DEV Community: Hirantha Sandaruwan Lakshan

Sri Lankan public holidays and Poya days as open data, and two date bugs to avoid

Hirantha Sandaruwan Lakshan — Thu, 29 Oct 2026 02:53:48 +0000

If you build anything for Sri Lanka (payroll, delivery dates, booking slots, a school timetable) you need the holiday list. It can't be computed. The full moon Poya days follow the lunar calendar, other dates move with religious calendars, and the government publishes the list for each year by gazette.

I needed it for my own tools, so I put it in one clean dataset and made it free to use. Here is what's in it, and two mistakes that are easy to make once you start doing date maths with it.

Where the data comes from

The source is the Central Bank of Sri Lanka's published bank-holiday lists for 2026 and 2027. Those follow the holidays gazetted under the Holidays Act, No. 29 of 1971 (the 2027 list is Extraordinary Gazette No. 2493/05). Every row was checked against the Central Bank's table, including the weekday it prints.

2027 has 25 holidays: 12 of them are Poya days and 8 are also mercantile holidays. 2026 has 26, with 13 Poya days.

The files

The repo is HiranthaSandaruwan/sri-lanka-public-holidays. Each holiday looks like this:

{
  "date": "2027-01-22",
  "weekday": "Friday",
  "name": "Duruthu Full Moon Poya Day",
  "poya": true,
  "mercantile": false,
  "bank": true,
  "public": true
}

poya is true for the Poya day itself, not for "the day following Vesak Poya", which is a separate holiday.
mercantile means it is also a mercantile holiday, the one that applies to private-sector employees.
bank and public are true for every row. Saturdays and Sundays are bank holidays too. A plain weekend isn't listed, but a holiday that falls on one is (2026-01-03, a Poya day, is a Saturday).

There is one file per year (data/holidays-2026.json, data/holidays-2027.json), data/holidays.json with all years, and data/holidays.csv:

date,weekday,name,poya,mercantile,bank,public
2026-01-03,Saturday,Duruthu Full Moon Poya Day,true,false,true,true
2026-01-15,Thursday,Tamil Thai Pongal Day,false,true,true,true

The calendars folder has the same list as three .ics files (every holiday, Poya days only, mercantile holidays only). They are all-day events marked as free time, so importing one into a calendar app won't make you look busy.

Reading it straight from GitHub:

const url = "https://raw.githubusercontent.com/HiranthaSandaruwan/sri-lanka-public-holidays/main/data/holidays-2027.json";
const holidays = await (await fetch(url)).json();

const offDays = new Set(holidays.map((h) => h.date));
const isHoliday = (iso) => offDays.has(iso);

A Set of dates is better than counting rows, because two holidays can share a day. On 2026-05-01, Vesak Full Moon Poya Day and May Day fall together, so 2026 has 26 rows but 25 different dates.

Bug 1: "today" in UTC is the wrong day

Sri Lanka is UTC+5:30 all year, with no daylight saving. A server running in UTC that does new Date().toISOString().slice(0, 10) gets the UTC date, which is still yesterday between midnight and 05:30 in Colombo.

That is enough to get a holiday wrong. At 01:30 on 24 November 2026 in Colombo, it is 20:00 on the 23rd in UTC. The 24th is Il Full Moon Poya Day, so the server would say "tomorrow is a holiday" on the holiday itself. Ask for the date in Sri Lanka's time zone instead:

const todayInColombo = () =>
  new Intl.DateTimeFormat("en-CA", { timeZone: "Asia/Colombo" }).format(new Date());

// The files are sorted by date, and ISO dates compare correctly as plain strings.
const next = holidays.find((h) => h.date >= todayInColombo());

The en-CA locale formats a date as YYYY-MM-DD, the same shape as the date field.

Bug 2: counting working days

A common question is how many working days fall between two dates. Walk the days, skip Saturdays, Sundays and holidays, and do the walking in UTC:

const DAY = 86_400_000;
const utc = (iso) => Date.parse(`${iso}T00:00:00Z`);

function workingDays(start, end, holidays) {
  const off = new Set(holidays.map((h) => h.date));
  let n = 0;
  for (let t = utc(start); t <= utc(end); t += DAY) {
    const d = new Date(t);
    const weekend = d.getUTCDay() === 0 || d.getUTCDay() === 6;
    if (!weekend && !off.has(d.toISOString().slice(0, 10))) n++;
  }
  return n;
}

workingDays("2027-04-01", "2027-04-30", holidays);                               // 19
workingDays("2027-04-01", "2027-04-30", holidays.filter((h) => h.mercantile));   // 20

Why UTC, when Sri Lanka has no daylight saving? Because the machine running the code might. If you build local dates with new Date(2027, 3, 1) and add 24 hours at a time, a clock change in the server's own time zone can push one step to 23:00 on the day before, and a date gets counted twice or skipped. UTC midnights never move.

