Cryptographically strong, hat-compatible, pseudo-random number generator
npm install cryptohathat-Compatible CSPRNG
This is a node.js package that implements
hat's main API, but uses a
cryptographically secure pseudo-random number generator
to generate the random identifiers. This is especially beneficial when using
older versions of V8)
that exhibit this bug.
In our use cases, cryptohat takes up to 3x more time to produce a random
string than hat. However, cryptohat also exposes an alternative API that
produces random numbers up to 2x faster than hat, while still using a
CSPRNG.
This package should work on any reasonably modern browser or node.js version.
Every commit is tested using continuous integration
on node.js 0.10 and above. Releases are also tested against the most recent
versions of Chrome,
Firefox,
Safari, and
Internet Explorer.
``bash`
npm install cryptohat@1.x --save
bower install cryptohat@1.x --save
cryptohat implements the hat(bits, base) API. The function takes in thehat.rack
desired number of bits of randomness and the
base/radix that will be used to
represent the returned random number, and returns a string. The returned
strings are guaranteed to have the same length for a given bits/base
combination. As long as you don't need the method, cryptohat can
be used as a drop-in replacement.
`javascript
var hat = require('cryptohat');
hat(); // '39a00e331acce7516a8ea69b85e191f0'
hat(); // '00549f7401ac0ba9ea1b791f50bc7b1e'
hat(53, 10); // '6738095220277140'
`
Pass in zero (0) for the base argument to get a number. This is significantly
faster than obtaining a string and converting it into a number. Keep in mind
that JavaScript numbers can accurately represent integers of at most 53 bits.
`javascript
var cryptohat = require('cryptohat');
cryptohat(53, 0); // 6738095220277140
cryptohat(32, 0); // 3840742823
cryptohat(63, 0); // RangeError: JavaScript numbers can accurately represent at most 53 bits
`
For maximum throughput, use the cryptohat.generator API. It takes exactly thehat
same arguments as the API, but it returns a generator function. Calling
the function yields random numbers or identifiers.
`javascript
var rng1 = cryptohat.generator(53, 0);
rng1(); // 4438236126178078
rng1(); // 187896805323588
var rng2 = cryptohat.generator(32, 10);
rng2(); // '0053668130'
rng2(); // '1939036909'
`
After cloning the repository, install the dependencies.
`bash`
npm install
node node_modules/.bin/bower install
Make sure the tests pass after making a change.
`bash`
SKIP_BENCHMARKS=1 npm test
When adding new functionality, make sure it has good test coverage and that it
does not regress the code's performance.
`bash`
SKIP_BENCHMARKS=1 npm run cov
npm test
When adding new functionality, also make sure that the documentation looks
reasonable.
`bash`
npm run doc
When modifying code around or inside feature detection blocks (combinations of
if and typeof), make sure the tests pass at least in Chrome and Firefox, by
opening test/index.html in the browsers.
`bash`
open test/index.html # On OSX.
xdg-open test/index.html # On Linux.
When testing against a browser in a VM (e.g., for Internet Explorer), spawn a
local Web server inside the source tree and visit it inside the VM
(e.g., http://10.0.2.2:8080/test/index.html).
`bash``
node node_modules/.bin/http-server
If you submit a
pull request,
Travis CI will run the test suite against your code
on the node versions that we support. Please fix any errors that it reports.
Copyright (c) 2016 Heap Inc., released under the MIT license.