Bugzilla

(In reply to Shane Caraveo (:mixedpuppy) from comment #15)
> @enndeakin or @gijs
> 
> Is there any reason not to add about:blanktab to handle this situation?  We have a behavior regression and we cannot allow extensions to inject into about:newtab.

A dedicated about: page was the original suggestion in bug 1676492 ; I don't recall why Neil chose to stick with adding a different mode to about:newtab instead, so I'll leave him to clarify that.

I'll add that if we do use a separate URL it may be wise to avoid a URL like `about:blanktab` (or, sadly, my preference of [`about:blanketyblank`](https://en.wikipedia.org/wiki/Blankety_Blank) because they prefix-match with `about:blank`. Of course in theory all about: page matching should be on full URLs minus query/ref bits, in practice because about URLs IIRC are not `nsStandardURL`s and don't QI to `nsIURL`, query parsing has to be done separately, and sometimes people are lazy (`url.startsWith(...)`) and so there may be lurking bugs in that area.

(In reply to Shane Caraveo (:mixedpuppy) from comment #15)
> @enndeakin or @gijs
> 
> Is there any reason not to add about:blanktab to handle this situation?  We have a behavior regression and we cannot allow extensions to inject into about:newtab.

A dedicated about: page was the original suggestion in bug 1676492 ; I don't recall why Neil chose to stick with adding a different mode to about:newtab instead, so I'll leave him to clarify that.

I'll add that if we do use a separate URL it may be wise to avoid a URL like `about:blanktab` (or, sadly, my preference of [`about:blanketyblank`](https://en.wikipedia.org/wiki/Blankety_Blank)) because they prefix-match with `about:blank`. Of course in theory all about: page matching should be on full URLs minus query/ref bits, in practice because about URLs IIRC are not `nsStandardURL`s and don't QI to `nsIURL`, query parsing has to be done separately, and sometimes people are lazy (`url.startsWith(...)`) and so there may be lurking bugs in that area.