Most bugs don’t live in the middle of the input range. Nobody’s code breaks on “$12.34”. It breaks on “”, on “0”, on “12.5” read as twelve dollars and five cents, on " 7.10 " pasted with a trailing space. Finding those inputs on purpose is the skill this lesson teaches.
\nSplitwise-lite’s amount field accepts what people type and turns it into integer cents:
\nfunction parseAmount(input) {\n const text = String(input).trim();\n if (!/^\\d+(\\.\\d{1,2})?$/.test(text)) {\n throw new Error(`Not a valid amount: \"${input}\"`);\n }\n const [whole, fraction = ''] = text.split('.');\n const cents = Number(whole) * 100 + Number(fraction.padEnd(2, '0'));\n if (cents === 0) throw new Error('Amount must be greater than zero');\n return cents;\n}\n\nfor (const input of ['12', '12.5', ' 7.10 ', '0.01']) {\n console.log(JSON.stringify(input), '->', parseAmount(input));\n}\nfor (const input of ['', '0', '-5', '12.345', 'ten']) {\n try { parseAmount(input); } catch (e) { console.log(JSON.stringify(input), '->', e.message); }\n}You can’t test every string, and you don’t need to. Equivalence partitioning means grouping inputs the code should treat the same way, and testing one representative of each group:
\n'12''12.5' (the dangerous one: is it 1250 or 1205?)'12.50'' 7.10 ''ten', '', '-5'Boundary value analysis then adds the inputs at the edges between groups, because that is where < and <= get mixed up and where {1,2} becomes {1,3}. For each limit, test just below, on, and just above:
Here the boundaries are zero (rejected) versus one cent (accepted), and two decimals (accepted) versus three (rejected). A checklist I run through for any input:
\n'', ' ''0', '0.01', '12345.67''1,000')'12.')Writing a separate test(...) for each of those is noisy. test.each takes a table and produces one named test per row:
import { describe, test, expect } from 'vitest';\nimport { parseAmount } from './parse-amount.js';\n\ndescribe('parseAmount', () => {\n test.each([\n ['12', 1200],\n ['12.5', 1250],\n ['12.50', 1250],\n ['0.01', 1],\n [' 7.10 ', 710],\n ])('parses %s as %i cents', (input, cents) => {\n expect(parseAmount(input)).toBe(cents);\n });\n\n test.each(['', '0', '0.00', '-5', '12.505', 'ten', '1,000'])(\n 'rejects %s',\n (input) => {\n expect(() => parseAmount(input)).toThrow();\n },\n );\n});%s and %i are filled from each row, so the report reads parses 12.5 as 1250 cents, and a failure names the exact input. For wider tables, rows can be objects and the name can use $field:
test.each([\n { input: '12.5', cents: 1250 },\n { input: '0.01', cents: 1 },\n])('parses $input as $cents cents', ({ input, cents }) => {\n expect(parseAmount(input)).toBe(cents);\n});The real win is maintenance. When a user reports that '12.' crashes the form, the regression test is one new row, not a new block of code.
You don’t find edge cases by staring at the code. Three sources do most of the work:
\n'12,50' is not a typo for them; decide whether you reject it with a helpful message or accept it, and write the row either way.How many rows are enough? One per class, plus both sides of each boundary, plus one row per past bug. For parseAmount that is about fifteen rows, a few milliseconds of runtime and a very good night’s sleep. If you find yourself wanting thousands of random inputs, look at property-based testing libraries such as fast-check, which generate inputs and shrink failures to the smallest example; for most app code, a hand-picked table is clearer.
In the exercise, build a table that exposes four realistic parsing bugs. Next up: functions that throw and functions that return promises.
\n","toc":[{"id":"sec-1","text":"The function under test","level":2},{"id":"sec-2","text":"Equivalence classes and boundaries","level":2},{"id":"sec-3","text":"Table-driven tests with test.each","level":2},{"id":"sec-4","text":"Where edge cases come from","level":2}],"takeaways":["Group inputs into classes that the code should treat the same way, then test one typical value per class plus the edges between classes.","Bugs cluster at boundaries, so test just below, on and just above every limit (zero, the maximum length, the decimal limit).","test.each turns a table of inputs and expected outputs into one named test per row, so adding a case is one line.","Name each row with %s placeholders or $field so a failure tells you which input broke."],"further":[{"title":"Vitest — test.each","url":"https://vitest.dev/api/test#test-each"},{"title":"MDN — Regular expressions","url":"https://developer.mozilla.org/en-US/docs/Web/JavaScript/Guide/Regular_expressions"}],"quiz":[{"q":"parseAmount accepts up to two decimal places. Which set of inputs best tests that rule?