The two answers differ because of who you count for. April 2027 has three holidays on weekdays: the day before the Sinhala and Tamil New Year (Tuesday 13th), New Year Day itself (Wednesday 14th) and Bak Full Moon Poya Day (Tuesday 20th). Bak Poya is a public and bank holiday but not a mercantile one, so for private-sector staff the month has one working day more. If you build payroll or leave tools, filter on mercantile rather than using every row.

Licence

The dataset is CC BY 4.0: use it for anything, including commercial work, with a credit link. The government can still amend the list by notification, so check the Sri Lanka holidays page before relying on a date months ahead.

If you spot a wrong date, open an issue on the repo. Which other Sri Lankan data would you like to have as a clean file like this?

This article was written with the help of AI. The code, numbers and sources come from my own project and were checked against it before publishing.

Turning Sri Lanka's expressway toll gazette into a lookup table

Hirantha Sandaruwan Lakshan — Wed, 28 Oct 2026 23:31:28 +0000

Sri Lanka's expressway tolls are set by gazette. The current tables for the Southern, Outer Circular and Katunayake expressways (E01, E02, E03) are in Gazette Extraordinary No. 2467/49 of 17 December 2025. The Central Expressway section (E04) has its own, No. 2329/17 of 24 April 2023. The Road Development Authority's expressway division publishes them on its toll rates page.

They are hard to read on a phone. You need the row for one interchange, the column for the other and the right table for your vehicle. So I turned them into a small lookup and a toll calculator. Here is what the data looks like in code and how I checked it.

The shape of the table

Each vehicle category has one table. The interchanges run down the side and along the top in the same order, from Katunayake (B4) to Hambantota Gate (A21), 29 in all. The fee from A to B is the same as from B to A, so the gazette prints only the lower half: a triangle.

That maps directly to an array of rows, where row n holds the fees from interchange n to every interchange before it:

type Matrix = readonly (readonly (number | null)[])[];

// Interchanges in the gazette's table order.
const SOUTH = [
  { id: "katunayake", name: "Katunayake", exit: "B4" },
  { id: "ja-ela", name: "Ja-Ela", exit: "B3" },
  { id: "kerawalapitiya", name: "Kerawalapitiya", exit: "B2" },
  // … 29 in all
];

// Category 1: row n = fees from interchange n+1 to interchanges 0..n
const CAT1: Matrix = [
  [200],                 // Ja-Ela → Katunayake
  [250, 200],            // Kerawalapitiya → Katunayake, Ja-Ela
  [400, 250, 200],       // Peliyagoda → …
  // …
];

The lookup is three lines once you sort the two indexes:

const lo = Math.min(a, b);
const hi = Math.max(a, b);
const fee = fees[hi - 1]?.[lo] ?? null;

No duplicated upper half to keep in step, and the code has the same shape as the source. That made the checking much easier.

`null` is information, not a gap

Some cells in the gazette are blacked out as "not functional". Kotalawala, Kapuduwa and Bedigama are half interchanges: you can only travel one way from them. In the matrix those cells are null, and the calculator says why there's no fee instead of guessing one:

if (fee === null) {
  return {
    ok: false,
    reason: `The gazette gives no fee between ${from.name} and ${to.name}: that interchange does not connect in this direction.`,
  };
}

It is tempting to "fill in" a missing cell with the nearest sensible number. For data people pay money against, I'd rather show a clear "no fee here".

Transcribing about 1,200 numbers without trusting yourself

Three triangles of 29 interchanges are 3 × 406 = 1,218 cells, plus the Central Expressway tables. One typo and someone gets a wrong toll.

So the tables were typed in twice, independently, and the two copies were compared cell by cell.

After that, tests check what a correct table has to look like, not just a few hand-picked values:

test("every table is complete, and categories never get cheaper upwards", () => {
  for (let i = 0; i < SOUTH.length; i++)
    for (let j = i + 1; j < SOUTH.length; j++) {
      const f = ["1", "2", "3"].map((c) => fee(SOUTH[i].id, SOUTH[j].id, c));
      if (f.every((x) => x === null)) continue;                 // a not-functional pair
      assert.ok(f.every((x) => x !== null));                    // otherwise all three categories exist
      assert.ok(f[0] <= f[1] && f[1] <= f[2]);                  // a bigger vehicle never pays less
    }
});

test("from Kottawa southwards, a farther exit never costs less", () => {
  for (const c of ["1", "2", "3"]) {
    let last = 0;
    for (const to of SOUTH.slice(kottawa + 1)) {
      const f = fee("kottawa", to.id, c);
      assert.ok(f >= last);
      last = f;
    }
  }
});

These invariants catch the kind of mistake a spot check misses: a digit dropped in the middle of a row, or two cells swapped.

