Stephen Hicks

High level response: the person who wrote all this code left Google a while back, and we don't really use ES modules internally at the moment, so this is pretty...

When the esModule is written, there's a check for [`goog.isDocumentLoading_()`](https://github.com/google/closure-library/blob/23056d424c15d31d45fec6df2c08b7c40b031306/closure/goog/base.js#L3311-L3319), and it only sets `defer_ = true` (and uses document.write) if it's currently loading. When the bootstrap is in a...

Okay, and now that repro is also printing (before, after, inside) in Chrome. So now I really have no idea what's going on. I wonder if this use case is...

I see. Thanks for reporting this, and expanding on what you're doing. Unfortunately, `goog.bootstrap` is pretty unmaintained, as you may have gathered from the scant documentation. The short story is...

It would appear to be the case. Looks like Google shut down code.google.com and just redirect to the same-id issues in GitHub, which is obviously wrong. FWIW, it's still available...

Seems like it should be fixable even without a FragmentData API, no? Then again, I haven't look at goog.Uri recently...

Yes and no - the behavior would be welcome, but I'm concerned about over-configurability and API design. The various XHR classes are a tangled mess that have been having their...

Can you point me to a (minimal) example demonstrating the breakage? (If you want to ping me internally, that's fine.) I think I know how to fix it, but if...

Alternatively, if we don't really have a good way to set up a regression test (we don't compile our tests internally, obviously/unfortunately), maybe you could send a PR that you've...

I don't know anything about `@types` - how are these typically kept up to date? What problems did you need to fix by hand?