3 ms·

Comparing Hashes in PHP the Wrong Way

2 points by sarciszewski 12y ago

sarciszewski 12y ago
Better: hash_equals() Best: use password_hash() and password_verify()
- opless 12y ago
  or use the "is this really equal" operator '==='
  - sarciszewski 12y ago
    http://blog.astrumfutura.com/2010/10/nanosecond-scale-remote-timing-attacks-on-php-applications-time-to-take-them-seriously/ http://blog.astrumfutura.com/2010/10/nanosecond-scale-remote... For cryptographic secrets, password hashes, MACs, digital signatures, etc. use a comparison operator that runs in constant time. Ergo, hash_equals()
opless 12y ago
This is one reason that I don't regard PHP as anything more than a toy language.
- sarciszewski 12y ago
  Your elitism is noted and disregarded.
  - ainiriand 12y ago
    Indeed. This 'toy language' has accomplished many millions of strong websites. It has some faults, but all languages has faults.
    - opless 12y ago
      Just because something is popular does not mean that it's "good". However it is "good enough" for many things. Much in the same way as BASIC was rubbish in the 8 but days. But the usage of BASIC was near universal because it was near universal. Php is on virtually every web hosting outfits web servers.
  - opless 12y ago
    Perhaps I'm being a bit harsh, but there should be nothing in a language that should surprise a user of it. The == and === operators are a prime example of this. There are many other reasons why the language is pretty awful to use, date handling not being consistent with the documentation and being generally broken is something else that springs to mind for example. Then again, not many languages get that right either. :-)
LarryMade2 12y ago
So both works out to something that looks like an exponent value, and then the non-strct comparison evaluates them both as 0. The simple solution is ===. A lot of luck having a hash with an 0e############ result.
- sarciszewski 12y ago
  https://twitter.com/voodooKobra/status/534724855103774721 https://twitter.com/voodooKobra/status/534724855103774721 Comparing /^0e[0-9]{30}$/ with /^[0-9a-f]$/ gives you a chance of about 2.939 x 10^-9 in 1 (or about 0.0000002939%) of occurring. This is a naive probability estimate. Compared to a birthday collision, we can expect 6.28 * 10^28 0e############################## cases to occur for every hash collision. Of course, there's also the possibility that 1e would act the same way, which effectively doubles the odds.

2 points by sarciszewski 12y ago

sarciszewski 12y ago

Better: hash_equals() Best: use password_hash() and password_verify()

opless 12y ago
or use the "is this really equal" operator '==='
- sarciszewski 12y ago
  http://blog.astrumfutura.com/2010/10/nanosecond-scale-remote-timing-attacks-on-php-applications-time-to-take-them-seriously/ http://blog.astrumfutura.com/2010/10/nanosecond-scale-remote... For cryptographic secrets, password hashes, MACs, digital signatures, etc. use a comparison operator that runs in constant time. Ergo, hash_equals()

opless 12y ago

This is one reason that I don't regard PHP as anything more than a toy language.

sarciszewski 12y ago
Your elitism is noted and disregarded.
- ainiriand 12y ago
  Indeed. This 'toy language' has accomplished many millions of strong websites. It has some faults, but all languages has faults.
  - opless 12y ago
    Just because something is popular does not mean that it's "good". However it is "good enough" for many things. Much in the same way as BASIC was rubbish in the 8 but days. But the usage of BASIC was near universal because it was near universal. Php is on virtually every web hosting outfits web servers.
- opless 12y ago
  Perhaps I'm being a bit harsh, but there should be nothing in a language that should surprise a user of it. The == and === operators are a prime example of this. There are many other reasons why the language is pretty awful to use, date handling not being consistent with the documentation and being generally broken is something else that springs to mind for example. Then again, not many languages get that right either. :-)

LarryMade2 12y ago

So both works out to something that looks like an exponent value, and then the non-strct comparison evaluates them both as 0. The simple solution is ===. A lot of luck having a hash with an 0e############ result.

sarciszewski 12y ago
https://twitter.com/voodooKobra/status/534724855103774721 https://twitter.com/voodooKobra/status/534724855103774721 Comparing /^0e[0-9]{30}$/ with /^[0-9a-f]$/ gives you a chance of about 2.939 x 10^-9 in 1 (or about 0.0000002939%) of occurring. This is a naive probability estimate. Compared to a birthday collision, we can expect 6.28 * 10^28 0e############################## cases to occur for every hash collision. Of course, there's also the possibility that 1e would act the same way, which effectively doubles the odds.