Skip to content

Integer overflow in DataInterface.js #42

@ZoranRavic

Description

@ZoranRavic

The forceReadVint() function doesn't make sense.

You have (result << 32) | getUint32() which doesn't make sense in javascript, since binary operations work on 31-bit signed integers.

This is because js uses 32-bit integers where 1 bit is reserved.

Also in javascript doing my_int << 32 actually returns my_int, as if you didn't do anything.
This is sometimes used to convert float to int, where doing 5.67 << 32 would return 5.

Note that (result << 16) | getUint16() also wont work, because it would require a 32-bit unsigned integer (16 bits for current result value and another 16 bits that you are adding), which you don't have.

I'm guessing you copied this code from a language that had support for 64-bit integers.

You can make it work by replacing << with multiplication and | with addition, which would treat the number as a double float.
This wont give you a full 64-bit int, but that shouldn't matter. Alternative would be to use BigInt.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions