I'm working on a new Ruby gem minitest_lucid
that improves the error output from minitest
assertions.
For testing Ruby hashes:
-
Not so good: default
assert_equal
behavior can produce a messy result. -
Better: same assertion, but with added call to
make_my_diffs_pretty!
(actually part ofminitest
itself), produces a better result. -
Best: same assertion, but with added
require 'minitest_lucid'
(the forthcoming gem), produces the best result.
Not So Good
Here's what you get from a failed assert_equal
for two hashes, each with eight entries:
1) Failure:
MinitestLucidTest#test_hash [C:/Users/Burdette/Documents/GitHub/minitest_lucid/test/minitest_lucid_test.rb:28]:
-------- expected
+++ actual
@@ -1 +1 @@
-{:tauro=>"Cia ina do ip ocat doat.", :loquens=>"Dua sarat rad noad maat caea.", :lor=>"Eser in dolo eaata labor ut.", :dolo=>"Ipaat paal doat iruat ala magabor.", :offab=>"Ut dolore ua consal vaba caea.", :moam=>"Sunt sed te coma teu alaaame."}
+{:laboru=>"Laboab vaga dat maaua in venima.", :dolo=>"Ipaat paal doat iruat ala magabor.", :loquens=>"dua sarat rad noad maat caea.", :lor=>"Eser in dolo eaata labor ut.", :tauro=>"cia ina do ip ocat doat.", :amcae=>"Utatu cilaa cit siat commag seqa."}
Better
Here's what you get if you add make_my_diffs_pretty!
to your test:
1) Failure:
MinitestLucidTest#test_hash [C:/Users/Burdette/Documents/GitHub/minitest_lucid/test/minitest_lucid_test.rb:28]:
-------- expected
+++ actual
@@ -1,6 +1,6 @@
-{:tauro=>"Cia ina do ip ocat doat.",
- :loquens=>"Dua sarat rad noad maat caea.",
- :lor=>"Eser in dolo eaata labor ut.",
+{:laboru=>"Laboab vaga dat maaua in venima.",
:dolo=>"Ipaat paal doat iruat ala magabor.",
- :offab=>"Ut dolore ua consal vaba caea.",
- :moam=>"Sunt sed te coma teu alaaame."}
+ :loquens=>"dua sarat rad noad maat caea.",
+ :lor=>"Eser in dolo eaata labor ut.",
+ :tauro=>"cia ina do ip ocat doat.",
+ :amcae=>"Utatu cilaa cit siat commag seqa."}
Best
And here's what you get from minitest_lucid
if you add require minitest_lucid
to your test:
1) Error:
MinitestLucidTest#test_hash:
Exception:
elucidation = {
:missing_pairs => {
:offab => 'Ut dolore ua consal vaba caea.',
:moam => 'Sunt sed te coma teu alaaame.',
},
:unexpected_pairs => {
:laboru => 'Laboab vaga dat maaua in venima.',
:amcae => 'Utatu cilaa cit siat commag seqa.',
},
:changed_values => {
:tauro => {
:expected => 'Cia ina do ip ocat doat.',
:got => 'cia ina do ip ocat doat.',
},
:loquens => {
:expected => 'Dua sarat rad noad maat caea.',
:got => 'dua sarat rad noad maat caea.',
},
},
:ok_pairs => {
:lor => 'Eser in dolo eaata labor ut.',
:dolo => 'Ipaat paal doat iruat ala magabor.',
},
}
C:/Users/Burdette/Documents/GitHub/minitest_lucid/test/minitest_lucid_test.rb:28:in `test_hash'
For the Record
Here's the actual test:
require 'minitest_lucid'
class MinitestLucidTest < Minitest::Test
def test_hash
expected = {
:tauro => 'Cia ina do ip ocat doat.',
:loquens => 'Dua sarat rad noad maat caea.',
:lor => 'Eser in dolo eaata labor ut.',
:dolo => 'Ipaat paal doat iruat ala magabor.',
:offab => 'Ut dolore ua consal vaba caea.',
:moam => 'Sunt sed te coma teu alaaame.',
}
actual = {
:laboru => 'Laboab vaga dat maaua in venima.',
:dolo => 'Ipaat paal doat iruat ala magabor.',
:loquens => 'dua sarat rad noad maat caea.',
:lor => 'Eser in dolo eaata labor ut.',
:tauro => 'cia ina do ip ocat doat.',
:amcae => 'Utatu cilaa cit siat commag seqa.',
}
assert_equal(expected, actual)
end
end
Implementation
The only change to test code needed is to require 'minitest_lucid'
(after, of course, installing the gem once it's released).
Scope
I'm expecting the new gem to treat all collections in the Ruby core and standard libraries:
Hash
Array
Set
Struct
Openstruct
Questions
Please feel free to opine:
Is there a better name for the gem? I definitely want the name to include the word minitest; it could include the word assertions, but then it would need a third word to indicate what the gem does.
I'm thinking to use gem
diff-lcs
for arrays, because it can identify similar but out-of-position sequences. Should there be an option to just do a straight-up slot-for-slot comparison?Are there other collections that should be considered?
Are there scalars (non-collections) whose messages deserve improvement?
Top comments (0)