test(time): expand is_rfc3339_fulltime coverage for offset, secfrac, and leap-second branches

[vtushar06]

I added 157 tests to test/time/rfc3339_fulltime_test.cc. Core already passes all of
them, so this is coverage, not a fix. I organised them by the branches in
src/core/time/rfc3339_fulltime.cc rather than by input shape, so every if in the
scanner has assertions on both sides.

What is new

• length guard (size < 9) - every length from 0 to 8, plus the 9-character
  minimum and the 8-character "valid partial-time, no offset" case.
• time-hour - 00, 01, 09, 10, 19, 20, 23 valid and 24, 25, 29, 30, 59, 90, 99
  invalid, plus a non-digit at each of the two positions individually.
• time-minute / time-second - the same treatment: the range boundary either
  side, and each digit position broken on its own.
• the second digit of each 2DIGIT field, specifically. A scanner that checks only
  the first digit and then computes (a - '0') * 10 + (b - '0') is not caught by an
  alphabetic probe, because 'a' - '0' is 49 and the field lands far out of range and
  is rejected anyway. The characters just above '9' are the ones that land it back
  INSIDE the range: ':' - '0' is 10, ';' is 11, '<' is 12. Those are the inputs
  that distinguish "checks both digits" from "checks the first".
• both : separators, each on its own - 10 wrong characters at position 2 alone
  and at position 5 alone, so neither check is load-bearing for the other.
• time-secfrac - the .-then-1*DIGIT rule from both sides: a single digit, an
  empty fraction with and without an offset, all-zero, 9 / 30 / 300 digits, a double
  dot, two fractions, a comma instead of a dot, a non-ASCII digit, and a sign.
• time-offset - Z, z, missing, and six wrong characters; plus trailing
  content after Z.
• time-numoffset - offset hour 23 valid and 24 / 99 invalid, offset minute 59
  valid and 60 / 99 invalid, the required : (including the colonless +0000 form),
  one-digit hour and minute, a truncated offset, a double sign, a dot separator,
  trailing content, and +19:00 (grammar-legal, beyond the usual ±18h convention).
• trailing and leading terminators - LF, CR, CRLF, TAB, space, NUL, NBSP,
  U+2028 and a BOM, in both positions.
• the section 5.7 leap-second rule - 23:59:60Z valid, and the offset arithmetic
  from both directions: 00:59:60+01:00 and 22:59:60-01:00 valid, the wrap-around
  cases, a leap second at a UTC time that is not 23:59 invalid, with a fraction, and
  at the maximum offset.

How I checked the expected values

I did not write the expected verdicts by hand. Every input was run through the real
is_rfc3339_fulltime, compiled straight from this branch's source, and the
EXPECT_TRUE / EXPECT_FALSE was emitted from what the function actually returned.
The whole file was then compiled and run again to confirm. sourcemeta_core_time_unit
reports 585 passed, 0 failed. clang-format is clean and the diff is additions only.

Eleven inputs contain an embedded NUL and are emitted as
std::string_view("...", N) rather than a plain literal, because a plain literal
truncates at the NUL and the test would silently assert on a shorter input. Where a
\x00 escape is followed by a hex digit the literal is also split - "\x00" "8:30:06Z"
rather than "\x008:30:06Z", since C++ hex escapes are greedy and the latter is the
single byte 0x08.

RFC references

• RFC 3339 section 5.6 - full-time = partial-time time-offset
• RFC 3339 section 5.7 - the time-second leap rule and the zone-offset shift
• RFC 5234 section 3.4 - DIGIT = %x30-39

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.

[augmentcode]

🤖 Augment PR Summary

Summary: This PR expands unit-test coverage for sourcemeta::core::is_rfc3339_fulltime, focusing on exercising every parsing/validation branch (including edge cases) without changing production behavior.

Changes:

• Adds a large set of new assertions to test/time/rfc3339_fulltime_test.cc (additions only).
• Covers the minimum-length guard (size < 9) across all short lengths.
• Expands validation coverage for time-hour, time-minute, and time-second ranges and non-digit probes (including tricky second-digit probes like ; / <).
• Adds targeted tests for both : separators, missing separators, and embedded NUL bytes via std::string_view.
• Extends coverage for time-secfrac (empty fraction, multiple fractions, very long fractions, non-ASCII digits, sign in fraction, etc.).
• Expands coverage for time-offset and numeric offsets (invalid forms, truncation, trailing content, Unicode minus, colonless offsets).
• Adds leading/trailing terminator/whitespace probes (CR/LF/CRLF/TAB/space/NUL/NBSP/U+2028/BOM).
• Strengthens leap-second coverage, including offset arithmetic and wraparound cases.

Technical Notes: Expected results are snapshots of the current implementation’s return values (coverage-focused); no changes to src/core/time/rfc3339_fulltime.cc are included.

_{🤖 Was this summary useful? React with 👍 or 👎}

[augmentcode]

Review completed. No suggestions at this time.

Comment augment review to trigger a new review at any time.

[cubic-dev-ai]

No issues found across 1 file

_{Re-trigger cubic}