Roland Illig

Results 83 comments of Roland Illig

IntelliJ knows the following token categories for coloring Go programs: - Bad character - Block comment - Braces - Brackets - Builtin function - Builtin type - Colon - Comma...

+1 This is probably related to #111, since the editor needs to know about the structure of the source code. As a side node, there is already the autocomplete feature,...

Self-contained example, extracted from https://github.com/rillig/pkglint: ~~~go package main import "fmt" type Autofix struct { autofixShortTerm // embedded with implicit name } type autofixShortTerm struct { actions []string } func (a...

[Still reproducible](https://github.com/golangci/golangci-lint/issues/826#issuecomment-564055536): ~~~text $ golangci-lint.exe --version golangci-lint has version v1.45.2 built from (unknown, mod sum: "h1:9I3PzkvscJkFAQpTQi5Ga0V4qWdJERajX1UZ7QqkW+I=") on (unknown) $ golangci-lint.exe run ./main.go main.go:10:2: `actions` is unused (structcheck) actions []string ^...

I thought about the improved diagnostic again, and I'm not going to implement it. * If the new message were to replace the current one, it wouldn't be beneficial to...

I won't implement this "parallel" mode myself, as it looks too complicated to get it right. If you want to prove the opposite, feel free to do so.

The first step, as I mentioned in https://github.com/microsoft/calculator/issues/1993, is to add an assertion at the end of `intrat` that the result is indeed an integer.

The wrong value is calculated in `scidisp.cpp:73`, in `result &= GetChopNumber()`. For some reason, `Rational(8396807.00000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000019316) & 0xffff_ffff_ffff_ffff` results in 0. Thinking further, the many zeroes are already wrong, the rational...

Minimal reproducer: ~~~c++ auto nom = CalcEngine::Rational(1ULL

At the end of `intrat`, there is an assertion missing that the result is actually an integer. The whole approach of doing limited-precision arithmetics including rounding when dealing with perfectly...