Issue #19857 has been updated by ioquatix (Samuel Williams). @jeremyevans0 I'm not proposing any change to `load`, only to `eval` and related methods. However, I actually believe your example is equally confusing because the coverage is not merged. If someone loads the same file several times, it's reasonable to expect coverage to be merged (and probably unexpected that it's not).
I agree with @mame (Yusuke Endoh) that if there is no way to ensure that the same code is evaluated in each evaluation, it is wrong to merge the results.
We have no way to disambiguate coverage, since by its nature, coverage refers to files and line offsets. In practice, After the program executes, coverage is displayed by loading those files and displaying information to the user. If those file paths and line numbers are not consistent, coverage will become meaningless. So if the user does not desire coverage, they should simply avoid setting the path and line number. This kind of consistency is also important for the Ruby debugger, so this is not something specific to coverage.
I'm not necessarily opposed to merging results, but as it is not possible to ensure result merging makes sense, it should be an option
This is already the case: it already has an option, `Coverage.start(eval: true)`. When you select this mode, coverage is already merged. If you have several evals for the same file, they all get merged into one coverage report. What this bug is trying to address, is the fact that when those eval ranges overlap, they get reset. In every case I can think of, this is undesirable and basically not reporting coverage correctly, as lines which are previously executed, with coverage, get reset to 0. ---------------------------------------- Bug #19857: Eval coverage is reset after each `eval`. https://bugs.ruby-lang.org/issues/19857#change-104486 * Author: ioquatix (Samuel Williams) * Status: Open * Priority: Normal * Assignee: ioquatix (Samuel Williams) * Backport: 3.0: DONTNEED, 3.1: DONTNEED, 3.2: REQUIRED ---------------------------------------- It seems like `eval` based coverage is reset every time eval is invoked. ```ruby #!/usr/bin/env ruby require 'coverage' def measure(flag) c = Class.new c.class_eval(<<~RUBY, "foo.rb", 1) def foo(flag) if flag puts "foo" else puts "bar" end end RUBY return c.new.foo(flag) end Coverage.start(lines: true, eval: true) # Depending on the order of these two operations, different coverage is calculated, because the evaluation of the code is considered different, even if the content/path is the same. measure(false) measure(true) p Coverage.result ``` Further investigation is required. -- https://bugs.ruby-lang.org/