BLI: factor out BLI_path_has_hidden_component, fix it and speed it up

Factors out utility function "does the path contain any hidden file/folder
components?" function out of innards of UI code (edit_file,
is_hidden_dot_filename) into a BLI function BLI_path_has_hidden_component
and then:
- Adds unit test coverage to it, which uncovered some inconsistencies
- Fix the behavior inconsistencies:
  - A path component that is just a dot (.), was not considered hidden.
    Unless it was the first folder component (now this is fixed).
  - A path component that ended in a tilde (~) was considered hidden.
    Unless it was the first folder component (now this is fixed).
- Speedup the function by not doing several recursive scans over the
  string; instead all the logic is done in a single string scan.

Synthetic HasHiddenComponents_Performance test: 6.0s -> 1.1s for 50M calls
on Mac M1 Max.

More real world test (setup as in #120494): out of whole build_catalog_tree
time, the time taken by BLI_path_has_hidden_component drops from 37% down
to 28%

Pull Request: https://projects.blender.org/blender/blender/pulls/120541

source/blender/blenlib/BLI_path_util.h

@@ -312,6 +312,17 @@ const char *BLI_path_basename(const char *path) ATTR_NONNULL(1) ATTR_WARN_UNUSED
 
 /** \} */
 
+/* -------------------------------------------------------------------- */
+/** \name Path Filtering Utilities
+ * \{ */
+
+/**
+ * Return true if any path component starts with a dot.
+ */
+bool BLI_path_has_hidden_component(const char *path) ATTR_NONNULL(1) ATTR_WARN_UNUSED_RESULT;
+
+/** \} */
+
 /* -------------------------------------------------------------------- */
 /** \name Path Append
  * \{ */

source/blender/blenlib/intern/path_util.cc

@@ -2013,3 +2013,39 @@ int BLI_path_cmp_normalized(const char *p1, const char *p2)
   }
   return result;
 }
+
+bool BLI_path_has_hidden_component(const char *path)
+{
+  bool component_start = true;
+  char cur_char = path[0];
+  char prev_char = '\0';
+
+  while (cur_char != '\0') {
+    char next_char = path[1];
+    /* If we're at a start of path component, current is '.'
+     * and next one is not '.', end or separator: hidden. */
+    if (component_start && cur_char == '.') {
+      if (!ELEM(path[1], '.', '\0', '/', '\\')) {
+        return true;
+      }
+    }
+
+    component_start = ELEM(cur_char, '/', '\\');
+    /* Separator, and previous was tilde: hidden. */
+    if (component_start && prev_char == '~') {
+      return true;
+    }
+
+    path++;
+    prev_char = cur_char;
+    cur_char = next_char;
+  }
+
+  /* Was a tilde right at end of path: hidden. */
+  if (prev_char == '~') {
+    return true;
+  }
+
+  /* Nothing was hidden. */
+  return false;
+}

source/blender/blenlib/tests/BLI_path_util_test.cc

@@ -11,6 +11,8 @@
 #include "BLI_string.h"
 #include "BLI_string_utils.hh"
 
+#define DO_PERF_TESTS 0
+
 /* -------------------------------------------------------------------- */
 /** \name Local Utilities
  * \{ */
@@ -1317,3 +1319,95 @@ TEST(path_util, Contains_Windows_case_insensitive)
 #endif /* WIN32 */
 
