[Huang-404-Q]

LineRanges::check returned AfterLastRange as soon as a line was past the from-end cutoff, without considering a larger absolute upper bound from another range. controller.rs then stopped reading, so a later absolute --line-range was never printed.

For a 10-line file:

bat multiline.txt --line-range=:-2 --line-range=10:10

before this change prints lines 1-8 and drops line 10. Ranges should be unioned, so line 10 must be printed as well.

The fix makes largest_absolute_upper_bound optional. The early exit still applies when there is no absolute upper bound (pure from-end ranges), but is skipped when a later line is still covered by an absolute range.

Tests

Added line_range_offset_from_end_and_later_absolute_range to tests/integration_tests.rs.

Verified with cargo +1.95.0-x86_64-pc-windows-gnu test --locked --test integration_tests:

• before: test fails, output is missing line 10
• after: test passes; all line-range unit and integration tests pass

[Huang-404-Q]

The i686-unknown-linux-gnu failure is in no_args_doesnt_break, which panics in wait-timeout-0.2.1/src/unix.rs:55 while spawning/waiting for the process. It does not touch line ranges; the line-range unit and integration tests passed (265 passed, 1 failed). This looks like the known 32-bit wait-timeout flake. The changelog entry has been added in 787e588.

[Huang-404-Q]

The rerun failed identically: no_args_doesnt_break panics in wait-timeout again (265 passed, 1 failed). The line-range tests pass on this target as well, so this is the same environment-specific wait-timeout flake rather than a regression from this change.