The cause of a wide disparity in battery life in Apple’s new MacBook Pro laptops that Consumer Reports found during testing has been discovered and eliminated, Apple said on Tuesday.
The new MacBook Pro was the first such Apple laptop not to receive Consumer Reports’ coveted “recommended” rating after the non-profit product evaluation group said battery life had varied from under four hours to almost 20 hours in its testing.
On Tuesday, Apple said the battery life fluctuations were due to “an obscure and intermittent bug” that related to reloading icons in the Safari web browser on the new laptops. The bug was triggered by a certain setting Consumer Reports used in its tests that is usually only used by web site developers, Apple said. The bug did not occur and battery results were consistently high when the special developer setting was not used, Consumer Reports agreed.
Get Data Sheet, Fortune’s technology newsletter.
“This is not a setting used by customers and does not reflect real-world usage,” Apple said in a statement. Apple (AAPL) said it fixed the bug and that Consumer Reports would rerun the tests.
Consumer Reports said on its website that it was redoing the tests after applying Apple’s bug fix and that the laptops would be given the recommended rating if the battery fluctuations were eliminated.
Consumer Reports also said it had communicated with Apple before publishing its original results last month and noted that many new MacBook Pro owners have complained online about episodes of short battery life.
“Both CR’s findings as well as these consumer posts have caused much discussion and debate in the tech press and on user forums,” the group said. “It’s unclear whether Apple’s bug fix for Safari will impact battery life issues for those individuals.”
But Consumer Reports also defended its continued use of the developer setting, which prevented the Safari browser from caching, or saving to local storage, the web sites visited during the battery tests.
For more on the new MacBook Pro, watch:
“Because people use laptops differently and because their usage can vary from day to day, our battery tests are not designed to be a direct simulation of a consumer’s experience,” the group said. “Rather, we look to control as many variables as possible, then perform a test that gives potential users a reasonable expectation of battery life when a computer’s processors, screen, memory, and antennas are under a light to moderate workload.”
Turning off the browser’s cache “allows us to collect consistent results across the testing of many laptops, and it also puts batteries through a tougher workout,” the group said.