There is one piece of the puzzle still missing - the files are being
sent wholesale instead of being streamed - and this might cause memory
issues. I haven't benchmarked yet, leaving this until we get some
response from the upstream issue
https://github.com/electron/electron/issues/41872 (hopefully it's just
some thing I missed).
But otherwise, ran export on a trivial library and it worked fine, so
everything is at least hooked up properly now.
## Description
Ran into error after clearing search bar of queries by hitting
backspace.
> The error is shown below
> Thank you to the team for considering
## Tests
**Error**
**Fix**
Fixed by wrapping the line-causing error around by an `if` block.
Attempt to solve the following error in the browser's console when trying to
make the request:
[error] download and save failed: TypeError: Failed to construct 'Request':
The `duplex` member must be specified for a request with a streaming body
This is only needed for ML indexing, let us revisit this once the ML flow is
done instead of adding a premature and possibly disk-space heavy optimization.
Fixes: https://github.com/ente-io/ente/issues/1440
There were two issues:
- It was a backward incompatible change to change the scheme from ente:// to
next://. Revert.
- We also need to setup a top level redirect when the user presses back without
making any changes to their subscription.
**Tested by**
Running payments app and desktop app locally, and verifying that after
interactions with the payments endpoint we go back to the desktop app's actual
contents correctly.
