Summary: | dev-python/pandas-1.2.1 fails test TestTimestampProperties.test_is_leap_year | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | Thomas Deutschmann (RETIRED) <whissi> |
Component: | Current packages | Assignee: | Gentoo Science Related Packages <sci> |
Status: | CONFIRMED --- | ||
Severity: | normal | CC: | mattst88, mgorny, python |
Priority: | Normal | Keywords: | TESTFAILURE |
Version: | unspecified | ||
Hardware: | All | ||
OS: | Linux | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Bug Depends on: | |||
Bug Blocks: | 772500 |
Description
Thomas Deutschmann (RETIRED)
![]() What's your pytest version? Where's the build log? *** Bug 779394 has been marked as a duplicate of this bug. *** The file you are trying to attach is 1465 kilobytes (KB) in size. Attachments cannot be more than 1000 KB. We recommend that you try to compress the file by using bzip2, gzip, tar or lzma alternatively you may want to store your attachment elsewhere and then paste the URL to this file on the attachment creation page in the appropriate text field, which you can access by clicking the "paste text as attachment" link. https://0x0.st/-be7.gz pytest --version pytest 6.2.2 > dt = Timestamp("2100-01-01 00:00:00", tz=tz) > ... > return time.localtime(timestamp + time.timezone).tm_isdst > E OverflowError: timestamp out of range for platform time_t Looks like you're building on a 32-bit x86 system. time_t is probably 32-bit and thus is unable to represent a date of 2100. > assert q1 == q2 > E assert 1.7000000000000002 == 1.7 > E +1.7000000000000002 > E -1.7 This sort of stuff looks like x87 floating-point precision differences. Basically unsolvable. Yes, x86 (32bit). I've tried to reproduce this in a 32-bit chroot, and something caused the OverflowErrors to be ignored. I'm wondering why they aren't ignored on your system. Kernel version differences? https://lkml.org/lkml/2020/1/29/355 says > As a result, linux-5.6, or my backport of the patches to 5.4 [1], should > be the first release that can serve as a base for a 32-bit system designed > to run beyond year 2038, with a few remaining caveats: > > - All user space must be compiled with a 64-bit time_t, which will be > supported in the coming musl-1.2 and glibc-2.32 releases, along with > installed kernel headers from linux-5.6 or higher. I see from Whissi's emerge --info that he's running v5.4. That's why I am not a fan of running stabilization in chroots: You aren't running with all arch limitations. QEMU/type 2 hypervisor is minimum requirement for proper arch testing if you don't have access to the hardware. (In reply to Matt Turner from comment #7) > Kernel version differences? https://lkml.org/lkml/2020/1/29/355 says > > > As a result, linux-5.6, or my backport of the patches to 5.4 [1], should > > be the first release that can serve as a base for a 32-bit system designed > > to run beyond year 2038, with a few remaining caveats: > > > > - All user space must be compiled with a 64-bit time_t, which will be > > supported in the coming musl-1.2 and glibc-2.32 releases, along with > > installed kernel headers from linux-5.6 or higher. > > I see from Whissi's emerge --info that he's running v5.4. Doubt it. The message explicitly said that the OverflowError did happen and the whole test was skipped because of it. What's your dev-python/cython version? |