Archived filenames are not read properly.

[agr]

Using zip.js from npm:

    "@zip.js/zip.js": "^2.6.12"

Very simple code:

    const url = 'https://globalcdn.nuget.org/packages/newtonsoft.json.13.0.1.nupkg';
    const httpReader = new zip.HttpRangeReader(url);
    httpReader.reader.forceRangeRequests = true;   // Accept-Ranges is not exposed through CORS request, so need some overrides
    const zipReader = new zip.ZipReader(httpReader);
    const entries = await zipReader.getEntries();

yields broken filenames in the entries array:

However, if I manually decode rawFilename, I get proper value:

How do I fix it?

[gildas-lormeau]

I tested your code in Firefox, Chrome, Microsoft Edge (on Win11, Ubuntu 20 and Android 12) and I cannot reproduce the issue, see https://jsfiddle.net/uvmyq3b9/. Do you run the code in a particular environment?

FYI, the filenames in the zip file of your code are encoded in CP437. Thus, the decoding of the data into UTF-8 is handled by the function exported in /lib/core/util/cp437-decode.js. It looks like CP437 is corrupted in your environment.

https://github.com/gildas-lormeau/zip.js/blob/876c084cebe48bcf6f37085481d2f839cbc2d4f9/lib/core/util/cp437-decode.js#L28-L30

[agr]

Windows 10, the code is loaded from nginx running on localhost. I use zip-full.js from the dist directory. Also tried all browsers, the behavior is the same everywhere. Will experiment more.

[agr]

Oh, maybe the file itself is not treated as UTF-8 when served by nginx...

[gildas-lormeau]

Can you reproduce the problem with the test on jsfiddle? The file should be served as binary data by nginx. If the encoding was invalid, zip.js wouldn't be able to find the entries in the file.

[agr]

No, jsfiddle works fine. This is the content type I get from unpkg.com: This is the content type I get from my nginx: That missing charset is likely the cause.

[gildas-lormeau]

Sorry, I thought you were referring to the zip file. It's strange though, that would mean the JS code is not parsed in UTF-8.

[agr]

Yes, nginx was not configured to specify charset and apparently browsers default to interpret JS code using some code page other than UTF-8, which causes JS interpreter to incorrectly read the value of CP437 (since it contains multi-byte (in UTF-8) characters right from the start, it treats individual bytes as a character, which eventually messes up decoding). I enforced UTF-8 in nginx and it fixed the issue.

[gildas-lormeau]

I'm glad to hear it! Maybe I should add a note in the documentation about this requirement.

[gildas-lormeau]

I added a note, I think I can close the issue now. Thank you for pointing this out.

[agr]

Replacing that "...".split("") with something like this:

