We were temporarily using .xml as the extension for exported .opml files as a work around for them not emailing properly with their proper extension. This causes issues with some mind map tools though so we've gone back to the .opml extension and we zip them if they are emailed.
Thanks to Roon for pointing this out.
There was a case where auto-script expected results that referred to the last parameter value could be removed when saving a bulk edit with no change. Thanks to Scott and his team for reporting this.
To simplify requirements traceability (and to boost performance) requirements are now reported as fulfilled once and only once in the first test case that fulfills the requirement.
It is possible in Hexawise to require values that you also indicate are not allowed. Probably the ideal case is to alert you to this fact and not allow it.
What was actually happening was an inconsistent priority as to which should prevail, the requirement or the constraint. This was resolved with the requirement always being fulfilled, even if it violates a constraint.
The link is very useful in explaining why value pairs are sometimes inferred and created for you, so it's good it's working again.
This was an interesting pairwise defect discovered by Scott. If the FIRST thing you ever did on a plan's Auto-Scripts page was enter text into the Start, Step 1 or Finish text areas, AND you did it using only the mouse, never touching the keyboard (right click and paste the text), then the text area would disappear and not be saved. Depending on exactly how you are breaking things down that's a 2-way or 3-way defect, but it's one of those you'd swear wouldn't possibly happen until you replicated it yourself.
Thanks Scott!
When inserting a parameter value replacement with the keyboard or link, it would insert it at the end of the step text rather than in the cursor position. This has been resolved.
The lines in the inferred value pair creation dialog have a little more breathing room and styling to make them easier to parse and read.
Thanks to Scott for requesting the better legibility.
The first step of an Auto-Script did not have the edit icon like the rest of the steps.
Thanks to Amit for reporting this defect.
Once you have 3 or more steps in your Auto-Script, there is a new link in the Auto-Script panel header that allows you to clear out your Auto-Script and start anew without the tedium of deleting every step one by one.
Thanks to Amit for the idea.
Removed the possibility of our most common IE-only JavaScript error from occurring (a stray console.log) by providing a stub that permits it without error.
If you have a lot of parameters in your plan, creating an auto-script can be tedious if the vast majority of the steps have the exact same format.
Relax! Auto-Auto-Scripts are here to take away the tedium.
If you add the parameter name and parameter value replacement (use the insert drop down in the top right of the step) of the very first parameter in your plan to the first and only step of your auto-script, a new button appears that lets your automatically generate your auto-scripts based on the pattern you used in the first step for the first paramater.
Additional steps are created for each additional parameter in your test plan, and these new steps follow the same pattern of language as your first step. Then you can delete any steps that don't apply.
Thanks to Amit for the idea and for testing and feedback on the initial version.
You can click a test case row to highlight it (for contemplation or discussion), but this highlighting was unintentionally "sticky". Now clicking another row will result in just 1 highlight.
There is a tooltip on "No possible value" test case values that explains, in a generic fashion, why there is no possible value. An upcoming enhancement will make this explanation more specific.
Hexawise would silently skip auto-script expected results that did not correspond to the parameters and parameter values specified in the imported plan or otherwise had invalid syntax. Now Hexawise does not allow the import and notifies you exactly where the problem occurred in the import file.
Hexawise would silently skip invalid and married pairs that did not correspond to the parameters and parameter values specified in the imported plan. Now Hexawise does not allow the import and notifies you exactly where the problem occurred in the import file.
There was an interesting pair-wise bug where saving a requirement name or expected result with a % in it, or an auto-script step or expected result with a % in it was broken, but only when saving via the warning dialog that informs you of unsaved changes.
A large set of tests (a couple thousand) would take a minute or more to render the final sets of parameter values once the tests had been generated (they are generated as parameter indexes). In many cases this was actually a fairly significant portion of the total computation time. This rendering time has been reduced about 6x.
This won't be noticeable on typical sets of generated tests (dozens to a couple hundred), but once you get beyond 500 or so generated tests there should be a noticeable improvement.
Now with 2% more accuracy. (That was a self-referential joke.)
Fixed a problem importing exported plans that contained certain special characters. Thanks to Chris for finding this issue and reporting it.
If you had just 1 parameter in a plan, then add more with bulk add, you would not be able to make invalid and married pairs with the original parameter until after a refresh.
Thanks to Scott for finding this one and pointing it out.
If you have already used a reset password link and click it again in your email, the wording of the message is now a bit more helpful.
There was an unintentional decay in how often requests were checked which meant that very often larger compute requests were completed but the the user interface still indicated they were in progress.
There is now a spinner indicating the edit activity is occurring, and the top panel is expanded if it was collapsed, and increased in size if it was sized very small, so that you can see where the editing will occur.