Attempt at fixing mpmc_bounded_queue on 32bit where integer overflow is a concern.

[toffaletti]

I'm not super happy with this, I think I've probably missed some edge case and the logic is not at all straight forward. Maybe I should put the overflow logic into its own 'else if' and heavily comment it? Suggestions welcome.

P.S. - I couldn't actually compile this code because master doesn't build for me for some reason. I get:

error: linking with `cc` failed with code 1
note: cc arguments: -L/home/jason/rust/x86_64-unknown-linux-gnu/stage1/lib/rustc/x86_64-unknown-linux-gnu/lib -m64 -o x86_64-unknown-linux-gnu/stage1/lib/rustc/x86_64-unknown-linux-gnu/lib/librustdoc-a437806a76c5f37a-0.9-pre.so x86_64-unknown-linux-gnu/stage1/lib/rustc/x86_64-unknown-linux-gnu/lib/lib.o -L/home/jason/rust/x86_64-unknown-linux-gnu/stage1/lib/rustc/x86_64-unknown-linux-gnu/lib -lstd-6c65cf4b443341b1-0.9-pre -L/home/jason/rust/x86_64-unknown-linux-gnu/stage1/lib/rustc/x86_64-unknown-linux-gnu/lib -lextra-a7c050cfd46b2c9a-0.9-pre -L/home/jason/rust/x86_64-unknown-linux-gnu/stage1/lib/rustc/x86_64-unknown-linux-gnu/lib -lsyntax-64629f7f0c6a9bc-0.9-pre -L/home/jason/rust/x86_64-unknown-linux-gnu/stage1/lib/rustc/x86_64-unknown-linux-gnu/lib -lrustc-d3cb8c2ccd84a7a7-0.9-pre -lsundown -Lx86_64-unknown-linux-gnu/rt/sundown -L/home/jason/rust/.rust -L/home/jason/rust -shared -lrt -ldl -lm -lmorestack -lrustrt -Wl,-rpath,$ORIGIN/. -Wl,-rpath,/home/jason/rust/x86_64-unknown-linux-gnu/stage1/lib/rustc/x86_64-unknown-linux-gnu/lib -Wl,-rpath,/usr/local/lib/rustc/x86_64-unknown-linux-gnu/lib
note: cc: error: x86_64-unknown-linux-gnu/stage1/lib/rustc/x86_64-unknown-linux-gnu/lib/lib.o: No such file or directory

error: aborting due to previous error

I've tried make clean and ./configure again as well as rm -rf tmp x86_64-unknown-linux-gnu.

[thestinger]

There are checked_{mul,sub,add} methods if you just need to do multiplication/addition/subtraction with an overflow check.

[huonw]

(@alexcrichton is fixing the build failure as we speak.)

[huonw]

Can you just use uint::next_power_of_two rather than importing it specially?

[toffaletti]

Easy enough

[toffaletti]

I feel better about this now. The changes are commented and I think the checks are more clear.

[catamorphism]

@brson @toffaletti Is this ready to merge?

[toffaletti]

@catamorphism I think it is ready now.

[brson]

I still don't understand this change, and I think somebody needs to in order to merge it.

[alexcrichton]

As with @brson, I'm uncomfortable merging this when it appears that there aren't that many people understanding what's going on here.

I would be comfortable landing this is someone else can be convinced of the correctness of the overflow handling, or possibly also if tests were added which exposed the bug as it stands today and shows that the changes fix it.

[alexcrichton]

Closing due to a lack of activity, but as always feel free to reopen with updates!

[toffaletti]

Sorry, I've been out of the country for a few weeks. Can revisit this if it ever becomes an issue.