RFR: 8252730: jlink does not create reproducible builds on different servers [v9]

[Ian Graves]

On Sat, 19 Sep 2020 16:04:36 GMT, Alan Bateman <alanb at openjdk.org> wrote:

>> Ian Graves has updated the pull request incrementally with one additional commit since the last revision:
>> 
>>   Using File.walk for copy
>
> test/jdk/tools/jlink/JLinkReproducible3Test.java line 112:
> 
>> 110:     }
>> 111: }
>> 112:
> 
> The update to ImageFileCreator looks good.
> 
> A few nits in JLinkReproducible3Test:
> 1. Invoking copyJDKs with two destination locations looks a bit strange. I think would be easier for future maintainers
> if it does one copy and is called twice. 2. Files.copy will create a directory, no need for Files.createDirectories.
> 3. Are you sure that bin/jlink is executable, I just wonder if the test runs with the patch have ended up depending the
> umask at the time. It is possible to specify COPY_ATTRIBUTES to the copy method and it will attempt to copy the
> permissions. 4. The variable naming is more like C style, so a bit inconsistent with the naming usually used in Java
> programs.

1. I agree that it's awkward and I'm not a huge fan. I'm okay doing it this way, but this is recreating the reproducer
for https://bugs.openjdk.java.net/browse/JDK-8252730 which is two identical JDK's in different directories. I can
document this in the test further or we can change it.

2-4: Noted. Will clean these up and loop back. Thanks!

-------------

PR: https://git.openjdk.java.net/jdk/pull/156