search

zndznd / crypto-js

Automatically exported from code.google.com/p/crypto-js
0 stars 0 forks source link

comparison against sjcl and found it ran much faster #59

Open GoogleCodeExporter opened 8 years ago

GoogleCodeExporter commented 8 years ago 
What steps will reproduce the problem?
1. Generate CryptoJS.HmacSHA256 signature 10000 times with the same data and key
2. Compare results
3. Notice a varying number of wrong results per execution

What is the expected output? What do you see instead?
Given `CryptoJS.HmacSHA256("1354733244", "6f19e4981d0e2aacfb7b38b80da34216")`
Expected: 0a99d375b3b8910f965acf19bfd216c061865932357252e307422e4bba8ac2db
Result: 403d3aa50ea53401169712d3d07a43d2cd90d434257b6e9a83ea38d11c24c7a0

What version of the product are you using? On what operating system?
v3.0.2 in Chrome 25.0.1337.0 dev on OS X 10.7.5

Please provide any additional information below.

See attached.

I also ran a comparison against sjcl (https://github.com/bitwiseshiftleft/sjcl) 
and found it didn't have this issue and ran much faster.

Original issue reported on code.google.com by Jessestu...@gmail.com on 5 Dec 2012 at 8:29

Attachments:

GoogleCodeExporter commented 8 years ago 
[deleted comment]
GoogleCodeExporter commented 8 years ago 
Update:

sjcl does have this issue (see attached), and it does not occur in FireFox 17 
suggesting it's an issue with Chrome 25 and not with either library.

Original comment by Jessestu...@gmail.com on 5 Dec 2012 at 9:23

Attachments:

GoogleCodeExporter commented 8 years ago 
Google's description for dev releases of Chrome says that it's "where ideas get 
tested (and sometimes fail); the Dev channel can be very unstable at times." So 
I don't think we need to worry too much about CryptoJS's results when tested on 
an unstable dev build.

I can, however, confirm the speed difference in the latest stable Chrome. I'll 
investigate where that performance comes from.

Original comment by Jeff.Mott.OR on 5 Dec 2012 at 9:49

GoogleCodeExporter commented 8 years ago 
Changing label from defect to enhancement since the bug report is invalid 
(browser bug in unstable build) but the performance comparison is a valid 
enhancement request.

Original comment by Jeff.Mott.OR on 7 Dec 2012 at 6:56

GoogleCodeExporter commented 8 years ago 
Updated summary to reflect the new focus of this issue.

Original comment by Jeff.Mott.OR on 3 Jan 2013 at 10:38