 /** \} */
+
+/* -------------------------------------------------------------------- */
+/** \name Tests for: #BLI_path_has_hidden_component
+ * \{ */
+
+TEST(path_util, HasHiddenComponents)
+{
+  /* No hidden components: */
+  EXPECT_FALSE(BLI_path_has_hidden_component(""));
+  EXPECT_FALSE(BLI_path_has_hidden_component(" "));
+  EXPECT_FALSE(BLI_path_has_hidden_component(".."));
+  EXPECT_FALSE(BLI_path_has_hidden_component("../.."));
+  EXPECT_FALSE(BLI_path_has_hidden_component("..\\.."));
+  EXPECT_FALSE(BLI_path_has_hidden_component("a/../b"));
+  EXPECT_FALSE(BLI_path_has_hidden_component("a\\..\\b"));
+  EXPECT_FALSE(BLI_path_has_hidden_component("a"));
+  EXPECT_FALSE(BLI_path_has_hidden_component("a~b"));
+  EXPECT_FALSE(BLI_path_has_hidden_component("a/b"));
+  EXPECT_FALSE(BLI_path_has_hidden_component("a\\b"));
+  EXPECT_FALSE(BLI_path_has_hidden_component("a/b/c"));
+  EXPECT_FALSE(BLI_path_has_hidden_component("a\\b\\c"));
+  EXPECT_FALSE(BLI_path_has_hidden_component("/a/b"));
+  EXPECT_FALSE(BLI_path_has_hidden_component("C:/a/b"));
+  EXPECT_FALSE(BLI_path_has_hidden_component("C:/\a\\b"));
+  EXPECT_FALSE(BLI_path_has_hidden_component("a.txt"));
+  EXPECT_FALSE(BLI_path_has_hidden_component("a/b.txt"));
+  EXPECT_FALSE(BLI_path_has_hidden_component("a\\b.txt"));
+  EXPECT_FALSE(BLI_path_has_hidden_component("/"));
+  EXPECT_FALSE(BLI_path_has_hidden_component("\\"));
+  EXPECT_FALSE(BLI_path_has_hidden_component("a."));
+  EXPECT_FALSE(BLI_path_has_hidden_component("a./b."));
+
+  /* Note: path component that is just a dot is not considered hidden. */
+  EXPECT_FALSE(BLI_path_has_hidden_component("."));
+  EXPECT_FALSE(BLI_path_has_hidden_component("a/."));
+  EXPECT_FALSE(BLI_path_has_hidden_component("a\\."));
+  EXPECT_FALSE(BLI_path_has_hidden_component("a/./b"));
+  EXPECT_FALSE(BLI_path_has_hidden_component("a\\.\\b"));
+  EXPECT_FALSE(BLI_path_has_hidden_component("./a"));
+  EXPECT_FALSE(BLI_path_has_hidden_component(".\\a"));
+
+  /* Does contain hidden components: */
+  EXPECT_TRUE(BLI_path_has_hidden_component(".a"));
+  EXPECT_TRUE(BLI_path_has_hidden_component(".a.b"));
+  EXPECT_TRUE(BLI_path_has_hidden_component("a/.b"));
+  EXPECT_TRUE(BLI_path_has_hidden_component("a\\.b"));
+  EXPECT_TRUE(BLI_path_has_hidden_component(".a/b"));
+  EXPECT_TRUE(BLI_path_has_hidden_component(".a\\b"));
+  EXPECT_TRUE(BLI_path_has_hidden_component(".a/.b"));
+  EXPECT_TRUE(BLI_path_has_hidden_component(".a\\.b"));
+  EXPECT_TRUE(BLI_path_has_hidden_component("a/.b.c"));
+  EXPECT_TRUE(BLI_path_has_hidden_component("a/.b/c.txt"));
+  EXPECT_TRUE(BLI_path_has_hidden_component("a\\.b\\c.txt"));
+
+  /* Tilde at end of path component: considered hidden. */
+  EXPECT_TRUE(BLI_path_has_hidden_component("~"));
+  EXPECT_TRUE(BLI_path_has_hidden_component("a~"));
+  EXPECT_TRUE(BLI_path_has_hidden_component("a/~/c"));
+  EXPECT_TRUE(BLI_path_has_hidden_component("a/b~/c"));
+  EXPECT_TRUE(BLI_path_has_hidden_component("a\\b~\\c"));
+  EXPECT_TRUE(BLI_path_has_hidden_component("~/b"));
+  EXPECT_TRUE(BLI_path_has_hidden_component("a~/b"));
+  EXPECT_TRUE(BLI_path_has_hidden_component("a~\\b"));
+}
+
+#if DO_PERF_TESTS
+
+#  include "BLI_timeit.hh"
+
+TEST(path_util, HasHiddenComponents_Performance)
+{
+  SCOPED_TIMER(__func__);
+  const char *test_paths[] = {
+      "test.txt",
+      "test/a_fairly_long/path/here/shall_we/ok.txt",
+      "test/a_fairly_long/path/here/.with_a_hidden_component/shall_we/ok.txt",
+      "test/.another_path_with_hidden_component/yes.txt",
+      "test/another/path/with/.hidden_filename",
+  };
+  const int RUN_COUNT = 10'000'000;
+  int hidden = 0;
+  for (int i = 0; i < RUN_COUNT; i++) {
+    for (int j = 0; j < ARRAY_SIZE(test_paths); j++) {
+      hidden += BLI_path_has_hidden_component(test_paths[j]) ? 1 : 0;
+    }
+  }
+  EXPECT_EQ(RUN_COUNT * 3, hidden);
+}
+
+#endif
+
+/** \} */

