Currently, when implementing a custom allocator for quickjs the library makes sure that allocations are 64 bit aligned on a 64 bit target and 32 bit aligned on a 32 bit target.
However, the msvc doesn't follow this rule, instead 64bit values are 64 bit aligned even on 32 bit target.
We should change to implementation to reflect this.
Currently, when implementing a custom allocator for quickjs the library makes sure that allocations are 64 bit aligned on a 64 bit target and 32 bit aligned on a 32 bit target. However, the
msvc
doesn't follow this rule, instead 64bit values are 64 bit aligned even on 32 bit target.We should change to implementation to reflect this.