'12', '12.5', '12.50', '12.505'","why":"Correct. No decimals, one, two (on the limit) and three (just over) cover both sides of the boundary."},{"text":"'12.50', '99.99', '0.01'","why":"These are all in the same class (exactly two decimals), so a bug that also accepted three decimals would pass every one."},{"text":"'abc', '', '-5'","why":"Good invalid inputs, but none of them is near the decimal limit, so they can’t tell two decimals from three."}],"answer":0},{"q":"What are the names of the tests this produces?
\ntest.each([\n ['12', 1200],\n ['0.5', 50],\n])('parses %s as %i cents', (input, cents) => {\n expect(parseAmount(input)).toBe(cents);\n});parses %s as %i cents, twice","why":"The placeholders are filled in from each row, which is exactly what makes table-driven failures readable."},{"text":"parses 12 as 1200 cents and parses 0.5 as 50 cents","why":"Correct. %s takes the first value of the row and %i the second, giving one named test per row."},{"text":"One test named parses 12, 0.5 as 1200, 50 cents","why":"test.each creates one test per row, not one test for the whole table; each row passes or fails on its own."}],"answer":1},{"q":"A form’s amount field rejects zero. Which bug does a test with input '0' catch that '1' and '-1' miss?
'-1' already covers negatives; zero adds nothing there."},{"text":"A parser that forgets to trim spaces.","why":"Trimming is about whitespace; none of these three inputs has any."},{"text":"A check written as cents < 0 instead of cents <= 0, which lets zero through.","why":"Correct. Zero sits exactly on the boundary, and only an input on the boundary tells < from <=."}],"answer":2}],"exercise":{"kind":"web","mode":"test","subject":"// Parses what a user typed in the amount field and returns integer cents.\n// Accepts: digits, optionally a dot and 1 or 2 decimals, surrounding spaces.\n// \"12\" -> 1200, \"12.5\" -> 1250, \" 7.10 \" -> 710\n// Throws an Error for: empty input, zero, negatives, letters,\n// commas and more than 2 decimals.\nfunction parseAmount(input) {\n const text = String(input).trim();\n if (!/^\\d+(\\.\\d{1,2})?$/.test(text)) {\n throw new Error(`Not a valid amount: \"${input}\"`);\n }\n const [whole, fraction = ''] = text.split('.');\n const cents = Number(whole) * 100 + Number(fraction.padEnd(2, '0'));\n if (cents === 0) throw new Error('Amount must be greater than zero');\n return cents;\n}\n","mutants":[{"id":"three-decimals","code":"// Parses what a user typed in the amount field and returns integer cents.\n// Accepts: digits, optionally a dot and 1 or 2 decimals, surrounding spaces.\n// \"12\" -> 1200, \"12.5\" -> 1250, \" 7.10 \" -> 710\n// Throws an Error for: empty input, zero, negatives, letters,\n// commas and more than 2 decimals.\nfunction parseAmount(input) {\n const text = String(input).trim();\n if (!/^\\d+(\\.\\d{1,3})?$/.test(text)) {\n throw new Error(`Not a valid amount: \"${input}\"`);\n }\n const [whole, fraction = ''] = text.split('.');\n const cents = Number(whole) * 100 + Number(fraction.padEnd(2, '0'));\n if (cents === 0) throw new Error('Amount must be greater than zero');\n return cents;\n}\n"},{"id":"pads-wrong-side","code":"// Parses what a user typed in the amount field and returns integer cents.\n// Accepts: digits, optionally a dot and 1 or 2 decimals, surrounding spaces.\n// \"12\" -> 1200, \"12.5\" -> 1250, \" 7.10 \" -> 710\n// Throws an Error for: empty input, zero, negatives, letters,\n// commas and more than 2 decimals.\nfunction parseAmount(input) {\n const text = String(input).trim();\n if (!/^\\d+(\\.\\d{1,2})?$/.test(text)) {\n throw new Error(`Not a valid amount: \"${input}\"`);\n }\n const [whole, fraction = ''] = text.split('.');\n const cents = Number(whole) * 100 + Number(fraction.padStart(2, '0'));\n if (cents === 0) throw new Error('Amount must be greater than zero');\n return cents;\n}\n"},{"id":"accepts-zero","code":"// Parses what a user typed in the amount field and returns integer cents.