Is there any way to calculate incrementally?

Please see this small example: https://jsfiddle.net/j1rp55oo/

You can check that Base64.encode(...) works correctly by decoding it's output using OpenSSL or any other base64 encoder. Base64.decode(...) on the other hand, returns complete garbage.

does this SHA256 compatible with php hash ?

Hi, I'm working on a project that I'd like to support SHA-384. Any plans to implement this?

Hi! The current example subfolders "client" and "server" make it look like this would be a client/server application. From the code it looks more like the "client" scripts are meant to run in web browsers, so how about we rename that folder to "web" or "html" or "browser-native"?

The "server" scripts seem meant to be run in a CommonJS environment such as Node.js, so how about we name that folder "cjs"? I recommend not to use "cli" or "node", because there are frameworks that enable browsers to load CommonJS modules (no build step required).

This package doesn't have a default file set. You can set it via jsdelivr, browser, or main field in package.json

which will make it more convenient! : )

Cloudflares new IUAM mode uses most of the library without respecting the license

Hi! https://www.npmjs.com/package/jshashes currently shows "none" as the license. I think you'll need a simple SPDX identifier string as the license field value. If I understand https://spdx.org/licenses/BSD-3-Clause.html correctly, the correct value would be "BSD-3-Clause".

I'm not sure what's wrong, but the digest is always "b613679a0814d9ec772f95d778c35fc5ff1697c493715653c6c712144292c5ad" when using the SHA256.HMAC_hex function.

There are 'var' declarations missing for a few variables, letting them leak to the global scope.

What does the Hashes.Helpers.urlEncode and Hashes.Helpers.urlDecode used for exactly? I mean do they mimic encodeURI or encodeURIComponent or are these encodes specifically for use within the hash (as they are thus called helper functions)

The 1.0.4 tag is missing from the repo.

The CRC32 checksum sometimes returns negative values.
You can use an unsigned shift (x >>> 0) to force JavaScript to treat the arithmetic as unsigned.

Can anyone tell me how can i use this code for aws signature generation.
It is step 3 of aws signature generation.
Link of aws doc : https://docs.aws.amazon.com/general/latest/gr/sigv4-calculate-signature.html

Signature generated does not match .......sample input is given in aws document.

This is a great package, and thank you so much for it! However, I think it could be improved if each algorithm was it's own file, for those of us who only need one or two different algorithms.

Then, the build process could concatenate all the files and your bower.json could remain the same, but overrides could be specified as necessary, e.g.

"overrides": {
  "jshashes": {
    "main": "sha256.js"
  }
}

What do you think?

I am developing an OpenHarmony application and have JavaScript compatibility issues. Can you support the OpenHarmony version? I am willing to participate in the development of adaptation and look forward to reply.

The README says "hex_hmac(string,key)" but in fact all the functions expect the arguments in the opposite order (key first, then data).

The bin/hashes.js in the NPM package (which seems to be missing from this repository) has the path to /usr/bin/env on the first line.

It should also probably have execute permissions.

I would be nice with an example of how to use HMAC. Are there any restrictions on salt? Length/format ?

Would be really nice if we could include only the hashing algorithms we need.

If you only need MD5, it would be nice to be able to exclude SHA1, SHA256, SHA512, CRC32 etc etc.

Base64 encode/decode of unicode chars first tries to convert to utf8, right idea but wrong implementation.

(new Hashes.Base64()).encode('张')
"5Q=="
(new Hashes.Base64()).setUTF8(false).encode(unescape(encodeURIComponent('张')))
"5byg"
window.btoa(unescape(encodeURIComponent('张')))
"5byg"

See https://developer.mozilla.org/en-US/docs/Web/JavaScript/Base64_encoding_and_decoding#The_.22Unicode_Problem.22

Implementation wise utf8Encode() returns same as unescape(encodeURIComponent()) so the problem is elsewhere. See https://github.com/davidchambers/Base64.js/blob/master/base64.js for window.atob() shim. I think len shoud be after utf8 encode.

Also base 64 decode is wrong.

(new Hashes.Base64()).decode('5byg')
"o("
(new Hashes.Base64()).setUTF8(false).decode('5byg')
"o("
decodeURIComponent(escape(window.atob("5byg")))
"张"

Whirlpool algorithm here: http://www.sunsean.com/Whirlpool.js

I'm executing:

new Hashes.SHA256().any('test', "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789`~!@#$%^&*()_-+={}|[]\\:\";'<>?,./")

which give me A-*A_[Hza*ek.e]hlq*%VwLgH0zDi-?4);_0tbQk,
but I'm expecting something like -*A_[Hza*ek.e]hlq*%VwLgH0zDi-?4);_0tbQk\, without the A prefix.

I'm not sure I'm expecting the right result, but I observed a lot of outputs beginning with the first character of my encoding (A).

Using SHA1, the result seems to be OK, but I'm suspecting the same issue appears with MD5 and SHA512.

Hello I see the following error when trying to use jsHashes on Mac OSX:::

$ hashes md5-b64 hello
env: node\r: No such file or directory

This is because you are using windows line endings. I changed the line endings to Unix line endings and it worked. I am not sure which line endings will support all platforms.. hence no pull request. Thanks

warning [email protected]: The engine "browser" appears to be invalid.

This appears when using various versions of Yarn (rather than npm) to install jshashes, either individually or in combination.

It happens in all environments, both in my local and in CI builds.

When googling the error string, it appears to be happening in many other builds as well.

It's non-blocking. But, FYI.

It seems so that it doesn't work in PhantomJS, it always returns the same hash for different outpusts (tested with MD5 and SHA1), works as expected in Chrome and Firefox

Digging on code, I found that.. and it seems strange
b64pad = (options && typeof options.pad === 'string') ? options.pda : '=', // base-64 pad character. Defaults to '=' for strict RFC compliance

Shouldn't options.pda be options.pad ?

If that's the case, i suspect same bug/typpo will repeat fro other Hash classes

I had problems running the minified version of jsHashes, so I ran the unminified version through CodeKit. JSHint reports over a thousands warnings on this file. Just thought I'd mention it :-P

npm install jshashes results in the following server directory structure under express:

node_modules/jshashes/server/lib

your use instructions suggest using:

var Hashes = require('./lib/hashes');

This throws a cannot find module error

https://github.com/h2non/jshashes/blob/master/hashes.js#L331. It seems that the 'input' start to parse the char from '1', but not '0', which cause the problem.

If I try to define in my component.json the dependency on jshashes as in the following:

{
    ...
    "jshashes": "master"
    ...
}

I get the following error:

bower jshashes#master                   ECMDERR Failed to execute "git ls-remote --tags --heads git://github.com/h2non/jsHashes.git", exit code of #128

It looks to me that the jshashes endpoint on bower is wrong and it should be:

git://github.com/h2non/jshashes.git

With the lowercase 'h'. To prove this I tried to define the dependency in this way:

{
    ...
    "jshashes": "git://github.com/h2non/jshashes.git#master"
    ...
}

And it worked fine. Is it possible to fix the definition on bower?

Thank you!

Hashes.Base64.decode parse errors

b64.decode(b64.encode(str))

Just thought you might like to see this!

http://jsperf.com/sha256/2

If you want I can write a TypeScript type declaration file for this package so that TS developers can utilise the type-checking feature of the language as they use this package. Let me know if you are interested.