(The source has one quirk worth knowing: for Aparekka to Kapuduwa, category 1 is blacked out but categories 2 and 3 are simply left blank. Both become null, and the test has a comment saying why.)

Keep the date next to the data

Tolls change. The gazette number and date live in the code next to the matrix, and the page shows them:

export const GAZETTE_SOUTH = { no: "2467/49", date: "2025-12-17" } as const;
export const GAZETTE_CENTRAL = { no: "2329/17", date: "2023-04-24" } as const;

When the next gazette comes out, there is one place to change, and the tests say whether the new table still makes sense.

A few results

From the current tables:

Trip	Category 1	Category 2	Category 3
Kottawa → Pinnaduwa (Galle)	Rs. 500	Rs. 900	Rs. 1,500
Kottawa → Godagama (Matara)	Rs. 650	Rs. 1,150	Rs. 1,950
Katunayake → Hambantota Gate	Rs. 1,300	Rs. 2,350	Rs. 3,950
Mirigama North → Kurunegala (E04)	Rs. 300	Rs. 450	Rs. 700

You can pick any two interchanges here: Sri Lanka expressway toll calculator.

If you've turned another gazette table into data, I'd like to hear how you checked it.

This article was written with the help of AI. The code, numbers and sources come from my own project and were checked against it before publishing.

Why Sinhala copied from a Sri Lankan gazette looks like "Y%S ,xld" (and how to turn it into Unicode)

Hirantha Sandaruwan Lakshan — Wed, 28 Oct 2026 20:50:54 +0000

Copy a line of Sinhala from a Sri Lankan government gazette PDF and paste it somewhere. You don't get Sinhala. You get this:

Y%S ,xld m%cd;dka;%sl iudcjd§ ckrcfha .eiÜ m;%h

That line is the gazette's own heading, ශ්‍රී ලංකා ප්‍රජාතාන්ත්‍රික සමාජවාදී ජනරජයේ ගැසට් පත්‍රය. The cause is that the text was typed in FM Abhaya, FM Malithi or another "FM" font.

I wrote a converter for this. Here is how these fonts work and what it takes to turn them into proper Unicode.

The FM fonts are glyph fonts, not Sinhala fonts

The FM fonts come from before Unicode Sinhala was usable everywhere. They don't contain Sinhala characters at all. They draw Sinhala shapes over ordinary ASCII and Latin-1 codes:

You type	FM font draws	The real character
`l`	ක	U+0D9A SINHALA LETTER ALPAPRAANA KAYANNA
`i`	ස	U+0DC3
`d`	ා	U+0DCF (the aela-pilla vowel sign)
`f`	ෙ	the kombuva, U+0DD9
`%`	the rakaransaya under a letter	් + ZWJ + ර

So the file stores l, and only the font makes it look like ක. Lose the font and you see Latin letters. Search engines, screen readers and translation tools also only ever see the Latin letters.

Step 1: a character map

The first part is a big lookup table. Consonants, independent vowels, dependent vowel signs, punctuation:

const CONS = { l: "ක", L: "ඛ", ".": "ග", p: "ච", c: "ජ", g: "ට", ";": "ත", o: "ද", k: "න",
               m: "ප", n: "බ", u: "ම", h: "ය", r: "ර", ",": "ල", j: "ව", Y: "ශ", i: "ස", y: "හ" /* … */ };
const SIGNS = { d: "ා", e: "ැ", E: "ෑ", s: "ි", S: "ී", q: "ු", Q: "ූ", a: "්" /* … */ };
const VOWELS = { w: "අ", b: "ඉ", W: "උ", t: "එ", T: "ඔ" /* … */ };

If the fonts were only that, a replace per character would do. They aren't, for three reasons.

Step 2: FM text is in visual order

In Sinhala, the kombuva (ෙ) is written to the left of the consonant it belongs to. FM text stores what you see, so the kombuva comes first. Unicode stores the logical order: consonant first, then the vowel sign.

A two-part vowel like ො (kombuva + aela-pilla) is worse: FM types the kombuva before the consonant and the aela-pilla after it, but in Unicode ො is one code point after the consonant.

