xxHash sometimes doesn't match the Hash from the C library from Google #16

Activity

1. illcom94 commented on Aug 20, 2015

   

   illcom94

   on Aug 20, 2015

   Author

   More actions

   I stepped through the code along with the C++ code from google and I've found the issue.
   xxHash.cs

   1. In ComputeHashInternal() the dataCount variable should be a long instead of an int. On large files, it overflows into negative.
   2. In PostProcess() you have an if statement "if ((remainder.Length % 8) > 4)" this ensures that your code for processing the last 4 bytes of a file will never be hit. It should be changed to " if ((remainder.Length % 8) == 4)".

   After these two changes, everything seems to match.
2. illcom94 commented on Aug 20, 2015

   

   illcom94

   on Aug 20, 2015

   Author

   More actions

   After running a few more files I had to make a modification to my #2 fix.
   The if statement should be "if ((remainder.Length % 8) >= 4)"
3. brandondahler commented on Aug 31, 2015

   

   brandondahler

   on Aug 31, 2015

   Owner

   More actions

   I'll take a look at as soon as I can. If you know how to submit a pull request with the changes you talked about above, it could significantly help expedite the process.
4. brandondahler

   added 3 commits that reference this issue on Sep 5, 2015

   Issue #16 - xxHash Deviations

   ...

   f9dd177

   Issue #16 - xxHash Deviations

   ...

   764edbe

   Issue #16 - xxHash Deviations

   ...

   165e3d5
5. brandondahler

   added a commit that references this issue on Sep 27, 2015

   Merge pull request #17 from brandondahler/bugfix/xxHash

   ...

   9a464b9
6. brandondahler commented on Sep 27, 2015

   

   brandondahler

   on Sep 27, 2015

   Owner

   More actions

   This issue has been addressed in release 1.8.1.
7. brandondahler

   closed this as completedon Sep 27, 2015