If this is a regression, when did it regress?

Oh, yes, it doesn't zoom any webpage.  This either happened since the previous official beta build of 54, or it's self-induced I need to reboot for some reason.

Rebooted and am still hitting the issue (phew)

<triage>@adlr do you know why this behavior might have changed?</triage>

This functionality was turned off on purpose in M36.  See  issue 323164 

That bug does not explain it, as it refers to a change in 2014, and I've been using this feature up until the beta where I filed that bug report.

Maybe it was only enabled in the PDF viewer?

You're right, it's probably PDF related.  Switching back to the previous title, since we don't expect the browser to zoom.

WAI?

The browser is definitely working as intended. It seems reasonable for the PDF viewer's behavior to match (I didn't know that it differed).

Does this break a workflow? It was disabled because it was easy to accidentally trigger and left the browser in a weird state with a non-obvious way to get out of it. Maybe that's less of a concern in the PDF viewer (e.g. perhaps users need to zoom more frequently and the UI makes it clearer what's going on).

I certainly used it a lot, since it makes navigating a hardware schematic or block diagram much easier: ctrl-2finger to zoom out, 2finger to navigate, then ctrl-2finger to zoom in on the point where the mouse cursor is.

I understand the consistency argument, although I hope there was more UX study going on behind the scenes than what  issue 323164  suggests, since it reads like the justification was "I don't hear anyone complaining on this obscure bug, so therefore nobody uses it."

I personally use ctrl+2finger zooming much more in the PDF than in general browser scenarios, but that's because I have a touch screen Chromebook and zoom webpages using the screen.  That doesn't work quite as well with PDFs since the increased render detail of browser zoom is needed.

Then there's the lack of pinch-to-zoom on the trackpad.

Basically, our zoom story is generally confusing and needs an overhaul, but we currently just appear to be removing something useful without improving the overall situation.

> I personally use ctrl+2finger zooming much more
> in the PDF than in general browser scenarios ...

It's been disabled in the browser for almost three years, right?

I looked through the recent commit history in //pdf and didn't see any obvious changes related to Ctrl+two-finger-scroll zoom.

Pinch-to-zoom via the touchpad has never been a thing, has it? That seems like it would be way too easy to accidentally trigger while doing two-finger scrolling and doesn't match any existing examples of pinch-to-zoom that I'm aware of (where you're using a touchscreen and effectively interacting directly with the thing you're trying to zoom).

> It's been disabled in the browser for almost three years, right?

Yep.

> I looked through the recent commit history in //pdf and didn't see any obvious changes related to Ctrl+two-finger-scroll zoom.

That's really bizarre; I've definitely been using that.

> Pinch-to-zoom via the touchpad has never been a thing, has it? That seems like it would be way too easy to accidentally trigger while doing two-finger scrolling and doesn't match any existing examples of pinch-to-zoom that I'm aware of (where you're using a touchscreen and effectively interacting directly with the thing you're trying to zoom).

Yes, this never happened, and I think was a conscious decision.  But I wonder when the last time our zoom UX has been considered as a whole...

FYI re: #13, we've enabled pinch zoom before ( issue 569939 ) but reverted it soon after ( issue 627183 ).

So we recently made the PDF viewer aware of pinch zooming so that it scales the content of the PDF, instead of its previous behaviour where the browser would just scale everything, including the control elements, off screen. This was introduced here: https://chromium.googlesource.com/chromium/src/+/6e1abbfb2450eedddb1ab128be1b31cc93104e41

Although that landed after this bug was filed, so it wouldn't be the cause.

Aside from the zooming methods it handles itself (the control elements and now pinches), the PDF viewer relies on the browser. So if the browser wouldn't zoom, neither would the PDF viewer.

Okay, I suppose this is all WAI then.