-Module-Name missing from sub-module manifests #2920
Assignees
Milestone
Comments
|
I just hit this too. The |
|
I think this should be fixed and released ASAP as 23.1 for example, given the fact that there is less than 20 days to JDK9 General Availability. |
|
Any progress on this issue? 23.1 was released today, but I do not see |
Closed
Merged
cpovirk
added a commit
to google/auto
that referenced
this issue
Oct 4, 2019
This makes it practical to use from code that uses modules. Compare to Guava: https://github.com/google/guava/blob/5496c37d4d904869297c2ced1f0d20e6f1507eaa/guava/pom.xml#L60 google/guava#2920 RELNOTES=AutoService: Set an Automatic-Module-Name. ------------- Created by MOE: https://github.com/google/moe MOE_MIGRATED_REVID=272882826
cpovirk
added a commit
to google/auto
that referenced
this issue
Oct 4, 2019
This makes it practical to use from code that uses modules. Compare to Guava: https://github.com/google/guava/blob/5496c37d4d904869297c2ced1f0d20e6f1507eaa/guava/pom.xml#L60 google/guava#2920 RELNOTES=AutoService: Set an Automatic-Module-Name. ------------- Created by MOE: https://github.com/google/moe MOE_MIGRATED_REVID=272882826
cpovirk
added a commit
to google/error-prone
that referenced
this issue
Oct 4, 2019
This makes it practical to use from code that uses modules. Compare to Guava: https://github.com/google/guava/blob/5496c37d4d904869297c2ced1f0d20e6f1507eaa/guava/pom.xml#L60 google/guava#2920 This is a step toward #1116 (though it doesn't address most artifacts, including javac). Modules users may still see problems because com.google.errorprone.annotations is split between error_prone_annotations and error_prone_type_annotations. I wonder if the best solution to that is going to be to merge the two into error_prone_annotations, turning error_prone_type_annotations into an empty "forwarding" artifact. That would require compiling some files for Java 7 and some for Java 8, but that's doable (if a little clumsy). It might cause problems from some tool somewhere, though. Given that error_prone_type_annotations isn't nearly as commonly used, maybe you can wait it out until it's time to drop Java 7 support? In any case, I don't think this CL makes things any worse. RELNOTES=Set Automatic-Module-Name for annotations package. ------------- Created by MOE: https://github.com/google/moe MOE_MIGRATED_REVID=272890641
cpovirk
added a commit
to google/error-prone
that referenced
this issue
Oct 4, 2019
This makes it practical to use from code that uses modules. Compare to Guava: https://github.com/google/guava/blob/5496c37d4d904869297c2ced1f0d20e6f1507eaa/guava/pom.xml#L60 google/guava#2920 This is a step toward #1116 (though it doesn't address most artifacts, including javac). Modules users may still see problems because com.google.errorprone.annotations is split between error_prone_annotations and error_prone_type_annotations. I wonder if the best solution to that is going to be to merge the two into error_prone_annotations, turning error_prone_type_annotations into an empty "forwarding" artifact. That would require compiling some files for Java 7 and some for Java 8, but that's doable (if a little clumsy). It might cause problems from some tool somewhere, though. Given that error_prone_type_annotations isn't nearly as commonly used, maybe you can wait it out until it's time to drop Java 7 support? In any case, I don't think this CL makes things any worse. RELNOTES=Set Automatic-Module-Name for annotations package. ------------- Created by MOE: https://github.com/google/moe MOE_MIGRATED_REVID=272890641
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Version 23 was supposed to add
Automatic-Module-Nameto the manifest file but it only did so forguava-parent: https://github.com/google/guava/blob/v23.0/pom.xml#L119Please add
Automatic-Module-Namefor all sub-modules otherwise JDK9 clients will end up with the wrong module name.The text was updated successfully, but these errors were encountered: