server/internal/registry: take over pulls from server package (#9485)

This commit replaces the old pull implementation in the server package
with the new, faster, more robust pull implementation in the registry
package.

The new endpoint, and now the remove endpoint too, are behind the
feature gate "client2" enabled only by setting the OLLAMA_EXPERIMENT
environment variable include "client2".

Currently, the progress indication is wired to perform the same as the
previous implementation to avoid making changes to the CLI, and because
the status reports happen at the start of the download, and the end of
the write to disk, the progress indication is not as smooth as it could
be. This is a known issue and will be addressed in a future change.

This implementation may be ~0.5-1.0% slower in rare cases, depending on
network and disk speed, but is generally MUCH faster and more robust
than the its predecessor in all other cases.

server/internal/client/ollama/trace.go

@@ -13,9 +13,13 @@ type Trace struct {
 	// Update is called during [Registry.Push] and [Registry.Pull] to
 	// report the progress of blob uploads and downloads.
 	//
-	// It is called once at the beginning of the download with a zero n and
-	// then once per read operation with the number of bytes read so far,
-	// and an error if any.
+	// The n argument is the number of bytes transferred so far, and err is
+	// any error that has occurred. If n == 0, and err is nil, the download
+	// or upload has just started. If err is [ErrCached], the download or
+	// upload has been skipped because the blob is already present in the
+	// local cache or remote registry, respectively. Otherwise, if err is
+	// non-nil, the download or upload has failed. When l.Size == n, and
+	// err is nil, the download or upload has completed.
 	//
 	// A function assigned must be safe for concurrent use. The function is
 	// called synchronously and so should not block or take long to run.