Replaced one-shot ZSTD_compress / ZSTD_decompress with streaming
API (ZSTD_compressStream2 + ZSTD_decompressStream) in both
data_compress and data_decompress. Interface unchanged.
Prepares for true streaming transfer where compression can be
interleaved with transmission.
- Reduced test data from 50MB to 25MB (test.py)
- Increased SSH socketpair buffer to 1MB via setsockopt (transport_ssh.c)
- Made chunk size configurable: new Config field, --chunk-size CLI flag,
sent over wire, and used in scanner instead of DESIRED_CHUNK_SIZE
- Fixed DESIRED_CHUNK_SIZE macro parentheses (chunk.h)
- Merge io.h/c back into protocol.h/c
- Change send_data to take Data* argument
- Remove redundant file_load_data from send_chunk
- Move compression into file_send_single_calls
- Move file_receive from multiprocessing.c to file.c
- Add check_ssh_localhost() that tests SSH connectivity and symlinks
fastsync-server into PATH if needed
- Add SSH_CASES with single-threaded, multithreaded, and compression tests
- Support no_server mode in run_single_test for SSH tests
- Skip SSH tests gracefully when SSH/localhost/fastsync-server unavailable
- Replace send()/recv() with read()/write() + io_set_fds() for fd abstraction
- Add client_connect_ssh() via socketpair + fork/exec; reap child in client_disconnect()
- Add server --stdio flag for SSH invocation
- Replace handle_arg() with rsync-style positional args (fastsync <src> <dest>)
- Add SSH vs TCP auto-detection via is_remote_dest()
- Add TransportType and ssh_destination to Config, fix uninitialized fields in
config_receive() (pre-existing multithreaded crash)
- Sender threads now wait for server STATUS_OK before disconnecting
- Add --help, error handling for sendfile+SSH combos
- Add pre-flight checks and Posix Args test case to test.py
- Add test_config_ssh_dest() unit test
Conflicts resolved:
- config_create signature: added both use_metadata and use_sendfile params
- config_send/config_receive: wire protocol includes both use_metadata and use_sendfile
- client.c: config_create call updated, arg parsing includes both -M/--preserve and -f/--sendfile
- test_config.c: all three config_create calls updated
- file_send_sendfile: use file->data->size instead of removed struct stat
- Remove FileReceive struct; use File everywhere with nullable FileMetadata*
- Remove struct stat from File; file size lives in Data->size (data_create_reserve)
- Add FileMetadata struct (mode, uid, gid, mtime) sent conditionally over wire
- Add -M / --preserve flag to client
- Restore permissions, ownership, timestamps on disk write
- Wire format uses per-file present flag for metadata (zero overhead when off)
- netem_apply now calls netem_reset() first to clear any leftover qdisc
- Rsync daemon startup polls TCP port instead of fixed sleep
- Dest dir cleared at start of run
Previously rsync ran as a direct local copy (rsync -aH source/ dest/), bypassing the
loopback interface entirely. This meant tc netem latency/loss/rate limits were never
applied to rsync, making the comparison fundamentally unfair.
Now rsync goes through the network via an rsync daemon on localhost:
rsync --daemon --no-detach --config=rsyncd.conf
rsync rsync://localhost:PORT/source/ dest/
The daemon is started before rsync tests and killed afterward. Both the rsync client
and FastSync client run under the same systemd-run disk I/O limits when applicable.
TapTap