Another interesting thing is that someone else reported what looks like the
same segfault, but this time with MSYS2:
https://github.com/google/protobuf/issues/2380
On Tue, Nov 15, 2016 at 3:54 PM, David Edge wrote:
> I will try and figure out where the segfault is happening and get back to
> you
I will try and figure out where the segfault is happening and get back to
you.
My compiler options included the following flag: -fsingle-precision-constant
Turning off this option got rid of all the precision errors. Every test now
passes prior to the segfault. However I still segfault at the i
Interestingly it kind of looks like your compiler is treating
floating-point literals as single-precision floats instead of doubles.
Judging from this snippet you posted for example:
Expected: 1.234567890123456789
Which is: 1.2345678806304932
The first number is the original text from the source