Fix integer overflow when parsing SOURCE_DATE_EPOCH

[mizdebsk]

Parsing SOURCE_DATE_EPOCH environment variable by Tstamp leads to integer overflow and produces wrong result.

The following example illustrates that:

<tstamp/>
<echo>TODAY is ${TODAY}</echo>

This produces the following output:

[tstamp] Honouring environment variable SOURCE_DATE_EPOCH which has been set to 1650585600
[echo] TODAY is January 16 1970

Which is incorrect, because timestamp 1650585600 corresponds to April 22 2022.

[jaikiran]

Hello @mizdebsk, thank you for contributing this fix.

Would you be interested in adding a test case for this change so that this doesn't regress in future. We have an existing test class for this task which you can add a new test to here https://github.com/apache/ant/blob/master/src/tests/junit/org/apache/tools/ant/taskdefs/TStampTest.java.

If you won't be able to contribute the test, that's fine too. Just let us know.

[mizdebsk]

Sure, I will add a test case.

[mizdebsk]

I've added a test case that demonstrates the issue.
Test case is more complex than the actual fix.
Does it mean I have to sign Apache CLA? I have not signed it yet.

The test case is implemented as an AntUnit test. JUnit test case would not work because the test requires setting environment variable SOURCE_DATE_EPOCH, which must be done by forking JVM - the is no setenv in Java.

Without the fix the test fails in the following way:

Target: testSourceDateEpoch  FAILED
    at line 107, column 77
    Message: Expected 'TODAY is April 22 2022' but was 'TODAY is January 16 1970'
    took 0.166 sec

And it passes with the fix:

Target: testSourceDateEpoch took 0.164 sec

[kleini]

We're having an issue, with this, too.

[tstamp] Honouring environment variable SOURCE_DATE_EPOCH which has been set to 1651903417
Date of 12/12/1969 07:10 PM results in negative milliseconds value relative to epoch (January 1, 1970, 00:00:00 GMT).

[jaikiran]

I've added a test case that demonstrates the issue.
Test case is more complex than the actual fix.
Does it mean I have to sign Apache CLA? I have not signed it yet.

Hello @mizdebsk, I just saw this message. The contributor guidelines don't state that you are expected to sign the CLA for the contribution https://infra.apache.org/contributors.html. However, I'll check internally once.

Thank you for that test case. It looks fine to me. I'll run it once internally and merge it (after the CLA decision).

[jaikiran]

I've received confirmation that CLA is not a must. I've also tested your changes in this PR and they look good and I've now merged it. Thank you for this fix and the test case and apologies for the delay in merging.