FM:       f  m  d  ;          (ෙ  ප  ා  ත)
Unicode:  ප  ො  ත             "පොත" (book)

So when the converter meets an f, it reads ahead: the consonant, then any signs typed after it, and decides which vowel the kombuva makes together with those signs:

// signs = the vowel-sign codes typed after the consonant
function kombuVowel(signs) {
  const [a, b] = signs;
  if (a === "d" && (b === "a" || b === "A")) return { vowel: "ෝ", rest: signs.slice(2) }; // f + C + d + a
  if (a === "d") return { vowel: "ො", rest: signs.slice(1) };                              // f + C + d
  if (a === "a" || a === "A") return { vowel: "ේ", rest: signs.slice(1) };                 // f + C + a
  if (a === "!") return { vowel: "ෞ", rest: signs.slice(1) };
  return { vowel: "ෙ", rest: signs };                                                       // plain kombuva
}

Two kombuvas in a row (ff) are the kombu deka, ෛ.

Step 3: rakaransaya and yansaya need a ZWJ

The little ර under a letter (rakaransaya, as in ප්‍ර) and the ය joined after one (yansaya, as in ක්‍ය) are one keystroke each in FM: % and H. In Unicode they are three characters: al-lakuna, a zero-width joiner (U+200D), then ර or ය. Without the ZWJ you get a visible al-lakuna and a separate letter instead of the joined form.

const ZWJ = "‍";
const RAKAR = "්" + ZWJ + "ර";
const YANS  = "්" + ZWJ + "ය";

They also have an ordering problem of their own: typists put the vowel sign before or after the %. Both of these turn up in real gazette text, and both are ශ්‍රී:

Y%S   and   YS%

So after a consonant the converter collects the rakaransaya/yansaya and the vowel signs separately, in any typed order, and writes them back out in Unicode order: consonant, then ්‍ර, then the vowel sign.

Step 4: ligatures

Some consonant + vowel pairs are a single glyph in FM, because the shape changes (§ is දී, ß is රි, ÷ is ඳු), and so are some conjuncts (Ë is ක්‍ෂ). Those map to several Unicode characters.

A trap here: a ligature can end in a sign, and a kombuva can come before it. ï is ම් with the al-lakuna built in, so fï should be මේ. If the ligature is treated as finished text, the kombuva can't combine with its al-lakuna and you get the wrong vowel. The fix was to split a ligature's trailing sign back off and treat it like a typed sign, so the kombuva logic above sees it.

The small things that bite

ර + ු is drawn differently. On ර the ු and ූ look like ැ and ෑ, so FM types රු as r + e and රූ as r + E. A naive map gives you රැ.
¾ is not a reph. It looks like one, but it is just the inline ර් glyph and stays where it is, so f¾ is රේ and ld¾hd,h is කාර්යාලය.
Normalise first. Several FM codes are Latin-1 letters with accents (ú is වි), and the same letter can arrive precomposed or as u + a combining accent. Calling input.normalize("NFC") before mapping makes both forms hit the same table entry.
English words get converted too. In an FM font every Latin letter is a Sinhala letter, so there is no safe way to skip English in mixed text. Paste Sinhala only. To avoid double-converting text that is already Unicode, I count Sinhala characters against Latin letters first.

Testing with real documents

Gazette PDFs make good test data. Their text layer is the FM codes and the printed page shows the Sinhala, so every line gives you an input and its expected output from a real document, not from my own typing:

import test from "node:test";
import assert from "node:assert/strict";

const Z = "‍";
test("Gazette heading: FM text layer to the printed Sinhala", () => {
  assert.equal(fmToUnicode("Y%S ,xld"), `ශ්${Z}රී ලංකා`);
  assert.equal(fmToUnicode("YS% ,xld"), `ශ්${Z}රී ලංකා`); // sign typed before the rakaransaya
  assert.equal(fmToUnicode("fmd;"), "පොත");
  assert.equal(fmToUnicode("isxy,"), "සිංහල");
});

Try it

The converter runs in the browser (nothing is uploaded): Sinhala FM to Unicode converter. Paste FM Abhaya text and you get Unicode you can put in a website, email or document.

If you find FM text it gets wrong, I'd like to see it. Leave it in the comments.

This article was written with the help of AI. The code, numbers and sources come from my own project and were checked against it before publishing.

DEV Community: Hirantha Sandaruwan Lakshan

Sri Lankan public holidays and Poya days as open data, and two date bugs to avoid

Where the data comes from

The files

Bug 1: "today" in UTC is the wrong day

Bug 2: counting working days

Licence

Turning Sri Lanka's expressway toll gazette into a lookup table

The shape of the table

null is information, not a gap

Transcribing about 1,200 numbers without trusting yourself

Keep the date next to the data

A few results

Why Sinhala copied from a Sri Lankan gazette looks like "Y%S ,xld" (and how to turn it into Unicode)

The FM fonts are glyph fonts, not Sinhala fonts

Step 1: a character map

Step 2: FM text is in visual order

Step 3: rakaransaya and yansaya need a ZWJ

Step 4: ligatures

The small things that bite

Testing with real documents

Try it

`null` is information, not a gap