test failures on macOS 10.12

[chdiza]

When trying to build from master on 10.12.4, I get the following test failure: https://gist.github.com/chdiza/762c86adaf40bfef2bdd889485eaa547

The test-suite.log is here: https://gist.github.com/chdiza/a28e34301fe058b8759ce092a525a704

This happens whether or not I build against the enclosed oniguruma.

[nicowilliams]

That's... strptime() being broken on OS X. Fantastic. I guess I get to write code to deal with that. Or not.

[nicowilliams]

We already have an issue open for OS X strptime() issues, #929.

See also http://stackoverflow.com/questions/31984002/why-does-strptime-behave-differently-on-osx-and-on-linux. The OS X man page for strptime is woefully brief. It seems that we need to compute tm_wday and tm_yday ourselves.

I don't think I'll be fixing this anytime soon, though I'll probably paper this over in the test.

[nicowilliams]

Ahhh, right, timegm() will set tm_wday and tm_yday.

[nicowilliams]

@chdiza I've pushed an attempt to fix this. Can you please test.

[chdiza]

@nicowilliams Thanks. I tested it and I still get the same test failure, with the same test-suite.log.

[nicowilliams]

@chdiza I've pushed c538237 to try once more to fix this. Can you check it?

[chdiza]

@nicowilliams Getting closer, at least! The same test fails, but this time the testsuite.log says something different:

====================================================
   jq 1.5rc2-243-gc538237-dirty: ./test-suite.log
====================================================

# TOTAL: 7
# PASS:  6
# SKIP:  0
# XFAIL: 0
# FAIL:  1
# XPASS: 0
# ERROR: 0

.. contents:: :depth: 2

FAIL: tests/optionaltest
========================

Testing '[strptime("%Y-%m-%dT%H:%M:%SZ")|(.,mktime)]' at line number 4
Testing 'last(range(365 * 199)|("1900-03-01T01:02:03Z"|strptime("%Y-%m-%dT%H:%M:%SZ")|mktime) + (86400 * .)|strftime("%Y-%m-%dT%H:%M:%SZ")|strptime("%Y-%m-%dT%H:%M:%SZ"))' at line number 9
*** Insufficient results for test at line number 11: last(range(365 * 199)|("1900-03-01T01:02:03Z"|strptime("%Y-%m-%dT%H:%M:%SZ")|mktime) + (86400 * .)|strftime("%Y-%m-%dT%H:%M:%SZ")|strptime("%Y-%m-%dT%H:%M:%SZ"))
Testing 'strftime("%A, %B %e, %Y")' at line number 14
2 of 3 tests passed (0 malformed)
FAIL tests/optionaltest (exit status: 1)

[chdiza]

Hopefully that's an easy fix. If not, instead of spending even more time on this it'd be just fine with me if this particular test is simply skipped on macOS.

[nicowilliams]

@chdiza Yeah, maybe.

Can you show us the output of this:

$ jq -n 'last(range(365 * 199)|("1900-03-01T01:02:03Z"|strptime("%Y-%m-%dT%H:%M:%SZ")|mktime) + (86400 * .)|strftime("%Y-%m-%dT%H:%M:%SZ")|strptime("%Y-%m-%dT%H:%M:%SZ"))' 

I mean, the test log implies it output nothing. But I'm curious if there were any errors.

[chdiza]

The output is:

jq: error (at <unknown>): invalid gmtime representation

[yoichi]

c538237 breaks 1st test of optional.test on macos

$ echo '"2015-03-05T23:51:47Z"' | jq 'strptime("%Y-%m-%dT%H:%M:%SZ")|strftime("%A, %B %e, %Y")'
"Wednesday, March  5, 2015"

It's Thursday, correctly.