Henning Koch
Henning Koch
No worries, thanks for the update.
Now that we're dropping IE11 support there may be a way to parse incomplete HTML fragments by using `` elements:  We would need to benchmark this...
This is really a limitation of the [`:layer`](https://unpoly.com/layer) pseudo-selector. It takes the topmost swappable element in your overlay and [derives a selector](https://unpoly.com/up.fragment.toTarget) from it. Since we're preferring class names before...
Needs reproduction. There is a [test ensuring we don't crash](https://github.com/unpoly/unpoly/commit/b2014c7e4038d95a92e85030e0044c560f5f5b73) with undefined `clientX` and `clientY`, but that one passes on Safari 15 and Mobile Safari 15.
Both `:origin` and `&` are supported, as [documented here](https://unpoly.com/origin). Even without `:origin` in the selector, Unpoly 2 tries to [take a given `{ origin }` into account](https://unpoly.com/fragment-placement#interaction-origin-is-considered) when matching fragments....
Re-opening this issue for a while in case someone wants to tackle that change.
I have never used `X-Up-Events` for a `GET` request, which is the only kind of request that is being cached.
What are some other things that would have helped you writing your protocol implemention? I didn't want to make the redirect persistence a requirement for new integrations, as I found...
The Jasmine specs are just a bunch of static HTML and JS files. Network communication to the server is mocked.
Wow @angelikatyborska, that must be the best prepared issue I've ever received :smile: Thanks for the detailed report and reproduction repo. I believe what happens is this: - When Unpoly...