Ensure correct file length and sample skip calculations on 32-bit systems by using uint64_t #892
Conversation
…tems by using uint64_t On 32-bit architectures, size_t may not be large enough to represent large file sizes or sample counts. This commit updates the file length computation and the return type of samplesToSkip() to use uint64_t, allowing accurate handling of large files and ensuring consistency across platforms.
|
Is there a failing test that this fixes?
|
|
Hi @jwmelto , I found some references, for example: https://stackoverflow.com/questions/918787/whats-sizeofsize-t-on-32-bit-vs-the-various-64-bit-data-models, which indicate that the C++ standard allows implementations to choose the representation of size_t, with range up to SIZE_MAX, which must be at least 65,535. To verify this, I printed the size of size_t and the value of SIZE_MAX on both platforms: PC (x86-64): ARM (32-bit): Replacing size_t with uint64_t solved this issue. |
Ensure correct file length and sample skip calculations on 32-bit systems by using uint64_t
On 32-bit architectures, size_t may not be large enough to represent large file sizes or sample counts. This update modifies the file length computation and the return type of samplesToSkip() to use uint64_t, allowing accurate handling of large files and ensuring consistency across platforms.