The sheet package still finishes well before the UI's tests.
- commit
- 3148d18100c3649ed66cb2e79575f84c65723be9
- parent
- b3554d3d8a1c76e4c585cd1bf55536c62740ac07
- author
- Dami <dami.etoile@gmail.com>
- kind
- test risk score 0.7 of 3, Jev's estimate
| M | docs/ | 4 | ++-- |
| M | internal/ | 4 | ++-- |
@@ -29,7 +29,7 @@ saved and saves them again, byte for byte
2929 ([File fixtures](releasing.md#file-fixtures)).3030 3131 `TestRandomEdits` (`internal/sheet/randedit_test.go`) checks what holds32-between features: from each of a thousand seeds it builds a workbook32+between features: from each of 3,000 seeds it builds a workbook3333 of two sheets and a notebook tab, makes up to 40 edits drawn from the3434 seed (entries, pastes, fills, series, sorts, inserted and deleted rows3535 and columns, formats, borders, merges, moves, notes, rules, names,@@ -43,7 +43,7 @@ scratch), and recalculating everything in place changes no value.
4343 Outputs' rows come from outside the undo history, so the test sends4444 them again after each edit, undo and reopening, as the UI does. A4545 failure names the fewest of the seed's edits that still fail. It takes46-about a second; `-randedit.seeds=20000` runs more, and46+a few seconds; `-randedit.seeds=20000` runs more, and4747 `FuzzRandomEdits` (in `make fuzz`) lets the fuzzer choose the edits.4848 4949 `TestRandomSharedEdits` (`randedit_shared_test.go`) makes the same@@ -26,10 +26,10 @@ import (
2626 // recalculation agreeing with the incremental one);2727 // - recalculates in place, everything at once, to the same values.2828 //29-// go test runs a fixed set of seeds in a second or two; -randedit.seeds29+// go test runs a fixed set of seeds in a few seconds; -randedit.seeds3030 // runs more, and FuzzRandomEdits lets the fuzzer choose the edits.3131 var (32- randSeeds = flag.Int("randedit.seeds", 1000, "seeds TestRandomEdits runs")32+ randSeeds = flag.Int("randedit.seeds", 3000, "seeds TestRandomEdits runs")3333 randSteps = flag.Int("randedit.steps", 40, "edits per seed")3434 )3535