Artifact [bb284f4363]

Artifact bb284f43638b11886172be12d5a6c65a168d799c9267f27c84dfa966e5ad3e8c:

• File internal/fsltype/julian_test.go — part of check-in [2c2e5e74b6] at 2026-10-09 15:35:38 on branch trunk — Store check-in, link and tag rows as fossil does (#258) (#261)

  * fix(manifest): store check-in, link and tag rows as fossil does (#258)

  A rebuilt or locally written repository differed from fossil's in three derived values: event and plink times in their last digits, event.omtime left NULL, and a valueless tag stored as '' instead of NULL.

  Fossil stores a tag time as SQLite's julianday() bound as a double, but writes event, plink, attachment and forumpost times into SQL text with its own printf("%.17g"), which emits 16 digits in long double arithmetic. TimeToJulian is now julianday() exactly (whole milliseconds over 86400000) and TimeToTextJulian ports fossil's printf conversion, so both match fossil bit for bit. The conversion follows builds where long double is a double (macOS, Windows); fossil itself differs across platforms in the last digit.

  Check-in events record omtime, tag writes bind NULL for an empty value, the readers of event and plink times under the ncruces driver rebuild the text form, and the CLI's tag listing reads NULL values.

  * test(db): sweep both stored julian forms; scan the text form from driver times (#258)

  * test(verify): compare text times bit for bit only where fossil computes them in double (#258) (user: [email protected], size: 2302) [annotate] [blame] [check-ins using]

package fsltype

import (
	"testing"
	"time"
)

// fossilEventTimes are event.mtime values the fossil 2.28 binary (macOS arm64)
// stored, with the D-card time of each in milliseconds: the first six from a
// `fossil rebuild` of the althttpd repository, the last two from commits.
var fossilEventTimes = []struct {
	millis int64
	stored float64
}{
	{1606831156077, 2459185.082824966}, // 2020-12-01T13:59:16.077
	{1606831367825, 2459185.085275753}, // 2020-12-01T14:02:47.825
	{1607520762942, 2459193.064385903}, // 2020-12-09T13:32:42.942
	{1606831563360, 2459185.087538889}, // 2020-12-01T14:06:03.360
	{1607520969888, 2459193.066781112}, // 2020-12-09T13:36:09.888
	{1607779347305, 2459196.057260475}, // 2020-12-12T13:22:27.305
	{1791557956502, 2461323.124496551}, // 2026-10-09T14:59:16.502
	{1791557956580, 2461323.124497454}, // 2026-10-09T14:59:16.580
}

// TextJulianDayFromMillis reproduces the doubles fossil stores for event
// times, bit for bit.
func TestTextJulianDayMatchesFossil(t *testing.T) {
	for _, c := range fossilEventTimes {
		if got := TextJulianDayFromMillis(c.millis); got != c.stored {
			t.Errorf("TextJulianDayFromMillis(%d) = %.17g, fossil stored %.17g",
				c.millis, got, c.stored)
		}
	}
}

// JulianDayFromMillis is SQLite's julianday(): whole milliseconds over
// 86400000. The first althttpd time is one where it and the text form
// differ, which is the difference #258 reported.
func TestJulianDayIsSQLiteJulianday(t *testing.T) {
	const millis = 1606831156077
	want := float64(millis+210866760000000) / 86400000.0
	if got := JulianDayFromMillis(millis); got != want {
		t.Fatalf("JulianDayFromMillis = %.17g, want %.17g", got, want)
	}
	if JulianDayFromMillis(millis) == TextJulianDayFromMillis(millis) {
		t.Fatal("test case does not tell the exact and text forms apart")
	}
}

// Both forms convert back to the same millisecond, so the timeline's
// (mtime, rid) cursor still round-trips.
func TestJulianFormsRoundTrip(t *testing.T) {
	for _, c := range fossilEventTimes {
		want := time.UnixMilli(c.millis).UTC()
		forms := []float64{JulianDayFromMillis(c.millis), TextJulianDayFromMillis(c.millis)}
		for _, jd := range forms {
			if got := JulianToTime(jd); !got.Equal(want) {
				t.Errorf("JulianToTime(%.17g) = %v, want %v", jd, got, want)
			}
		}
	}
}