Fix SandboxClassLoader to obey parent-first contract #2905
Conversation
xian
force-pushed
the
classloader-correctness
branch
from
378f8f9
to
fe9c357
Compare
xian
force-pushed
the
classloader-correctness
branch
2 times, most recently
from
b5e5fc0
to
697cc52
Compare
Supports PowerMock and experimental Mockito mode which enables mocking final classes/methods. Supports Package.getPackage().
xian
force-pushed
the
classloader-correctness
branch
from
697cc52
to
6b22a72
Compare
| // See https://github.com/robolectric/robolectric/issues/521 | ||
| if (name.startsWith("android.R")) { | ||
| // android.R and com.android.internal.R classes must be loaded from the framework jar | ||
| if (name.matches("(android|com\\.android\\.internal)\\.R(\\$.+)?")) { |
There was a problem hiding this comment.
name.startsWith("android.R) || name.startsWith("com.android.internal.R") may be more clear here.
There was a problem hiding this comment.
We're pretty confident there'll never be an android.Riboflavin or com.android.internal.Roustabout or something? :-)
There was a problem hiding this comment.
Yeah, I think so, although, you can never say never :-D I think this will go away when we check the jar as the source of truth for what to acquire though anyway, so I'm not fussed about this :-)
| ReflectionHelpers.setStaticField(androidBuildVersionClass, "SDK_INT", sdkConfig.getApiLevel()); | ||
| ReflectionHelpers.setStaticField(androidBuildVersionClass, "RELEASE", sdkConfig.getAndroidVersion()); | ||
|
|
||
| PackageResourceTable systemResourceTable = sdkEnvironment.getSystemResourceTable(getJarResolver()); | ||
| PackageResourceTable systemResourceTable = (sdkEnvironment).getSystemResourceTable(getJarResolver()); |
There was a problem hiding this comment.
extraneous parens around (sdkEnvironment) ?
| @@ -107,6 +107,11 @@ public void testVersionConfiguration() { | |||
| .isEqualTo("4.4_r1"); | |||
| } | |||
|
|
|||
|
|
|||
| @Test public void assertThatz() throws Exception { | |||
There was a problem hiding this comment.
Want to do this more? I'm guessing that it is checking that things are loaded from the right classloader, but I'm a little confused as its an assert for a primitive rather than an object?
There was a problem hiding this comment.
Fixed the test name. I came across a classloader conflict because org.junit.Assert loads some hamcrest things.
|
Dogfood green! :-) |
|
Yey!! :-)
…On 22 February 2017 at 14:30, Christian Williams ***@***.***> wrote:
Merged #2905 <#2905>.
—
You are receiving this because your review was requested.
Reply to this email directly, view it on GitHub
<#2905 (comment)>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/AGBaeA6cJm0pTkdmWEgGGYdFKWxIeyWpks5rfLcKgaJpZM4L9_5n>
.
|
|
Well done. |
| @@ -75,8 +75,7 @@ | |||
| private final SdkPicker sdkPicker; | |||
| private final ConfigMerger configMerger; | |||
|
|
|||
| private TestLifecycle<Application> testLifecycle; | |||
| private DependencyResolver dependencyResolver; | |||
| private transient DependencyResolver dependencyResolver; | |||
There was a problem hiding this comment.
Why is dependencyResolver transient?
Fixes #2208.
Fixes #2677.