[ORC-1661] [C++] Better handling when TZDB is unavailable

• Type: Improvement
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 2.1.0
• Component/s: C++
• Assignee: Gang Wu
• Reporter: Gang Wu
• Created: 21/Mar/24 01:18
• Updated: 10/May/24 18:13
• Resolved: 23/Apr/24 02:35

Description

When /usr/share/zoneinfo is unavailable and TZDIR env is unset, creating C++ ORC reader will crash on Windows. We need to better deal with this case. See context from the Apache Arrow community: https://github.com/apache/arrow/issues/36026 and https://github.com/apache/arrow/issues/40633

 

We could perhaps do following things:

• Make sure it does not crash when TZDB is missing (on Windows). The prerequisite is to enable running C++ unit tests on Windows.
• More TZDB search locations. (e.g. $CONDA_PREFIX/usr/share/zoneinfo)
• Do not eagerly reading TZDB when timestamp types are not required.
• Add a runtime config to set TZDB location.
• Ship TZDB with the ORC library while building on Windows?

Sub-Tasks

• [C++] Enable TestTimezone.testMissingTZDB on Windows: Closed
• [C++] Find tzdb without TZDIR when in conda-environments: Closed
• [C++] Avoid using std::filesystem: Closed
• [C++] Do not access TZDB if there is no timestamp type: Closed