Bugzilla

See slack DM for some musings; but tl;dr, that seems maybe-fine, if we can get a fix for the other bug landed very soon (though I don't know how feasible that is, without knowing contenteditable code/edge-cases super-well; and I don't want you/reviewers to be under pressure to get that fix written/reviewed/landed).

Whatever route we take, the primary goals at this point are:
* to address the Nightly usability-regression ASAP, so that gmail, gdoc-comments, fastmail, etc. aren't broken (bug 1925606, bug 1926087) and users aren't inadvertently sending emails with missing spaces and/or being frustrated by mysteriously disappearing spaces while typing emails.
* ...and to avoid shipping that regression to beta (next week).

I think that means we have to do one of the following in the next day or two (the sooner the better): either back out this bug's code-changes, or land a fix for bug 1925635.  (Either of those should address the regression.)

The backout option seems trivially safe; we could leave the WPT changes in-place if it's administratively annoying to back them out, and we could just annotate them as known-failures or skipped.  (That's a bit of manual work but probably not too much trouble.)

Whereas I'm less sure about the timeline and risk for the "land a fix for bug 1925635" option.  Hence, I'd personally lean towards the backout option, unless you've got an extremely-high-confidence patch nearly ready for bug 1925635.

See slack DM for some musings; but tl;dr, that seems maybe-fine, if we can get a fix for the other bug landed very soon (though I don't know how feasible that is, without knowing contenteditable code/edge-cases super-well; and I don't want you/reviewers to be under pressure to get that fix written/reviewed/landed).

Whatever route we take, the primary goals at this point are:
* to address the Nightly usability-regression ASAP, so that gmail, gdoc-comments, fastmail, etc. aren't broken (bug 1925606, bug 1926087) and Nightly users aren't inadvertently sending emails with missing spaces and/or being frustrated by mysteriously disappearing spaces while typing emails for much longer.
* ...and to avoid shipping that regression to beta (next week).

I think that means we have to do one of the following in the next day or two (the sooner the better): either back out this bug's code-changes, or land a fix for bug 1925635.  (Either of those should address the regression.)

The backout option seems trivially safe; we could leave the WPT changes in-place if it's administratively annoying to back them out, and we could just annotate them as known-failures or skipped.  (That's a bit of manual work but probably not too much trouble.)

Whereas I'm less sure about the timeline and risk for the "land a fix for bug 1925635" option.  Hence, I'd personally lean towards the backout option, unless you've got an extremely-high-confidence patch nearly ready for bug 1925635.