Should KSPropertyDeclarationImpl.findOverridee use singleOrNull? #174
Comments
|
After changing this to |
|
i think we need to move it to be similar to how functions are implemented: |
yigit
added a commit
to yigit/ksp
that referenced
this issue
Nov 27, 2020
This CL updates KSPropertyDecl.findOverridee to use the same infra that is used by KSFunctionDeclaration. Fixes: google#174
neetopia
pushed a commit
that referenced
this issue
Nov 30, 2020
This CL updates KSPropertyDecl.findOverridee to use the same infra that is used by KSFunctionDeclaration. Fixes: #174
copybara-service bot
pushed a commit
to androidx/androidx
that referenced
this issue
Dec 1, 2020
This CL fixes another case on variance inheritance for overrides where if there is an overridden method w/ star projected type argument, we should not use start projection but instead pick covariant (out). To reproduce it, i've expanded method spec helper test to cover methods that are not overridden. Unfortunately, it revelaled an issue where if a kotlin interface overrides another interface with generics, it is not possible to generate kotlin override for it if the type argument is a final type. (b/174313780). Only one test case hits it so I've added a way to ignore them for now (we cannot fix it until room generates kotlin sources). I've also hit some KSP bugs: google/ksp#175 Overrides might fail when overriding with generics. Luckily, findOverridee method finds it so I've added a workaround by calling findOverridee. We could possibly only call it instead of calling both but resolver bug will be eventually fixed in KSP hence I kept that call. google/ksp#174 Since we call findOverridee for properties now, it might trigger another issue in KSP. Added a try catch for it now. google/ksp#173 KSP might throw when resolving deserialized type alias descriptors. Added a try catch for it where we hit the issue. (reading annotations from an element that also has @throws annotation). google/ksp#164 This issue also affects get getJvmName for which we have safeGetJvmName. In this CL, it happens when finding overridee for getter/setter methods in java. Added a try catch for it for now. Bug: 174313780 Bug: 160322705 Test: MethodSpecHelperTest Change-Id: Ie6347f46027c317ed7784e8d2a0c834012fb4b52
copybara-service bot
pushed a commit
to androidx/androidx
that referenced
this issue
Dec 9, 2020
This version is important because it also moves KSP's kotlin dependency to 1.4.20. I had to specify more versions for testing because we were getting kotlin compiler dependencies from kotlin compile testing, which is still on 1.4.10. Now compiler-processing-testing artifact explicitly specifies those version to match whatever AndroidX is using. I've also removed workarounds for the following issues as they are fixed. google/ksp#170 google/ksp#174 Bug: 160322705 Test: existing tests Change-Id: If948bcb762d1fad4911cf833a85cda6a36e7c787
neetopia
pushed a commit
that referenced
this issue
Dec 23, 2020
This CL updates KSPropertyDecl.findOverridee to use the same infra that is used by KSFunctionDeclaration. Fixes: #174
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
ksp/compiler-plugin/src/main/kotlin/com/google/devtools/ksp/symbol/impl/kotlin/KSPropertyDeclarationImpl.kt
Line 107 in 638d9a1
This now fails for me on an overridden val. The way this code is written now, looks like the last '?' in this part of the statement
.single { it.overriddenDescriptors.isEmpty() }?is never going to happen, sincesinglenever returns null.The text was updated successfully, but these errors were encountered: