updated README and vmem documentation

README.md

@@ -46,7 +46,7 @@ This library currently contains these functionalities (listed by header files):
 	- Semaphore
 	- Signal
 - [timer.h](include/sx/timer.h): Portable high-res timer, wrapper over [sokol_time](https://github.com/floooh/sokol)
-- [virtual-alloc.h](include/sx/virtual-alloc.h): Portable virtual memory allocator and low-level portable virtual memory management functions
+- [vmem.h](include/sx/vmem.h): Page based virtual memory allocator
 - [math.h](include/sx/math.h): 
 	- Standard floating-point
 	- Vector (2,3,4)
@@ -61,14 +61,18 @@ This library currently contains these functionalities (listed by header files):
 	- Basic file operations
 	- Path manipulation functions (c-string)
 - [bheap.h](include/sx/bheap.h): Binary heap implementation
-- [tlsf-alloc.h](include/sx/tlsf-alloc.h): Tlsf (Two-Level Segregated Fit memory) memory allocator. Wrapper over [Mathew Conte's implementation](http://tlsf.baisoku.org)
 - [simd.h](include/sx/simd.h): portable 128bit simd math intrinsics. currently there are two implementations: reference and SSE. ARM Neon will be added soon.
 - [ringbuffer.h](include/sx/ringbuffer.h): Basic ring-buffer (circular buffer)
 - [lockless.h](include/sx/lockless.h): lockless data structures. 
   - Self-contained single-producer-single-consumer queue
 - [linear-buffer.h](include/sx/linear-buffer.h): Helper custom memory allocator useful for allocating structures with arrays of data in a single allocation call
 - [bitarray.h](include/sx/bitarray.h): utility data structure to hold arbitary number of bits
 
+## Changes
+### v1.0.0 (May 2020)
+- [BREAKING]: Removed `virtual-alloc.h` and replaced it with the new `vmem.h`. virtual-alloc was not Suitable for general memory allocator. So I ditched it completely and replaced with a cleaner, better virtual memory allocation scheme. which is page based
+- [BREAKING]: Removed `tlsf-alloc` and all of it's dependencies. In order to cut the maintainable code and reduce repo size. I decided to remove this and leave it for the user to use the open-source [tlsf allocator](https://github.com/mattconte/tlsf). 
+
 ## Build
 ### Current supported platforms
 

include/sx/vmem.h

@@ -1,30 +1,32 @@
 //
-// Copyright 2018 Sepehr Taghdisian (septag@github). All rights reserved.
+// Copyright 2020 Sepehr Taghdisian (septag@github). All rights reserved.
 // License: https://github.com/septag/sx#license-bsd-2-clause
 //
-// virtual-alloc.h - v1.0 - Virtual memory allocator
+// vmem.h - v1.0 - Virtual memory page allocator
 //
 // Uses OS specific virtual memory functions
-//
-// virtualalloc: standard sequential virtual memory allocator
-//               Virtual memory can be reserved in large sizes and commited later on demand
-//               Very useful for large buffer allocations, instead of allocating directly from heap
-//               Memory is only allocate sequentially and will not be reused
-//               so the limitation is not to hit the 'reserve_sz' max
-//
-//      sx_virtualalloc_init        Initialize and reserves virtual memory for the allocator
-//                                  Reserved size will not be commited until it's alloacated and
-//                                  touched 'reserved_sz' value will grow to OS-specific aligned
-//                                  page size
-//
-//      sx_virtualalloc_release     Releases memory commited and reserved by virtual-allocator
-//
-// NOTE: When you use the sx_alloc object for allocating, the memory will be commited and you will
-//       get a valid pointer. But still the memory is not allocated from heap until you touch the
-//       contents of the memory block.
-//       Also the allocation sizes will grow to OS-specific aligned page size (for example 4k),
-//       .. so try to allocate in larger chunks from this type of allocator
-//
+// the vmem set of functions can be used to reserve and commit protected pages of virtual memory 
+// from the system. It works by reserving maximum number of pages. So the pointer range will be 
+// reserved, but not actually mapped to physical memory.
+// After init, you can use commit_page/commit_pages functions to start using specific pages. 
+// Pages are defined by their index. So, if you have for example, 4096 maximum pages. You can commit
+// and free pages from 0..4095, and you should keep track of free/allocated pages in your program.
+// Page memory size is OS/System dependent. You can actually get it's size with `sx_os_pagesz` 
+//
+//  Functions: 
+//      sx_vmem_get_bytes(num_pages)        returns number of bytes allocated for N pages
+//      sx_vmem_get_needed_pages(bytes)     returns number of needed pages to N bytes
+//                                          you can pass the return value to sx_vmem_init function
+//      sx_vmem_init                        initialize the vmem context and reserve `max_pages`
+//      sx_vmem_release                     Releases the vmem context and frees all memory
+//      sx_vmem_commit_page/pages           Commits `num_pages` that you need to actually allocate 
+//                                          and use. Returns the pointer to the start of pages
+//                                          start_page_id: Zero based index of the page you want
+//                                          num_pages: Number of pages starting from `start_page_id`
+//      sx_vmem_free_page/pages             Free a single or a range of pages to be used later again
+//                                          Parameters are same as commit functions
+//      sx_vmem_commit_size                 Returns total commited bytes. 
+//                                          Basically num_pages*page_size
 #pragma once
 
 #include "sx.h"

sx.sublime-project

@@ -0,0 +1,21 @@
+{
+    "folders": [
+        {
+            "path": ".",
+            "folder_exclude_patterns": [
+                "build",
+                ".vscode",
+                ".clangd",
+                "bin",
+                "build-*"
+            ],
+            "file_exclude_patterns": [
+                ".gitignore",
+                ".travis.yml",
+                ".gitmodules",
+                ".clang-format",
+                "compile_commands.json"
+            ]
+        }
+    ]
+}