[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]

Re: [pygame] Closing issue 211 with big-endian CPU test



Hi,

On 15-10-25 04:19 PM, Ian Mallett wrote:
On Sun, Oct 25, 2015 at 4:27 PM, Lenard Lindstrom <len-l@xxxxxxxxx <mailto:len-l@xxxxxxxxx>> wrote:

    Is there a need to continue big-endian support?

âBi-endian architectures are still around (ARM, MIPS in particular). AFAICT little endian won the war, but big endian is still around mainly for compatibility. The Raspberry Pi, for example, is almost universally configured to little endian, despite being a bi processor.
If Pygame is built for one of these platforms configured as big-endian, its big-endian code will be tested then.


Also, notionally, programs could swap endian-ness when operating on certain memory regions for performance improvements. Though, I don't know of any modern programs that do that.
Pygame's endian support is set at compile time.

â

    I mean, we leave in the big-endian specific code, but just leave
    it untested. Worry about it only if the need arises.

I like this solution. Making endianness go both ways with a compile flag isn't really extra effort.

    Lenard Lindstrom

âIan