Hi James, > Hi Manuel, > > On Wed, Sep 20, 2017 at 10:41 PM, <[email protected]> wrote: > > What do you mean? What do you think is wrong? > > I just meant that even on little endian architectures, the consuming > code is reading only 32 bits of a 64-bit value. It doesn't break > anything, because it happens to be reading the "right" 32 bits, and > the "wrong" 32 bits happen to be zero. If they weren't zero, that is > if the code were to process a value greater than 4294967295 (or > 2147483647 for signed values), then the code would do the wrong thing > even on little endian architectures. Thanks for the explanation but I'm not certain to know which code you are refereeing to. The code that you have patched recently?
-- Manuel