source/blender/editors/space_file/filelist.cc

@@ -614,46 +614,6 @@ void filelist_setsorting(FileList *filelist, const short sort, bool invert_sort)
 
 /* ********** Filter helpers ********** */
 
-/* True if filename is meant to be hidden, eg. starting with period. */
-static bool is_hidden_dot_filename(const char *filename, const FileListInternEntry *file)
-{
-  if (filename[0] == '.' && !ELEM(filename[1], '.', '\0')) {
-    return true; /* ignore .file */
-  }
-
-  int len = strlen(filename);
-  if ((len > 0) && (filename[len - 1] == '~')) {
-    return true; /* ignore file~ */
-  }
-
-  /* filename might actually be a piece of path, in which case we have to check all its parts. */
-
-  bool hidden = false;
-  char *sep = (char *)BLI_path_slash_rfind(filename);
-
-  if (!hidden && sep) {
-    char tmp_filename[FILE_MAX_LIBEXTRA];
-
-    STRNCPY(tmp_filename, filename);
-    sep = tmp_filename + (sep - filename);
-    while (sep) {
-      /* This happens when a path contains 'ALTSEP', '\' on Unix for e.g.
-       * Supporting alternate slashes in paths is a bigger task involving changes
-       * in many parts of the code, for now just prevent an assert, see #74579. */
-#if 0
-      BLI_assert(sep[1] != '\0');
-#endif
-      if (is_hidden_dot_filename(sep + 1, file)) {
-        hidden = true;
-        break;
-      }
-      *sep = '\0';
-      sep = (char *)BLI_path_slash_rfind(tmp_filename);
-    }
-  }
-  return hidden;
-}
-
 /* True if should be hidden, based on current filtering. */
 static bool is_filtered_hidden(const char *filename,
                                const FileListFilter *filter,
@@ -675,7 +635,7 @@ static bool is_filtered_hidden(const char *filename,
 
 #ifndef WIN32
   /* Check for unix-style names starting with period. */
-  if ((filter->flags & FLF_HIDE_DOT) && is_hidden_dot_filename(filename, file)) {
+  if ((filter->flags & FLF_HIDE_DOT) && BLI_path_has_hidden_component(filename)) {
     return true;
   }
 #endif
@@ -3131,7 +3091,7 @@ static int filelist_readjob_list_dir(FileListReadJob *job_params,
 
 #ifndef WIN32
       /* Set linux-style dot files hidden too. */
-      if (is_hidden_dot_filename(entry->relpath, entry)) {
+      if (BLI_path_has_hidden_component(entry->relpath)) {
         entry->attributes |= FILE_ATTR_HIDDEN;
       }
 #endif