A little dive in the original feature would reveal a bunch of unfortunate time-bombs of behaviors taking dependencies on _write time to the registry_, adding arbitrary retry loops to overcome time resolution, and circumventing the standard path in the platform.
> dependencies on _write time to the registry_, adding arbitrary retry loops to overcome time resolution
For what it's worth, around this time Microsoft's own API (which will only set Edge as default) messed this up occasionally if you ran it on the edge of a minute. When I implemented the equivalent feature for Firefox I checked the time, only one retry instead of Chrome's 5 and it's still not perfect, though my approach doesn't write the potentially out of date values first. That was already sufficiently stable that I couldn't make it fail.
A little dive in the original feature would reveal a bunch of unfortunate time-bombs of behaviors taking dependencies on _write time to the registry_, adding arbitrary retry loops to overcome time resolution, and circumventing the standard path in the platform.
https://chromium-review.googlesource.com/c/chromium/src/+/37...
The conspiracy theory is wild, and seems to be based largely on the changing of the name, but if you read the patch, it's dependent on the file.
The comments on the review that were rejected by the author could easily be the conditions that changed and a plausible cause.
The whole thing smells like sensationalist FUD, particularly given the content of this patch.