- Jan 17, 2020
-
-
Lars Knoll authored
The C++ standard does the right thing by default.
-
Lars Knoll authored
-
Lars Knoll authored
-
Lars Knoll authored
-
Lars Knoll authored
-
Lars Knoll authored
-
Lars Knoll authored
-
Lars Knoll authored
-
Lars Knoll authored
This now seems to pass all test cases. Can probably be cleaned up and optimized a bit more.
-
Lars Knoll authored
-
Lars Knoll authored
Start fixing QMultiHash, requires more work.
-
- Jan 15, 2020
-
-
Lars Knoll authored
-
Lars Knoll authored
And some small cleanup for keys() and values().
-
Lars Knoll authored
-
Lars Knoll authored
-
- Jan 14, 2020
-
-
Lars Knoll authored
-
Lars Knoll authored
-
- Jan 13, 2020
-
-
Lars Knoll authored
-
Lars Knoll authored
-
Lars Knoll authored
-
Lars Knoll authored
Avoid holes in the bucket list to make the hash stable for long term insertions/erases. This makes the whole table stable with regards to continuous inserts/deletes. Add a test case for a hash with continuous inserts/removals.
-
- Jan 07, 2020
-
-
Lars Knoll authored
-
Lars Knoll authored
We need to remember the deleted bucket, to be able to break out of the loop
-
- Jan 06, 2020
-
-
Lars Knoll authored
-
Lars Knoll authored
Fix errors in QMultiHash::erase() and add equal_range() methods. Move the growth policy into the QHashPrivate namespace.
-
Lars Knoll authored
And a smaller cleanup when calling insert(QHash).
-
Lars Knoll authored
-
- Jan 03, 2020
-
-
Lars Knoll authored
Inline code into class definition, add some typedefs for STL compatibility.
-
Lars Knoll authored
It doesn't seem to make a difference in performance
-
Lars Knoll authored
This requires the prepend optimization for QVector
-
- Dec 23, 2019
-
-
Lars Knoll authored
-
- Dec 20, 2019
-
-
Lars Knoll authored
Extend it to 48 bits for 64 bit platforms, and properly handle the limits on 32 bits.
-
Lars Knoll authored
-
Lars Knoll authored
-
Lars Knoll authored
-
Lars Knoll authored
-
Lars Knoll authored
Avoids a spike in memory usage while rehashing.
-
Lars Knoll authored
use if for integer and pointer based keys, as they don't have very good hashing functions and often a quite uneven distribution.
-
Lars Knoll authored
-
Lars Knoll authored
-