Slather 2.8.5 generates Cobertura file wtih incorrect hit count for condition lines

[sfloam]

When running slather coverage -x --output-directory Coverage/ --arch arm64 --scheme MyScheme whe noticed that the Cobertura file generated by Slather had lines where the hits="0" but the condition-coverage="100%". If the condition-coverage is at 100%, it would have had to have hit that line.

If the hits="0", how can the condition-converage be 100%?

 <line number="23" branch="true" hits="13" condition-coverage="100% (3/3)">
  <conditions>
    <condition number="0" type="jump" coverage="100%"/>
  </conditions>
</line>
<line number="24" branch="false" hits="13"/>
<line number="25" branch="true" hits="13" condition-coverage="0% (0/2)">
  <conditions>
    <condition number="0" type="jump" coverage="0%"/>
  </conditions>
</line>
<line number="26" branch="true" hits="0" condition-coverage="0% (0/2)"> 
  <conditions>
    <condition number="0" type="jump" coverage="0%"/>
  </conditions>
</line>
<line number="27" branch="true" hits="0" condition-coverage="100% (1/1)">  <-------- How can hits=0 with condition-coverage 100%
  <conditions>
    <condition number="0" type="jump" coverage="100%"/>
  </conditions>
</line>

	22 |  for (SomeElement *childNode in node.children) {
	23 |	  if ([childNode.name isEqualToString:SomeThing]) {
	24 |	    [self _doSomething:childNode];
	25 |	  } else {
	26 |	    SomeMacroForPrintingWithIfInIt(@"someMessage %@", childNode.name);
	27 |	  }
	28 |	}	

Note the problematic line is line 27 having a hit=0 with a condiiton-coverage=100%.

[jarrodlombardo-EventBase]

Slather doesn't analyze the coverage data but instead translates what your test environment detected into a report, so the underlying bug (if there is one) is that the test environment is recording incorrect data.

Combining the report and code you provided:

	23 |	  if ([childNode.name isEqualToString:SomeThing]) { // hits=13, 3/3 branches
	24 |	    [self _doSomething:childNode];                  // hits=13
	25 |	  } else {                                          // hits=13, 0/2 branches
	26 |	    SomeMacroForPrintingWithIfInIt(@"someMessage %@", childNode.name); // hits=0, 0/2 branches
	27 |	  }                                                 // hits=0, 1/1 branches

on line 23, the three branches should result in moving the Program Counter (PC) to line 24 or the else on line 25 (Something in line 23 is compound condition, but I don't know what based on your pseudocode). Based on hit count (23, 24, and 25 all have 13 hits), the PC only ever moves to line 24 though, so something is wrong with the branch detection in the test environment here
Line 24 gets hit 13 times and the PC then advances to the beginning of line 25.
The } on line 25 gets hit 13 times and the PC then advances to line 28.
The else on line 25 never gets hit. The possible branch results are moving the PC to line 26 or to after the } on line 27.
Line 26 never gets hit. Something in line 26 is branching, but I don't know what based on your pseudocode.
The } on line 27 never gets hit. The single branch after the } is always/never covered so the condition coverage could be 0 or 100 wihtout any difference in meaning.

The above is based on interpreting your altered code. The only way to 100% understand what is happening here is to look at the actual code, put a breakpoint at the start of line 23 and to carefully step through every time that BP is hit while the tests are running.