bug 1510785 puts all XBL related stuff including tests under MOZ_XBL directive, and has a patch D45616 that disables all XBL related tests. Bz pointed out that numbers of tests might be relevant for shadow DOM and thus need to be rewritten to use shadow DOM.

It might be a good idea to track this work separately from bug 1510785: the work volume might be large to mix with other stuff and it seems there's no point to keep XBL tests running after we have no XBL binding left in the tree (textbox is the last one bug 1513325), so we can remove them safely without any interference with bug 1510785.

Unblocking Bug 1397874 since I'd like to track this type of work in Bug 1566221 specifically.

Hello, i was browsing the code repo and noticed that there are a lot of tests that declare bindings https://dxr.mozilla.org/mozilla-central/search?q=%3Cbindings+-path%3A*.rs+-path%3A*.cpp&redirect=false

(In reply to Vlad Stanimir from comment #2)

Hello, i was browsing the code repo and noticed that there are a lot of tests that declare bindings https://dxr.mozilla.org/mozilla-central/search?q=%3Cbindings+-path%3A*.rs+-path%3A*.cpp&redirect=false

These are being ported to use shadow dom when applicable in Bug 1510785 / https://phabricator.services.mozilla.com/D45616

(In reply to Brian Grinstead [:bgrins] from comment #3)

(In reply to Vlad Stanimir from comment #2)

Hello, i was browsing the code repo and noticed that there are a lot of tests that declare bindings https://dxr.mozilla.org/mozilla-central/search?q=%3Cbindings+-path%3A*.rs+-path%3A*.cpp&redirect=false

These are being ported to use shadow dom when applicable in Bug 1510785 / https://phabricator.services.mozilla.com/D45616

for some reason the phab didn't show me the whole history of the revision, so if Brendan is about to finish all the conversion there, then let's keep that bug to remove tests that live under MOZ_XBL flag.

Converting is happening in Bug 1583314. We could use this bug to actually remove the tests once we start to rip out XBL code.

At this point we can remove any test that is marked with skip-if = !xbl. This should make further code removal easier.

Depends on D50650

Depends on D50651

Depends on D50652

It could be interesting to look at how code coverage data will change after this, to see if there is opportunity for even more dead code removal.

(In reply to Marco Castelluccio [:marco] from comment #13)

It could be interesting to look at how code coverage data will change after this, to see if there is opportunity for even more dead code removal.

I like that idea! I'm planning to track any code removal bugs that we can do in Bug 1566221. Could you help track down the changes in ccov? I was planning to land this in pieces but I could hold off and only push dom/ changes until once layout/ is ready if that makes it easier.

I thought about it some more, and I think we've already landed the changeset that would have caused the ccov changes in Bug 1583314. At that point we disabled MOZ_XBL so all of these tests became effectively skip-if=true. So it'd be interesting to see if we detect the GetWholeText removal (in Bug 1591588) in ccov after those changesets landed.

https://hg.mozilla.org/mozilla-central/rev/2f9d771a4e4d
https://hg.mozilla.org/mozilla-central/rev/0c434beb621c
https://hg.mozilla.org/mozilla-central/rev/ccbd65a92619

(In reply to Brian Grinstead [:bgrins] from comment #15)

I thought about it some more, and I think we've already landed the changeset that would have caused the ccov changes in Bug 1583314. At that point we disabled MOZ_XBL so all of these tests became effectively skip-if=true. So it'd be interesting to see if we detect the GetWholeText removal (in Bug 1591588) in ccov after those changesets landed.

The if branch was taken before 1 and not taken after 2.

From a quick look at the results of a quick and dirty script, I also see a few differences in dom/base/FragmentOrElement.cpp and dom/base/nsGlobalWindowInner.cpp (e.g. SetXBLInsertionPoint 3 is covered before and not after). Analyzing the differences is a bit of a manual process though, as clearly there are a few differences that are simply due to intermittency and a few other which are due to other changes which landed at the same time.

https://hg.mozilla.org/mozilla-central/rev/f86d070215fd