3 ms·

Florian continues to comment on the post, posts this link for confirmation of the source files still being in Android source http://android.git.kernel.org/?p=p

Florian continues to comment on the post, posts this link for confirmation of the source files still being in Android source http://android.git.kernel.org/?p=platform/libcore.git;a=tree;f=support/src/test/java/org/apache/harmony/security/tests/support/acl;h=569f09ad5547efdf94fc000cc5a4e7ffa0764c63;hb=refs/heads/froyo-plus-aosp http://android.git.kernel.org/?p=platform/libcore.git;a=tree...

yuhong 16y ago
The nice thing about open source is that the public can investigate what exactly is happening here, which is exactly what I am trying to do.
yuhong 16y ago
Clue: Go up to the support directory and click the HEAD link and notice the difference.
- orangecat 16y ago
  Nice. Specific commit: http://android.git.kernel.org/?p=platform/libcore.git;a=commitdiff;h=a13e4f2c06c4673a1d6fe52409432ccc576c2fe3 http://android.git.kernel.org/?p=platform/libcore.git;a=comm...
  - yuhong 16y ago
    Note however that deleting the files don't suddenly make the copyright infringement claims invalid.
    - orangecat 16y ago
      Sure, but it appears that there was no intent and that the actual damages from the infringement are roughly $0. It may result in a small penalty, but it's in no way a threat to Android's future.
      - yuhong 16y ago
        > It may result in a small penalty Especially considering they did register their copyright on Java 5. I know because I have read the court exhibits.
      - kingkilr 16y ago
        Yup, statutory damages for copyright infringement suck for individuals, but for someone like Google they're not awful (AFAIR it's something like $7,500 or $150,000 depending on something per count).
        nohat 16y ago
        True, but if they did ship something tainted, then that could be a very large count.
catch23 16y ago
I think Florian was looking at Froyo, while the author was looking at master. The offending source files are not in Gingerbread: http://android.git.kernel.org/?p=platform/libcore.git;a=tree;f=support/src/test/java/org/apache/harmony/security/tests/support/acl;h=801cdcf96fa4e6c12398782b2e936dbcc311b6c8;hb=refs/heads/gingerbread http://android.git.kernel.org/?p=platform/libcore.git;a=tree...
brown9-2 16y ago
His response seems to be along the line of "how do we know the unit test code didn't get put on a device and shipped?": For this one as well as quote #3, what's missing is an analysis of the code actually shipped with Android-based devices. There are many such devices. Some may be more security-oriented (hence more likely to make use of the "acl" and PolicyNodeImpl type of classes) than others. So we're talking about a totally unsubstantiated claim here. There may be build scripts, and some files may not be in them by default. That doesn't mean that no maker of an Android-based device actually decided to build them into a product, especially if Apache license headers appeared to allow it. If this scenario did in fact occur, then isn't it the device manufacturer's fault and problem, and not Google's responsibility? Which party builds/packages the Android distributions for the actual phone devices that the OS will appear on?

yuhong 16y ago

The nice thing about open source is that the public can investigate what exactly is happening here, which is exactly what I am trying to do.

yuhong 16y ago

Clue: Go up to the support directory and click the HEAD link and notice the difference.

orangecat 16y ago
Nice. Specific commit: http://android.git.kernel.org/?p=platform/libcore.git;a=commitdiff;h=a13e4f2c06c4673a1d6fe52409432ccc576c2fe3 http://android.git.kernel.org/?p=platform/libcore.git;a=comm...
- yuhong 16y ago
  Note however that deleting the files don't suddenly make the copyright infringement claims invalid.
  - orangecat 16y ago
    Sure, but it appears that there was no intent and that the actual damages from the infringement are roughly $0. It may result in a small penalty, but it's in no way a threat to Android's future.
    - yuhong 16y ago
      > It may result in a small penalty Especially considering they did register their copyright on Java 5. I know because I have read the court exhibits.
    - kingkilr 16y ago
      Yup, statutory damages for copyright infringement suck for individuals, but for someone like Google they're not awful (AFAIR it's something like $7,500 or $150,000 depending on something per count).
      - nohat 16y ago
        True, but if they did ship something tainted, then that could be a very large count.

catch23 16y ago

I think Florian was looking at Froyo, while the author was looking at master. The offending source files are not in Gingerbread: http://android.git.kernel.org/?p=platform/libcore.git;a=tree;f=support/src/test/java/org/apache/harmony/security/tests/support/acl;h=801cdcf96fa4e6c12398782b2e936dbcc311b6c8;hb=refs/heads/gingerbread http://android.git.kernel.org/?p=platform/libcore.git;a=tree...

brown9-2 16y ago

His response seems to be along the line of "how do we know the unit test code didn't get put on a device and shipped?": For this one as well as quote #3, what's missing is an analysis of the code actually shipped with Android-based devices. There are many such devices. Some may be more security-oriented (hence more likely to make use of the "acl" and PolicyNodeImpl type of classes) than others. So we're talking about a totally unsubstantiated claim here. There may be build scripts, and some files may not be in them by default. That doesn't mean that no maker of an Android-based device actually decided to build them into a product, especially if Apache license headers appeared to allow it. If this scenario did in fact occur, then isn't it the device manufacturer's fault and problem, and not Google's responsibility? Which party builds/packages the Android distributions for the actual phone devices that the OS will appear on?