kodie / md5-file

return an md5sum of a given file

MD5 file

Get the MD5-sum of a given file, with low memory usage, even on huge files.

Installation

npm install --save md5-file

Usage

As a module

const md5File = require('md5-file')

/* Async usage */
md5File('LICENSE.md').then((hash) => {
  console.log(`The MD5 sum of LICENSE.md is: ${hash}`)
})

/* Sync usage */
const hash = md5File.sync('LICENSE.md')
console.log(`The MD5 sum of LICENSE.md is: ${hash}`)

As a command line tool

$ md5-file LICENSE.md
ad1faf9381e43c471dc381c17a4ee4b6

API

md5File(path: string) => Promise<string>

Asynchronously get the MD5-sum of the file at path.

Returns a Promise that will be resolved with a string containing the MD5-sum.

md5File.sync(path: string) => string

Synchronously get the MD5-sum of the file at path.

License

MIT

md5-file's Issues

Async version is slow

It seems async version is slower than it needs to be. I haven't compares it with sync version because I'm using async one in multiple concurrent operations. They seem to be a lot slower than they need to be.

I notice that in sync version the file is only read upto a certain BUFFER_SIZE. This seems not to be the case in async version - it seems to read the whole file. Could that be the cause?

Looking for new maintainer

I don't use this module anymore, and to be honest I don't even use Node.js or GitHub much anymore, and I don't have the time to update it, review issues or pull requests, so I am looking for a new maintainer. I won't just pick anyone, as people are using the module, so I have to ensure it goes into good hands. If no one suitable comes forward (in a reasonable amount of time) I will deprecate the module in npm and archive the repo in GitHub.

Invalid MD5 value

Hello,
Seems like running this package on this file:
http://www.tekdefense.com/downloads/malware-samples/0.exe.zip

results with MD5 hash of 33 characters - which is invalid.

Is this a known issue? am i missing something?

update: it's 32 characters long... my bad :)

TypeScript interface

Is TypeScript interface in the roadmap?

Merge promise support with callbacks

In the main function, the use of a callback can be detected and can be used to either return a promise or to execute it.

"File not found" error for .asar files

When processing files with .asar file extension, md5-file module reports a "file not found error".

promise.js file is not included in module

Hello,

I installed the package with npm install --save md5-file and got ^3.1.0, but the promise.js file is not included, so require('md5-file/promise') does not work.

You can use pipe

You can .pipe the fileStream directly to the hash instead of using update.

fileStream.pipe(sum)

Breaking change between 4.0.0 and 5.0.0

Using md5-file to md5 files with a callback on 4.0.0 works as below:

md5File(filename, (err, hash) => {
   console.log("starting hash on " + filename);
   if (err) {
      console.log(err);
   } else {
      console.log(hash);
   }
});

On 5.0.0, this no longer works, the code within the callback is never executed.

Change api to have two separate functions

I think it would be nicer if the api looked like this:

// Async
md5File('LICENSE.md', (err, hash) => {})

// Sync
md5File.sync('LICENSE.md')

This seems to be where the community landed, and there are a lot of high profile packages (see below) following this style. By following this style, we adhere to the Principle_of_least_astonishment.

• rimraf
• mkdirp
• glob
• del

It would also allow us to have more strict arguments validation 🙌

Support modern browser runtime?

Hi Rory and Linus,

Thanks for this great repo!

I am wondering whether you would like to add modern browser support? Maybe utilising FileReader API to handle fs work?

It would be great to see that happen. I can contribute with some PR too :)

Best,
Blake

There is no promise.js in the dir after install with npm.

There isn't a promise.js file under the dir.
I installed it with npm i --save-dev md5-file.

doesn't work with files > 1GB

Thanks for making this package. It was useful to get me started.

FYI, using fs.readFile doesn't work for > 1 GB files since it reads the entire contents of the file into memory and the max buffer size is 1 GB.

This can work using a streaming approach such as discussed here:

http://stackoverflow.com/questions/18658612/obtaining-the-hash-of-a-file-using-the-stream-capabilities-of-crypto-module-ie

I used createReadStream piped to a hash object successfully to get the MD5 hash of 2GB files.
Of course it's not for a security context, simply a convenience for a download check.

Thanks again for your contributions to the community.

Reverse Process

Hey,so i was wandering if there is a way to convert the generated hash back to the file again.
Can the process be reversed? :)

A question about uniqueness

Is it possible that two files that are not same have same hash result?

Include the path of the file when generating the file's hash to prevent duplicate hashes when files have no content of their own.

I noticed that files with no content of their own all had the same hash. Would it be possible to include the file path when generating their hashes to guarantee that all files, including those with no content, all have unique hashes?

Not working on windows

Not working in windows.

But after commenting the first line which is #!/usr/bin/env node of the node_modules\md5-file\index.js it works fine.