Failed test: BytesTest.testCopyToWithUnicode

[ctubbsii]

Recently fixed Jenkins builds (was pointing to old git repo, prior to move to gitbox) and saw a failed test for org.apache.fluo.api.data.BytesTest.testCopyToWithUnicode:

Error Message

expected:<15> but was:<10>
Stacktrace

java.lang.AssertionError: expected:<15> but was:<10>
	at org.apache.fluo.api.data.BytesTest.testCopyToWithUnicode(BytesTest.java:339)

[cjmctague]

Interesting. Which encoding is Jenkins using? The test called .getBytes() which is using the default system encoding. The result of 15 is assuming the encoding is UTF-8 I believe. Locally I'm getting the result of 15 and the only way I can produce a result of 10 is by using the encoding .getBytes("CP1252")

[ctubbsii]

CP1252 appears to be the default for Windows. Jenkins may have some builders which run on Windows.

[ctubbsii]

Our Jenkins jobs are configured to run on "Hadoop||ubuntu" nodes, which should all be Linux, so I'm really not sure why this is happening. Either way, the test could be fixed so that line 334 call to getBytes() passes in StandardCharsets.UTF_8.

[cjmctague]

Error Message

expected:<c[†𝔊]e> but was:<c[â� ð���]e>
Stacktrace

org.junit.ComparisonFailure: expected:<c[†𝔊]e> but was:<c[â� ð���]e>
	at org.apache.fluo.api.data.BytesTest.testCopyToWithUnicode(BytesTest.java:343)

Broke another test with the fix in 3302fd6
https://builds.apache.org/view/F/view/Fluo/job/Fluo/org.apache.fluo$fluo-api/138/testReport/org.apache.fluo.api.data/BytesTest/testCopyToWithUnicode/