├── .github
    └── FUNDING.yml
├── README.md
└── dlmalloc.cc


/.github/FUNDING.yml:
--------------------------------------------------------------------------------
 1 | # These are supported funding model platforms
 2 | 
 3 | github: greg7mdp
 4 | patreon: # Replace with a single Patreon username
 5 | open_collective: # Replace with a single Open Collective username
 6 | ko_fi: # Replace with a single Ko-fi username
 7 | tidelift: # Replace with a single Tidelift platform-name/package-name e.g., npm/babel
 8 | community_bridge: # Replace with a single Community Bridge project-name e.g., cloud-foundry
 9 | liberapay: # Replace with a single Liberapay username
10 | issuehunt: # Replace with a single IssueHunt username
11 | otechie: # Replace with a single Otechie username
12 | custom: # Replace with up to 4 custom sponsorship URLs e.g., ['link1', 'link2']
13 | 


--------------------------------------------------------------------------------
/README.md:
--------------------------------------------------------------------------------
 1 | # dlmalloc
 2 | A C++ version of Doug Lea's excellent malloc() implementation.
 3 | 
 4 | I have been having issues with the malloc() implementation on Windows, where the memory usage can grow unreasonably large with some patterns of successive reallocations. So I decided to investigate Doug Lea's [malloc implementation](http://g.oswego.edu/dl/html/malloc.html). It is very easy to use, a [single C file](ftp://g.oswego.edu/pub/misc/malloc.c) that you just compile and link along with your source code, and you are all set.
 5 | 
 6 | Using Doug's malloc.c worked beautifully. For my use case, it was faster than Windows default malloc (I'm using Visual C++ 2015), and the memory usage was considerably reduced.
 7 | 
 8 | I wanted to understand the code, as I would like to use its core logic to make a custom C++ allocator for my [sparse hash table](https://github.com/greg7mdp/sparsepp). However, I found the source code very difficult to follow. A lot of it is implemented with preprocessor macros, and is invisible to debuggers. Macros are of course necessary when coding in C for ultimate performance, but they sure obfuscate the code.
 9 | 
10 | Also, I'd like my C++ allocator to be header only, but shoving all these macro definitions into a header file would pollute the global name space, which is not acceptable.
11 | 
12 | Therefore I decided to convert Doug's original malloc.c (v2.8.6) to C++, and this is what you find here. I have not modified what the code does, just how it is specified and organized (modulo any typo or bug I may have introduced of course).
13 | 
14 | Interestingly, the C++ version is just as fast as the original C version! In my tests (Visual C++ 2015, Windows 10), the C++ version was even very slightly faster.
15 | 
16 | While Linus may be opposed to C++, I still though that sharing this C++ implementation could be useful for people who, like me, would like to examine the inner workings of one of the best malloc implementation ever written.
17 | 
18 | ### License
19 | 
20 | Both the original source code from Doug Lea, and my changes are released to the public domain, as explained at http://creativecommons.org/publicdomain/zero/1.0/.
21 | 


--------------------------------------------------------------------------------