const CP437 = [String.fromCharCode(0),String.fromCharCode(9786),String.fromCharCode(9787),String.fromCharCode(9829),String.fromCharCode(9830),String.fromCharCode(9827),String.fromCharCode(9824),String.fromCharCode(8226),String.fromCharCode(9688),String.fromCharCode(9675),String.fromCharCode(9689),String.fromCharCode(9794),String.fromCharCode(9792),String.fromCharCode(9834),String.fromCharCode(9835),String.fromCharCode(9788),String.fromCharCode(9658),String.fromCharCode(9668),String.fromCharCode(8597),String.fromCharCode(8252),String.fromCharCode(182),String.fromCharCode(167),String.fromCharCode(9644),String.fromCharCode(8616),String.fromCharCode(8593),String.fromCharCode(8595),String.fromCharCode(8594),String.fromCharCode(8592),String.fromCharCode(8735),String.fromCharCode(8596),String.fromCharCode(9650),String.fromCharCode(9660),String.fromCharCode(32),String.fromCharCode(33),String.fromCharCode(34),String.fromCharCode(35),String.fromCharCode(36),String.fromCharCode(37),String.fromCharCode(38),String.fromCharCode(39),String.fromCharCode(40),String.fromCharCode(41),String.fromCharCode(42),String.fromCharCode(43),String.fromCharCode(44),String.fromCharCode(45),String.fromCharCode(46),String.fromCharCode(47),String.fromCharCode(48),String.fromCharCode(49),String.fromCharCode(50),String.fromCharCode(51),String.fromCharCode(52),String.fromCharCode(53),String.fromCharCode(54),String.fromCharCode(55),String.fromCharCode(56),String.fromCharCode(57),String.fromCharCode(58),String.fromCharCode(59),String.fromCharCode(60),String.fromCharCode(61),String.fromCharCode(62),String.fromCharCode(63),String.fromCharCode(64),String.fromCharCode(65),String.fromCharCode(66),String.fromCharCode(67),String.fromCharCode(68),String.fromCharCode(69),String.fromCharCode(70),String.fromCharCode(71),String.fromCharCode(72),String.fromCharCode(73),String.fromCharCode(74),String.fromCharCode(75),String.fromCharCode(76),String.fromCharCode(77),String.fromCharCode(78),String.fromCharCode(79),String.fromCharCode(80),String.fromCharCode(81),String.fromCharCode(82),String.fromCharCode(83),String.fromCharCode(84),String.fromCharCode(85),String.fromCharCode(86),String.fromCharCode(87),String.fromCharCode(88),String.fromCharCode(89),String.fromCharCode(90),String.fromCharCode(91),String.fromCharCode(92),String.fromCharCode(93),String.fromCharCode(94),String.fromCharCode(95),String.fromCharCode(96),String.fromCharCode(97),String.fromCharCode(98),String.fromCharCode(99),String.fromCharCode(100),String.fromCharCode(101),String.fromCharCode(102),String.fromCharCode(103),String.fromCharCode(104),String.fromCharCode(105),String.fromCharCode(106),String.fromCharCode(107),String.fromCharCode(108),String.fromCharCode(109),String.fromCharCode(110),String.fromCharCode(111),String.fromCharCode(112),String.fromCharCode(113),String.fromCharCode(114),String.fromCharCode(115),String.fromCharCode(116),String.fromCharCode(117),String.fromCharCode(118),String.fromCharCode(119),String.fromCharCode(120),String.fromCharCode(121),String.fromCharCode(122),String.fromCharCode(123),String.fromCharCode(124),String.fromCharCode(125),String.fromCharCode(126),String.fromCharCode(8962),String.fromCharCode(199),String.fromCharCode(252),String.fromCharCode(233),String.fromCharCode(226),String.fromCharCode(228),String.fromCharCode(224),String.fromCharCode(229),String.fromCharCode(231),String.fromCharCode(234),String.fromCharCode(235),String.fromCharCode(232),String.fromCharCode(239),String.fromCharCode(238),String.fromCharCode(236),String.fromCharCode(196),String.fromCharCode(197),String.fromCharCode(201),String.fromCharCode(230),String.fromCharCode(198),String.fromCharCode(244),String.fromCharCode(246),String.fromCharCode(242),String.fromCharCode(251),String.fromCharCode(249),String.fromCharCode(255),String.fromCharCode(214),String.fromCharCode(220),String.fromCharCode(162),String.fromCharCode(163),String.fromCharCode(165),String.fromCharCode(8359),String.fromCharCode(402),String.fromCharCode(225),String.fromCharCode(237),String.fromCharCode(243),String.fromCharCode(250),String.fromCharCode(241),String.fromCharCode(209),String.fromCharCode(170),String.fromCharCode(186),String.fromCharCode(191),String.fromCharCode(8976),String.fromCharCode(172),String.fromCharCode(189),String.fromCharCode(188),String.fromCharCode(161),String.fromCharCode(171),String.fromCharCode(187),String.fromCharCode(9617),String.fromCharCode(9618),String.fromCharCode(9619),String.fromCharCode(9474),String.fromCharCode(9508),String.fromCharCode(9569),String.fromCharCode(9570),String.fromCharCode(9558),String.fromCharCode(9557),String.fromCharCode(9571),String.fromCharCode(9553),String.fromCharCode(9559),String.fromCharCode(9565),String.fromCharCode(9564),String.fromCharCode(9563),String.fromCharCode(9488),String.fromCharCode(9492),String.fromCharCode(9524),String.fromCharCode(9516),String.fromCharCode(9500),String.fromCharCode(9472),String.fromCharCode(9532),String.fromCharCode(9566),String.fromCharCode(9567),String.fromCharCode(9562),String.fromCharCode(9556),String.fromCharCode(9577),String.fromCharCode(9574),String.fromCharCode(9568),String.fromCharCode(9552),String.fromCharCode(9580),String.fromCharCode(9575),String.fromCharCode(9576),String.fromCharCode(9572),String.fromCharCode(9573),String.fromCharCode(9561),String.fromCharCode(9560),String.fromCharCode(9554),String.fromCharCode(9555),String.fromCharCode(9579),String.fromCharCode(9578),String.fromCharCode(9496),String.fromCharCode(9484),String.fromCharCode(9608),String.fromCharCode(9604),String.fromCharCode(9612),String.fromCharCode(9616),String.fromCharCode(9600),String.fromCharCode(945),String.fromCharCode(223),String.fromCharCode(915),String.fromCharCode(960),String.fromCharCode(931),String.fromCharCode(963),String.fromCharCode(181),String.fromCharCode(964),String.fromCharCode(934),String.fromCharCode(920),String.fromCharCode(937),String.fromCharCode(948),String.fromCharCode(8734),String.fromCharCode(966),String.fromCharCode(949),String.fromCharCode(8745),String.fromCharCode(8801),String.fromCharCode(177),String.fromCharCode(8805),String.fromCharCode(8804),String.fromCharCode(8992),String.fromCharCode(8993),String.fromCharCode(247),String.fromCharCode(8776),String.fromCharCode(176),String.fromCharCode(8729),String.fromCharCode(183),String.fromCharCode(8730),String.fromCharCode(8319),String.fromCharCode(178),String.fromCharCode(9632),String.fromCharCode(32)];

