Encoding errror on s390x

[voxik]

Running sqlite3 test suite on s390x, I observe the following errors:

$ ruby -I~/build/BUILD/rubygem-sqlite3-2.5.0-build/sqlite3-2.5.0/usr/lib64/gems/ruby/sqlite3-2.5.0:lib:test -e 'Dir.glob "./test/test_*.rb", &method(:require)'
---
:ruby: ruby 3.4.1 (2024-12-25 revision 48d4efcb85) +PRISM [s390x-linux]
:gem:
  :version: 2.5.0
:sqlite:
  :compiled: 3.47.2
  :loaded: 3.47.2
  :packaged: false
  :precompiled: false
  :sqlcipher: false
  :threadsafe: true
Run options: --seed 4291
# Running:
.............................................................................................................................................S.......................S.......................................................................................................................................F..FF..................
Finished in 1.602496s, 202.1845 runs/s, 366.3035 assertions/s.
  1) Failure:
SQLite3::TestPragmas#test_encoding_uppercase [test/test_pragmas.rb:81]:
Expected: #<Encoding:UTF-16LE>
  Actual: #<Encoding:UTF-16BE>
  2) Failure:
SQLite3::TestPragmas#test_encoding_lowercase [test/test_pragmas.rb:97]:
Expected: #<Encoding:UTF-16LE>
  Actual: #<Encoding:UTF-16BE>
  3) Failure:
SQLite3::TestPragmas#test_encoding_objects [test/test_pragmas.rb:113]:
Expected: #<Encoding:UTF-16LE>
  Actual: #<Encoding:UTF-16BE>
324 runs, 587 assertions, 3 failures, 0 errors, 2 skips
You have skipped tests. Run with --verbose for details.

These test cases were introduced by #575, however I assume that the issue is just exposed by the tests.

[voxik]

Looking closer at the test case, I would not be surprised if "UTF-16" is interpreted as Encoding::UTF_16BE on BE machines.

[flavorjones]

Thanks, yes in those cases sqlite interprets it as the native endian encoding. Should be fixed in #616