Malthe Borch e47ca7ab40 Use a hash table to index existing vtables (#5314)
* Use a hash table to index existing vtables

This allows for quick deduplication even in situations where there
might be thousands of vtables due to 'combinatoric explosion'.

This fixes issue #5301.

* Refactor 0-offset trimming

* Improve deduplication benchmark

The routine now generates a set of realistic logical layouts and
uses a timer function that randomly picks a layout for each iteration.

The benchmark runs in batches of # of logical layouts = 1, 10, 100, 1000.

(Note that due to alignment, the actual number of vtables is usually slightly
higher.)
2019-05-06 15:00:02 -07:00
2018-10-22 15:41:12 -07:00
2017-02-08 10:15:26 -08:00
2019-02-21 20:06:04 +01:00
2019-04-24 11:34:53 -07:00
2019-04-24 11:34:53 -07:00
2018-10-22 15:41:12 -07:00
2019-04-15 11:38:00 -07:00
2019-05-02 13:57:58 -07:00
2019-04-24 11:34:53 -07:00
2019-04-24 11:34:53 -07:00
2018-09-21 10:42:00 -07:00
2019-02-14 22:05:46 +01:00

logo FlatBuffers

Join the chat at https://gitter.im/google/flatbuffers Build Status Build status

FlatBuffers is a cross platform serialization library architected for maximum memory efficiency. It allows you to directly access serialized data without parsing/unpacking it first, while still having great forwards/backwards compatibility.

Go to our landing page to browse our documentation.

Supported operating systems

  • Windows
  • MacOS X
  • Linux
  • Android
  • And any others with a recent C++ compiler.

Supported programming languages

  • C++
  • C#
  • C
  • Dart
  • Go
  • Java
  • JavaScript
  • Lobster
  • Lua
  • PHP
  • Python
  • Rust
  • TypeScript

and more in progress...

Contribution

To contribute to this project, see CONTRIBUTING.

Licensing

Flatbuffers is licensed under the Apache License, Version 2.0. See LICENSE for the full license text.


S
Description
FlatBuffers: Memory Efficient Serialization Library
Readme Apache-2.0 73 MiB
Languages
C++ 37.7%
Rust 13.5%
Swift 7.7%
Python 6.9%
C# 5.6%
Other 28.5%