Move the pattern matching from gensquashfs to dir_tree_iterator_t

A simple unit test is added that mainly checks for the behavior of
recursing into a sub-tree and only matching the children at the end,
but not reporting the parents that don't match. The behavior is
inteded to immitate the `find` command.

Signed-off-by: David Oberhollenzer <david.oberhollenzer@sigma-star.at>

3 files changed, 128 insertions, 1 deletions

diff --git a/lib/util/Makemodule.am b/lib/util/Makemodule.am
index 37eac93..15c5b88 100644
--- a/lib/util/Makemodule.am
+++ b/lib/util/Makemodule.am
@@ -96,11 +96,16 @@ test_dir_tree_iterator2_LDADD = libutil.a libcompat.a
 test_dir_tree_iterator2_CPPFLAGS = $(AM_CPPFLAGS)
 test_dir_tree_iterator2_CPPFLAGS += -DTESTPATH=$(top_srcdir)/lib/util/test/testdir
 
+test_dir_tree_iterator3_SOURCES = lib/util/test/dir_tree_iterator3.c
+test_dir_tree_iterator3_LDADD = libutil.a libcompat.a
+test_dir_tree_iterator3_CPPFLAGS = $(AM_CPPFLAGS)
+test_dir_tree_iterator3_CPPFLAGS += -DTESTPATH=$(top_srcdir)/lib/util/test/testdir
+
 LIBUTIL_TESTS = \
 	test_str_table test_rbtree test_xxhash test_threadpool test_ismemzero \
 	test_canonicalize_name test_filename_sane test_filename_sane_w32 \
 	test_sdate_epoch test_hex_decode test_base64_decode test_dir_iterator \
-	test_dir_tree_iterator test_dir_tree_iterator2
+	test_dir_tree_iterator test_dir_tree_iterator2 test_dir_tree_iterator3
 
 check_PROGRAMS += $(LIBUTIL_TESTS)
 TESTS += $(LIBUTIL_TESTS)
diff --git a/lib/util/src/dir_tree_iterator.c b/lib/util/src/dir_tree_iterator.c
index 55fbabb..a63209b 100644
--- a/lib/util/src/dir_tree_iterator.c
+++ b/lib/util/src/dir_tree_iterator.c
@@ -216,6 +216,23 @@ retry:
 		}
 	}
 
+	if (it->cfg.name_pattern != NULL) {
+		if (it->cfg.flags & DIR_SCAN_MATCH_FULL_PATH) {
+			ret = fnmatch(it->cfg.name_pattern,
+				      ent->name, FNM_PATHNAME);
+		} else {
+			const char *name = strrchr(ent->name, '/');
+			name = (name == NULL) ? ent->name : (name + 1);
+
+			ret = fnmatch(it->cfg.name_pattern, name, 0);
+		}
+
+		if (ret != 0) {
+			free(ent);
+			goto retry;
+		}
+	}
+
 	*out = ent;
 	return it->state;
 fail:
diff --git a/lib/util/test/dir_tree_iterator3.c b/lib/util/test/dir_tree_iterator3.c
new file mode 100644
index 0000000..7b14190
--- /dev/null
+++ b/lib/util/test/dir_tree_iterator3.c
@@ -0,0 +1,105 @@
+/* SPDX-License-Identifier: GPL-3.0-or-later */
+/*
+ * dir_tree_iterator2.c
+ *
+ * Copyright (C) 2023 David Oberhollenzer <goliath@infraroot.at>
+ */
+#include "config.h"
+
+#include "util/dir_tree_iterator.h"
+#include "sqfs/error.h"
+#include "util/test.h"
+#include "compat.h"
+
+static int compare_entries(const void *a, const void *b)
+{
+	const dir_entry_t *const *lhs = a;
+	const dir_entry_t *const *rhs = b;
+
+	return strcmp((*lhs)->name, (*rhs)->name);
+}
+
+int main(int argc, char **argv)
+{
+	dir_entry_t *ent[17];
+	dir_iterator_t *dir;
+	dir_tree_cfg_t cfg;
+	size_t i;
+	int ret;
+	(void)argc; (void)argv;
+
+	/********** match name **********/
+	memset(&cfg, 0, sizeof(cfg));
+	cfg.name_pattern = "file_x*";
+
+	dir = dir_tree_iterator_create(TEST_PATH, &cfg);
+	TEST_NOT_NULL(dir);
+
+	for (i = 0; i < 3; ++i) {
+		ret = dir->next(dir, &ent[i]);
+		TEST_NOT_NULL(ent[i]);
+		TEST_EQUAL_I(ret, 0);
+		printf("READ %s\n", ent[i]->name);
+	}
+
+	ret = dir->next(dir, &ent[3]);
+	TEST_NULL(ent[3]);
+	TEST_ASSERT(ret > 0);
+
+	dir = sqfs_drop(dir);
+
+	qsort(ent, 3, sizeof(ent[0]), compare_entries);
+
+	printf("After sort:\n");
+	for (i = 0; i < 3; ++i)
+		printf("%s\n", ent[i]->name);
+
+	TEST_STR_EQUAL(ent[0]->name, "dirb/dirx/file_x0");
+	TEST_ASSERT(S_ISREG(ent[0]->mode));
+	TEST_STR_EQUAL(ent[1]->name, "dirb/dirx/file_x1");
+	TEST_ASSERT(S_ISREG(ent[1]->mode));
+	TEST_STR_EQUAL(ent[2]->name, "dirb/dirx/file_x2");
+	TEST_ASSERT(S_ISREG(ent[2]->mode));
+
+	for (i = 0; i < 3; ++i)
+		free(ent[i]);
+
+	/********** match path **********/
+	memset(&cfg, 0, sizeof(cfg));
+	cfg.name_pattern = "dir*/file_*0";
+	cfg.flags |= DIR_SCAN_MATCH_FULL_PATH;
+
+	dir = dir_tree_iterator_create(TEST_PATH, &cfg);
+	TEST_NOT_NULL(dir);
+
+	for (i = 0; i < 3; ++i) {
+		ret = dir->next(dir, &ent[i]);
+		TEST_NOT_NULL(ent[i]);
+		TEST_EQUAL_I(ret, 0);
+		printf("READ %s\n", ent[i]->name);
+	}
+
+	ret = dir->next(dir, &ent[3]);
+	TEST_NULL(ent[3]);
+	TEST_ASSERT(ret > 0);
+
+	dir = sqfs_drop(dir);
+
+	qsort(ent, 3, sizeof(ent[0]), compare_entries);
+
+	printf("After sort:\n");
+	for (i = 0; i < 3; ++i)
+		printf("%s\n", ent[i]->name);
+
+	TEST_STR_EQUAL(ent[0]->name, "dira/file_a0");
+	TEST_ASSERT(S_ISREG(ent[0]->mode));
+	TEST_STR_EQUAL(ent[1]->name, "dirb/file_b0");
+	TEST_ASSERT(S_ISREG(ent[1]->mode));
+	TEST_STR_EQUAL(ent[2]->name, "dirc/file_c0");
+	TEST_ASSERT(S_ISREG(ent[2]->mode));
+
+	for (i = 0; i < 3; ++i)
+		free(ent[i]);
+
+	return EXIT_SUCCESS;
+}