There were graphical glitches if the browser window was sized very small.
Thanks to Melissa for letting us know about the issue.
When using the bulk add feature with some parameters that had greater than 8 values, a partial update would be applied before the handy warning dialog. Canceling the add would leave you with some (the small) parameters bulk added, and bypassing the warning would leave you with some parameters added in duplicate.
This is a classic pair-wise defect. To trigger it requires you to have params with > 8 values that occur after your smaller params in the plan, and to perform your adds using bulk add rather than single add.
This demonstrates that pairwise designed tests are only as good as the variability that's been identified. This is probably the most heavily tested part of Hexawise, with a strong battery of tests that have been automated, but the order of the large parameters compared to the small ones was never identified as a variation.
The bottom of the explanatory modal dialog was cut off leaving no white space. White space is back.
There is now a setting in the export dialog to pick your specific version of Microsoft Excel for Excel and HP QC exports so that any line breaks you have in your expected results or auto-scripts will be respected by Excel.
Thanks to Robin and Yue for nearly concurrent prompting in the forums asking for this feature.
Continuing the theme from last week, we revisited the sharing dialog itself and spruced up the UI to be more pleasing to the eye and changed the wording to be easier to understand.
Our automated error reporting alerted us to some uncommon but recurring errors being experienced in the requirements page and these were addressed.
Words more good. Me happy.
Some users of up to date version of the Firefox browser have experienced warning dialogs for the last couple weeks. These are resolved.
Thanks to John for alerting us to this issue.
Sharing plans in Hexawise has always been project based, and while that's not changing, we did make it much easier to share your plans. There is now a share link in the UI that takes you directly to the project sharing options, or if you're in a private plan, gives you the option to make a project for the plan.
A couple of the achievements on the path to Hexawise Guru were not ticking off properly. Thanks to Ed for pointing this out in the forums.
Hexawise.com is sporting a brand new website. Check out the new look!
You can now update a parameter name (change only the parameter name in the single parameter update, not using bulk update) and it will update the parameter reference in any requirements, value pairs and auto-scripts, rather than warning you and then deleting them.
After making a bulk edit with the only change being the order of the parameters, cached generated test cases would be displayed with the wrong parameter headers.
Thanks to Mary for reporting the bug.
A return on investment calculator is now available at http://app.hexawise.com/hexawise-roi.
Improved the performance rendering plans with lots of value pairs.
The project permissions dialog had some weird behavior when mousing around it in Internet Explorer. Classic pair-wise defect.
Thanks to Ravinder for the bug report.
In some cases after doing a bulk edit the parameters would be shown out of order until the browser was refreshed.
This was a classic regression after performance improvements were made to bulk editing. Parameter order wasn't checked in the test cases.
There is now a drop down list in the header of the Auto-Scripts preview test selection panel that allows you to use tests other than 2-way to preview your Auto-Scripts.
When updating your password, there is now a checklist of password requirements to be fulfilled.
Doing a bulk parameter edit on a very large plan is now about 10x faster.
If you were not very relaxed when you using the "Relax" button, you might have noticed the button wasn't disabled and there was no spinner while the system was relaxing. Now there is.
Fixed a rare case where submitting a quiz would fail. It was an unusual 2-way pairing that triggered the defect. Submit a quiz with one a more answers blank, and you have a previous unsuccessful attempt at passing the quiz.
The moral of this defect, consider "history" when designing your test inputs.