From 23432a3add6a745a1789d5db591b299a1f452dac Mon Sep 17 00:00:00 2001
From: scivision <scivision@users.noreply.github.com>
Date: Tue, 1 Oct 2024 22:12:25 -0400
Subject: [PATCH] non-java normalize, more than 10x faster

---
 +stdlib/canonical.m   |  2 +-
 +stdlib/normalize.m   | 46 +++++++++++++++++++++++++++++++++++++++++--
 +stdlib/relative_to.m |  4 ++--
 3 files changed, 47 insertions(+), 5 deletions(-)

diff --git a/+stdlib/canonical.m b/+stdlib/canonical.m
index 947e7fa..36127cf 100644
--- a/+stdlib/canonical.m
+++ b/+stdlib/canonical.m
@@ -45,7 +45,7 @@
   else
     % for non-existing path, return normalized relative path
     % like C++ filesystem weakly_canonical()
-    c = stdlib.normalize(c);
+    c = stdlib.normalize(c, use_java);
     return
   end
 end
diff --git a/+stdlib/normalize.m b/+stdlib/normalize.m
index 2caac93..29d22ca 100644
--- a/+stdlib/normalize.m
+++ b/+stdlib/normalize.m
@@ -1,4 +1,4 @@
-function n = normalize(p)
+function n = normalize(p, use_java)
 %% normalize(p)
 % normalize(p) remove redundant elements of path p
 % path need not exist, normalized path is returned
@@ -10,9 +10,51 @@
 % https://docs.oracle.com/en/java/javase/21/docs/api/java.base/java/nio/file/Path.html#normalize()
 arguments
   p (1,1) string
+  use_java (1,1) logical = false
 end
 
-n = stdlib.posix(java.io.File(p).toPath().normalize());
+if use_java
+  n = stdlib.posix(java.io.File(p).toPath().normalize());
+else
+
+  n = stdlib.posix(p);
+
+  % use split to remove /../ and /./ and duplicated /
+  parts = split(n, "/");
+  i0 = 1;
+  if startsWith(n, "/")
+    n = "/";
+  elseif ispc && strlength(n) >= 2 && isletter(extractBetween(n, 1, 1)) && extractBetween(n, 2, 2) == ":"
+    n = parts(1);
+    i0 = 2;
+  else
+    n = "";
+  end
+
+  for i = i0:length(parts)
+    if parts(i) == ".."
+      if n == ""
+        n = parts(i);
+      elseif endsWith(n, "..")
+        n = n + "/" + parts(i);
+      else
+        j = strfind(n, "/");
+        if isempty(j)
+          n = "";
+        else
+          n = extractBefore(n, j(end));
+        end
+      end
+    elseif all(parts(i) ~= [".", ""])
+      if n == ""
+        n = parts(i);
+      else
+        n = n + "/" + parts(i);
+      end
+    end
+  end
+
+end
 
 if(strlength(n) == 0), n = "."; end
 
diff --git a/+stdlib/relative_to.m b/+stdlib/relative_to.m
index 56d94a4..6033363 100644
--- a/+stdlib/relative_to.m
+++ b/+stdlib/relative_to.m
@@ -5,8 +5,8 @@
 end
 
 % must remove trailing slashes
-base = stdlib.normalize(base);
-other = stdlib.normalize(other);
+base = stdlib.normalize(base, true);
+other = stdlib.normalize(other, true);
 
 if base == other
   r = ".";