\n// Accepts: digits, optionally a dot and 1 or 2 decimals, surrounding spaces.\n// \"12\" -> 1200, \"12.5\" -> 1250, \" 7.10 \" -> 710\n// Throws an Error for: empty input, zero, negatives, letters,\n// commas and more than 2 decimals.\nfunction parseAmount(input) {\n const text = String(input).trim();\n if (!/^\\d+(\\.\\d{1,2})?$/.test(text)) {\n throw new Error(`Not a valid amount: \"${input}\"`);\n }\n const [whole, fraction = ''] = text.split('.');\n const cents = Number(whole) * 100 + Number(fraction.padEnd(2, '0'));\n if (cents < 0) throw new Error('Amount must be greater than zero');\n return cents;\n}\n"},{"id":"no-trim","code":"// Parses what a user typed in the amount field and returns integer cents.\n// Accepts: digits, optionally a dot and 1 or 2 decimals, surrounding spaces.\n// \"12\" -> 1200, \"12.5\" -> 1250, \" 7.10 \" -> 710\n// Throws an Error for: empty input, zero, negatives, letters,\n// commas and more than 2 decimals.\nfunction parseAmount(input) {\n const text = String(input);\n if (!/^\\d+(\\.\\d{1,2})?$/.test(text)) {\n throw new Error(`Not a valid amount: \"${input}\"`);\n }\n const [whole, fraction = ''] = text.split('.');\n const cents = Number(whole) * 100 + Number(fraction.padEnd(2, '0'));\n if (cents === 0) throw new Error('Amount must be greater than zero');\n return cents;\n}\n"}],"starter":{"js":"describe('parseAmount', () => {\n test.each([\n ['12', 1200],\n ['12.50', 1250],\n ])('parses %s as %i cents', (input, cents) => {\n expect(parseAmount(input)).toBe(cents);\n });\n\n test.each(['', 'ten', '-5'])('rejects %s', (input) => {\n expect(() => parseAmount(input)).toThrow();\n });\n});\n"},"solution":{"js":"describe('parseAmount', () => {\n test.each([\n ['12', 1200],\n ['12.5', 1250],\n ['12.50', 1250],\n ['0.01', 1],\n [' 7.10 ', 710],\n ])('parses %s as %i cents', (input, cents) => {\n expect(parseAmount(input)).toBe(cents);\n });\n\n test.each(['', 'ten', '-5', '1,000', '12.505'])('rejects %s', (input) => {\n expect(() => parseAmount(input)).toThrow();\n });\n\n test.each(['0', '0.00'])('rejects a zero amount: %s', (input) => {\n expect(() => parseAmount(input)).toThrow('greater than zero');\n });\n});\n"},"checks":[{"id":"passes","suite":"subject"},{"id":"catches-three-decimals","suite":"three-decimals"},{"id":"catches-pads-wrong-side","suite":"pads-wrong-side"},{"id":"catches-accepts-zero","suite":"accepts-zero"},{"id":"catches-no-trim","suite":"no-trim"}],"title":"A table that breaks parseAmount","prompt":"parseAmount(input) turns what a user typed into integer cents. Its rules are in the subject’s comment.
The starter has two test.each tables that all pass but catch nothing interesting: every row sits safely inside a class. Add rows (and tables if you like) until your suite catches four hidden bugs: one accepts three decimals, one reads '12.5' as 1205 cents, one accepts zero, and one forgets to trim spaces.
'12.5' must be 1250, not 1205. Zero sits on the boundary between rejected and accepted: try '0'.","Add ['12.5', 1250] and [' 7.10 ', 710] to the valid table, and '12.505' and '0' to the invalid ones."],"explanation":"Every bug here hides at a boundary or in a class the starter never visited. '12.505' sits one step past the two-decimal limit, so only it can tell {1,2} from {1,3}. '12.5' is the one-decimal class, where padding on the wrong side turns $12.50 into $12.05. '0' sits exactly on the zero boundary, the only input that separates === 0 from < 0. ' 7.10 ' is formatting noise that users paste every day. Matching the message ('greater than zero') in the zero test also proves the right rule rejected it, not a regex accident.