Recently I discovered a strange bug in dmd -cov, that I finally got some time today to reduce to a smallish test case. However, I can't seem to get rid of the dependency on std.stdio; anybody has any idea how to reduce this code further?

	https://issues.dlang.org/show_bug.cgi?id=17613

This is a very strange bug where decreasing code coverage actually increases the reported coverage percentage. :-P


T

-- 
LINUX = Lousy Interface for Nefarious Unix Xenophobes.

On 07/06/2017 10:58 AM, H. S. Teoh via Digitalmars-d wrote:
> Recently I discovered a strange bug in dmd -cov, that I finally got some
> time today to reduce to a smallish test case. However, I can't seem to
> get rid of the dependency on std.stdio; anybody has any idea how to
> reduce this code further?
>
> 	https://issues.dlang.org/show_bug.cgi?id=17613
>
> This is a very strange bug where decreasing code coverage actually
> increases the reported coverage percentage. :-P
>
>
> T
>

Sorry, not helping but rambling below...

Yep, not surprising that conditional expression has such issues. The last time I checked Python code for coverage, they were missing uncovered branch of the conditional expression.

[And saying this without testing.] Additionally, what about lazily evaluated sub-expressions like function arguments? So, per-line coverage is not sufficient anyway.

Ali