Fix bugs related to 64-bit int conversion, fix #212

[mkorbel1]

Description & Motivation

• Fix issues described in 64-bit width LogicValues can be incorrect if their integer value is larger than the maximum positive int value in Dart #212
• Add extension with utility on BigInt for unsigned conversion

Related Issue(s)

Fix #212

Testing

Added new tests covering changes

Backwards-compatibility

Is this a breaking change that will not be backwards-compatible? If yes, how so?

No

Documentation

Does the change require any updates to documentation? If so, where? Are they included?

No

[eric-norige]

There's got to be a better way to get at this; maybe 32 bits at a time can be safely extracted from the bigint through bit shifting?

[mkorbel1]

Sorry, looks like I just barely merged before your comment!

You're right, this appears to work well also, and is probably better? What do you think?

  int toIntUnsigned(int width) {
    if (width > LogicValue._INT_BITS) {
      throw Exception('Cannot convert to BigInt when width $width'
          ' is greater than ${LogicValue._INT_BITS}');
    } else if (width == LogicValue._INT_BITS) {
      const maskWidth = 32;
      final mask = _BigLogicValue._maskOfWidth(maskWidth);
      return (this & mask).toInt() |
          (((this >> maskWidth) & mask).toInt() << maskWidth);
    } else {
      return toInt();
    }
  }

[eric-norige]

Thanks for noticing my late comment. I see no correctness issues in this code. Unfortunately it's a bit more complex than I expected, but I don't see a nice way to clean that up. Despite version control history recording why these changes were made, I still suggest adding a comment explaining that this intentionally complex code is done this way to avoid signed conversion.
Finally, I worry a bit about what happens when this all is run in a javascript environment where _INT_BITS isn't 64. If _INT_BITS isn't even, even adjusting maskWidth won't be able to fix this code. But maybe that's a problem for the future.

[mkorbel1]

Good point, I'll add a comment.

There's already a lot of problems with javascript that I'm noticing. For example, in this PR I disabled the lint checks for large constants already (like 0xffffffffffffffff) that won't work with javascript. I think there's going to be a number of surprises when compiling to js, unfortunately.