Cast in BIT_initCStream

[heroicpixel]

Hi team!
We used AI in writing this issue description.

We noticed a small casting detail in the implementation of BIT_initCStream.
Currently, during the offset subtraction, there is an intermediate cast to core::ffi::c_ulong before casting to isize:

(*bitC).endPtr = ((*bitC).startPtr)
    .add(dstCapacity)
    .offset(-(::core::mem::size_of::<BitContainerType>() as core::ffi::c_ulong as isize));

On 64-bit Windows (LLP64 data model), core::ffi::c_ulong is 32-bit whereas isize is 64-bit. Casting the usize returned by size_of to a 32-bit c_ulong before converting it to a 64-bit isize could technically result in truncation if the type size were ever extremely large.

Would it be possible to safely simplify this by casting directly to isize?

To Reproduce
N/A

Expected behavior
N/A

Screenshots and charts
N/A

Desktop (please complete the following information):

• OS: Windows (64-bit LLP64 targets)
• Compiler: MSVC / Rustc targeting x86_64-pc-windows-msvc
• Build system: cargo / build.rs

Additional context
Direct link to the line:

libzstd-rs-sys/lib/common/bitstream.rs

Line 33 in 3dc8c85

.offset(-(::core::mem::size_of::<BitContainerType>() as core::ffi::c_ulong as isize));

Thank you so much for your time and hard work on this project!

[bjorn3]

BitContainerType will never be large enough to not fit in either c_ulong or isize.