It should return a bool instead of a PRInt32. In addition, I do not really see the point with the second assert.
Could look up blame on it to see whether there was a point....
http://hg.mozilla.org/mozilla-central/diff/59a540e22014/content/html/content/src/nsHTMLFormElement.cpp Refactored from an if checking for parents and if no parents, calling the method to an asserts with this comment removed: > // We check that the parent of both content nodes is non-null here, because > // newly created form control elements may not be in the parent form > // element's DOM tree yet. This prevents an assertion in CompareTreePosition. > // See Bug 267511 for more information. Looking at the code, this should not happen. At bind to tree, we are attached to the form if we found it as a parent so this method should not be called with an isolated element. I'm not sure if we should keep these asserts because it sounds a bit useless today and I guess it might have been a real issue a few years ago. For control1 != control2, I guess we can keep it, that sounds sane.
You need to log in before you can comment on or make changes to this bug.