feat: add --files-from/-0, --filter, -C and -F filter layer
CI / lint (pull_request) Successful in 23s
CI / sanitizers (address) (pull_request) Successful in 40s
CI / sanitizers (undefined) (pull_request) Successful in 39s
CI / fuzz-build (pull_request) Successful in 17s
CI / coverage (pull_request) Successful in 34s
CI / build-and-test (pull_request) Successful in 1m28s
CI / valgrind (pull_request) Successful in 36s

Implement the RSYNC_COMPAT Phase-2 filter/parser feature group:
- --files-from=FILE (repeatable) plus -0/--from0 NUL delimiters: parse the
  source file list relative to the source root into a shared read-only
  allow-set; the scanner transfers listed files and the whole subtree of
  listed directories and prunes everything else in single- and
  multithreaded mode. Absolute/'..' entries and missing files are hard
  CLI errors.
- --filter=RULE: rsync-style +/- rules (anchored '/', dir-only trailing '/',
  word include/exclude forms) evaluated first-match-wins with a default of
  include, as an independent layer from legacy --exclude/--include.
  Unsupported directives (merge/hide/... ) are rejected explicitly. -f stays
  sendfile.
- -C/--cvs-exclude: well-known rsync CVS default exclude set.
- -F: per-directory .rsync-filter files read during traversal and applied to
  the owning directory's subtree (single + parallel), never transferred.
- Delete manifest still derives from what was actually sent.

Client-only config fields; no wire/protocol change. Adds unit coverage
(CLI parse, allow-set and filter scanning single+parallel) and integration
tests (TestFilesFrom, TestFilters). RSYNC_COMPAT matrix rows updated:
5 rows move to Implemented (Summary 62/3/5/1/76 = 147).
This commit is contained in:
2026-09-06 13:43:34 +02:00
parent 8577f95550
commit f4c15a7b78
15 changed files with 1917 additions and 64 deletions
+23
View File
@@ -2,6 +2,8 @@
#define SCANNER_H
#include "chunk.h"
#include "file_list.h"
#include "filter.h"
#include "protocol.h"
#include "queue.h"
#include <dirent.h>
@@ -27,8 +29,18 @@ typedef struct {
bool copy_unsafe_links;
bool checksum;
bool one_file_system;
/* Phase 2 (files-from / filter layer). All pointers are shared read-only
* across scanner instances and worker threads; ownership stays with the
* caller (client_send). */
const FileListSet* file_list; /* --files-from allow-set, or NULL */
const FilterRuleList* base_filters; /* command-line + -C rules, or NULL */
bool per_dir_filters; /* -F: read .rsync-filter per directory */
} ScannerOptions;
/* Internal per-scanner filter state. FilterNode chains represent the ordered
* per-directory .rsync-filter rules that apply below a directory. */
typedef struct FilterNode FilterNode;
typedef struct {
Queue* directories;
DIR* current_dir;
@@ -51,6 +63,16 @@ typedef struct {
bool one_file_system;
dev_t root_dev;
bool failed;
/* Phase 2 (files-from / filter layer). */
char* root_path; /* transfer root (fs path) for rel computation */
char* current_rel; /* rel path of the open directory ("" == root) */
bool at_seed_dir; /* next open is the seed directory */
FilterNode* seed_node; /* inherited context of the seed dir, or NULL */
FilterNode* current_node; /* filter context of the open directory */
ArrayList* filter_nodes; /* owned FilterNode arena (may be NULL) */
const FileListSet* file_list;
const FilterRuleList* base_filters;
bool per_dir_filters;
} DirectoryScanner;
typedef struct {
@@ -68,6 +90,7 @@ typedef struct {
int completed;
Chunk* initial_chunk;
ProtocolSession* allocation_session;
FilterNode* root_filter_node; /* root .rsync-filter context (owned by ps) */
} ParallelScanner;
DirectoryScanner* directory_scanner_create(const char* root_directory, bool use_metadata,