> Also, would be neat if the patch was improved to not fail but lower > optimization level and emit a warning.
I liked the idea so I started playing around a bit. While I haven't found a way to achieve that yet (simply setting environment variables doesn't seem to work and there doesn't seem to be a facility for setting arbitrary rustflags from within build.rs) I found out that your patch will likely not work on Debian infra. Apparently, the `cfg` checks in build.rs (and build.rs only) apply to the build machine, not to the target, and on Debian's infra the i386 binaries are usually if not exclusively built on amd64 machines, so the empty main function will be built instead. I'm going to send you a MR to fix this. As for lowering the optimization level from within the build script, I'm not sure this is possible given what I wrote above. If I come up with something I'll send you a MR :-) And if you know a way to do it let me know! > It would be helpful if you could test lower optimization levels than the > aggressive default might succeed. This is zero by the way, any optimization level greater than zero builds unsound code and fails tests.