removes the dependency on the served code page at least in this case.

[gildas-lormeau]

I agree but it makes the built files significantly bigger.

[agr]

Well, I guess a string with \x codes should also work, give me a few moments to produce one.

[gildas-lormeau]

That would be useful only when using built files. I provide them but, in 2022, I expect most developers will import the library as an ES module. The HTML5 specification mandates the module files to be parsed in UTF-8 (see step 14 here: https://html.spec.whatwg.org/#fetch-a-single-module-script) so I'm tempted to say this issue is out of the scope of the library.

[agr]

const CP437 = "\u0000\u263a\u263b\u2665\u2666\u2663\u2660\u2022\u25d8\u25cb\u25d9\u2642\u2640\u266a\u266b\u263c\u25ba\u25c4\u2195\u203c\u00b6\u00a7\u25ac\u21a8\u2191\u2193\u2192\u2190\u221f\u2194\u25b2\u25bc !\"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\\]^_`abcdefghijklmnopqrstuvwxyz{|}~\u2302\u00c7\u00fc\u00e9\u00e2\u00e4\u00e0\u00e5\u00e7\u00ea\u00eb\u00e8\u00ef\u00ee\u00ec\u00c4\u00c5\u00c9\u00e6\u00c6\u00f4\u00f6\u00f2\u00fb\u00f9\u00ff\u00d6\u00dc\u00a2\u00a3\u00a5\u20a7\u0192\u00e1\u00ed\u00f3\u00fa\u00f1\u00d1\u00aa\u00ba\u00bf\u2310\u00ac\u00bd\u00bc\u00a1\u00ab\u00bb\u2591\u2592\u2593\u2502\u2524\u2561\u2562\u2556\u2555\u2563\u2551\u2557\u255d\u255c\u255b\u2510\u2514\u2534\u252c\u251c\u2500\u253c\u255e\u255f\u255a\u2554\u2569\u2566\u2560\u2550\u256c\u2567\u2568\u2564\u2565\u2559\u2558\u2552\u2553\u256b\u256a\u2518\u250c\u2588\u2584\u258c\u2590\u2580\u03b1\u00df\u0393\u03c0\u03a3\u03c3\u00b5\u03c4\u03a6\u0398\u03a9\u03b4\u221e\u03c6\u03b5\u2229\u2261\u00b1\u2265\u2264\u2320\u2321\u00f7\u2248\u00b0\u2219\u00b7\u221a\u207f\u00b2\u25a0 ".split("");

That should work.

[gildas-lormeau]

It should indeed work but when I build the project (npm run build), terser transforms the string back in minified files, i.e. they are strictly identical.

[nenge123]

If you are worried about file coding, Preferably. new Function(new TextDecoder().decode(new Uint8array(await (await fetch("zip.js")).arrayBuffer())))();

or

script.src = new File([new TextDecoder().decode(new Uint8array(await (await fetch("zip.js")).arrayBuffer()))],"zip.js",{type:"text/javascript"})

[gildas-lormeau]

Thank you @nenge123 for the suggestion.

I might be able to detect in zip.js when this error occurs and use TextDecoder as a fallback. I'll do some tests.

[gildas-lormeau]

It's fixed in the version 2.6.13 I just published.