Implement basic Web GPU server-side to achieve the draw calls replaying

[yorkie]

This pull request introduces a new GPU command buffer allocation and iteration system, along with assertion handling utilities and an initial implementation of the encoding context for GPU command buffers. The changes are focused on adding robust memory management for GPU command buffers, platform/compiler-specific assertion handling, and foundational code for encoding command buffers. Below are the most important changes grouped by theme.

GPU Command Buffer Allocation and Iteration

• Added new CommandAllocator and CommandIterator classes in command_allocator.cpp and command_allocator.hpp to manage allocation and iteration of GPU command buffer blocks. These classes support efficient memory management, block acquisition, and command traversal, with careful attention to alignment and error handling. [1] [2]
• Updated CMake configuration to use GLOB_RECURSE for collecting all common source files, simplifying build configuration for recursive file inclusion.

Assertion Handling Utilities

• Added platform- and compiler-specific implementation of BreakPoint() and HandleAssertionFailure() in assert.cpp and declared their interfaces in assert.hpp. This enables consistent assertion failure handling and debugging across multiple architectures and compilers. [1] [2]

GPU Command Buffer Encoding Context

• Introduced the initial version of EncodingContext in encoding_context.cpp, providing lifecycle management, command acquisition, and pass management for GPU command buffer encoding. This lays the groundwork for more advanced encoding and validation features.