This commit introduces a new adaptor trait for SeekTarget that works around
frustrating issues with lifetimes. It wraps the arguments in a newtype wrapper
that lives on the stack to avoid the lifetime getting extended to the caller
of the method.
This allows us to introduce a PathSuccessor object that can be passed as the
end argument of remove_between to remove a whole subtree.
Previously, the workspace view was on the stack when we were computing the
available actions, which excluded it. This is a stopgap. We should find a
better solution ASAP.
This avoids issues with the parent view being on the stack when we want to
interact with the delegate from the picker. Still have several picker usages
to convert.
I use unsafe code to transmute an AnyViewHandle to a ViewHandle<T> when
the type matches. Because ViewHandle<T> is repr(transparent) to the
wrapped AnyViewHandle, this is safe.
Any buffers we requested but that haven't been fully sent will cause
outstainding open requests to hang. If we re-request them, any
waiting open requests will resume when the requested buffers finish
being created.
Co-authored-by: Max Brunsfeld <max@zed.dev>
Co-authored-by: Mikayla Maki <mikayla@zed.dev>
It's possible that the host was disconnected when attempting to notify
guests of a file save, so we need to transmit this in order to correctly
update the file's mtime.
Next failing seed OPERATIONS=200 SEED=6894
We need to know the most recent scan id we have actually completed. This is to
handle the case where a guest disconnects when we're in the middle of streaming
worktree entries to them. When they reconnect, they need to report a scan_id
from before we started streaming the entries, because we have no record of when
the stream was interrupted.
Next failure:
SEED=5051 ITERATIONS=1 OPERATIONS=200 cargo test --release --package=collab random -- --nocapture
When a host sends a buffer to a guest for the first time, they record that
they have done so in a set tied to that guest's peer id. When the guest
reconnects and syncs buffers, they do so under a different peer id, so we
need to be sure we track which buffers we have sent them to avoid sending
them the same buffer twice, which violates the guest's assumptions.
We were seeing non-deterministic behavior in randomized tests when
generating backtraces took enough time to cause transactions to group
in some cases, but not group in others.
Tests will need to explicitly opt into grouping if they want it by
setting the interval explicitly. We have tests in the text module that
currently test the history grouping explicitly, but I'm not sure
it's needed elsewhere.
Previously, we were waiting to decrement the invite_count until a user
confirmed their email address, which created weird situations where we would
email people only to have them get a 500 when trying to sign up. Now, we
decrement the invite_count upon sending the email and always honor the
invitation.
Co-Authored-By: Joseph Lyons <joseph@zed.dev>
Co-Authored-By: Max Brunsfeld <max@zed.dev>
A bug caused the assumptions of this method to be violated. We will fix that in the next commit, but we want to be more conservative in our assumptions here going forward.
Co-Authored-By: Antonio Scandurra <me@as-cii.com>
I add /usr/lib/swift as an rpath, which seems to fix the issue even though
there doesn't seem to be a relevant library at that location on my machine.
Based on my research, wondering if `-Wl,-weak-lswiftCompatibilityConcurrency`
is also required for this to work on older OSes, but holding back for now.
If the user already exists, we return none. This will allow the web frontend
to avoid reporting a "join alpha" user event but also not error.
Co-Authored-By: Max Brunsfeld <max@zed.dev>
Co-Authored-By: Joseph Lyons <joseph@zed.dev>
Nathan Sobo
