[rudycelekli]

Why

A MySQL TIME is a signed duration from -838:59:59 to 838:59:59, not a time of day. TIMEDIFF, SEC_TO_TIME and TIME columns that hold elapsed hours all return it. The MySQL engine read it as chrono::NaiveTime, and that read has two problems.

• 24 hours or more. sqlx refuses the conversion. The text fallback cannot decode TIME either, so the cell is null. TIMEDIFF('2024-01-02 06:15:00', '2024-01-01 00:00:00') came back as null instead of 30:15:00.
• Negative values. The conversion drops the sign. A delivery that was 1h30 early, -01:30:00, came back as 01:30:00, so the model reads it as 1h30 late.

What changes

Cell::Time reads sqlx::mysql::types::MySqlTime, which covers the full range. It is written the way MySQL prints it: an optional -, zero-padded HH:MM:SS, and .ffffff only when there are microseconds. Ordinary times of day, such as 08:05:00, read the same as before.

The sign is taken from sign(). In sqlx 0.8.6, MySqlTime::is_negative() returns sign.is_positive(). Using it would invert every value, and the test catches that.

How it was checked

• A new live test, a_time_past_a_day_or_below_zero_reads_as_itself_mysql, selects two TIMEDIFFs and CAST('08:05:00' AS TIME). It expects {"late":"30:15:00","early":"-01:30:00","start":"08:05:00"}.
• UTOPIA_TEST_MYSQL_URL=mysql://root:pw@127.0.0.1:13306/sales cargo test -p utopia-server --bins query_engine::mysql, run against mariadb:11.4: 8 passed.
• With the fix reverted, the test fails with left: ["{\"late\":null,\"early\":\"01:30:00\",\"start\":\"08:05:00\"}"].
• cargo fmt --all --check and cargo clippy -p utopia-server --all-targets -- -D warnings pass.

Before review

• Every commit is signed off (git commit -s)
• cargo fmt --all --check, cargo clippy and cargo test pass for the touched crate

[WaylandYang]

The fix looks right, and good catch on is_negative. CI has no MySQL, so the live test never runs there. Please add a unit test of time_text built with MySqlTime::new (covering >24h, negative and microseconds). Please also keep chrono's fractional style (.500, not .500000) so existing output doesn't change.

[rudycelekli]

Rebased on dev after #1410 and #1413. time_text now writes fractions the way chrono NaiveTime does (.500, and .000500 only when it is not whole milliseconds), so output for a TIME within a day does not change. A new unit test, a_time_is_written_as_the_duration_it_holds, builds MySqlTime::new values over 24 h, negative and fractional, so it runs in CI without MySQL.