This only affected tests cases with more than one any possible value (marked by purple italics) where those same values were involved in value pair rules. The algorithm that ensured that selected values replacing any values did not violate value pair rules was a one-pass algorithm, so it was not able to adjust other any value replacements, if those values were the cause of there being 'no possible value' that could be used. The algorithm now takes multiple passes, each pass with variability in the order it attempts to replace values, to ensure if there is a set of any value replacements that can work, it will be found.
Thanks to many users for pointing this issue out over the last few months, but special thanks to Cortney for identifying and isolating a clean cut case that led us to a definitive fix.
Before this action did not clear the prompt text. Now it does.
Many additional, subtle improvements to the Auto-Scripts UI around step re-ordering, handling of the scroll position when adding steps, and auto-advancing the next parameter to be inserted.
There were some rare corner cases that would result in gaps in step number ordering or having a 'Step 0'. These have been resolved.
The presence of requirements no longer limits you to computing just 2-way tests.
You'll now see an indication that a requirement is satisfied by a test case in the Create Tests screen, not just in exports.
In some cases of any's, and nones and compound expected results, the expected result would be applied when it shouldn't be. This was not the expected result.
Considerable improvements and fixes in how defining auto-scripts works, particularly in the area of fast and efficient defaults and keyboard handling.
A classic pair-wise defect only triggered by the interaction of ranged values (a rarely used feature) and expected results or requirements that included that value (an advanced feature).
Extra blank rows in this sheet would cause an import failure. Thanks to Ed for finding and reporting this issue.
Sometimes additional parameter values were bolded.
They used to be added only the first time the requirement was satisfied.
If a plan has enough requirements and a small enough number of values on a parameter to get an any value for that parameter in the requirement then the computed test data was failing to parse and causing the completed computation to appear to take forever.
This was a classic pair-wise failure, only triggered by relatively high number of requirements and an unusually low number of values in an early parameter.
This is now working properly.
This was a classic pair-wise failure where "any" value expected results work fine and importing expected results from Excel works fine, but the combination of the two resulted in a defect.
Any modification to a parameter involved in an "any" expected result would give an expected result deletion warning. These warnings were not accurate as nothing short of removing the parameter will invalidate an "any" expected result.
Expected results from the expected results sheet in the Hexawise Excel file are now imported when importing a plan from Excel. Export a plan with expected results to see the format of the expected result sheet.
Requirements from the requirements sheet in the Hexawise Excel file are now imported when importing a plan from Excel. Export a plan with requirements to Excel to see the format of the expected result sheet.
Some additional cases of reserved XML characters causing Excel format problems. Thanks to Justin G. for pointing out the issue.
'Any value' paramater values used in test case title and description in HP QC exports now show their selected value rather than 'Any value'. Thanks to Justin G. for pointing out the issue.
Parameter values that contribute to a requirement or expected result being met are bolded by Hexawise. When the value was in a defined numeric range, the value bolding was omitted. Now it's not.
Thanks to Justin G. for the suggestion.
Tooltips tell you where each Expected Result came from.
Works now. Thanks to Justin G. for pointing out the issue.
Test case numbering starts with a leading 0 to match HP QC's expectations. Thanks to Justin G. for the suggestion.
Can no longer go to advanced tabs (Expected Results, Requirements and Auto-Scripts) without at least 2 parameters defined to avoid confusion.