From 41114dd5b04b1087830cf5d55582a04566dbb003 Mon Sep 17 00:00:00 2001
From: crusaderky <crusaderky@gmail.com>
Date: Fri, 12 Dec 2025 16:39:59 +0000
Subject: [PATCH 1/2] Don't obfuscate URLError

---
 ml_datasets/test/test_util.py | 25 +++++++++++++++++++++++++
 ml_datasets/util.py           | 13 +++++++++----
 2 files changed, 34 insertions(+), 4 deletions(-)
 create mode 100644 ml_datasets/test/test_util.py

diff --git a/ml_datasets/test/test_util.py b/ml_datasets/test/test_util.py
new file mode 100644
index 0000000..3c96330
--- /dev/null
+++ b/ml_datasets/test/test_util.py
@@ -0,0 +1,25 @@
+import pytest
+from urllib.error import HTTPError, URLError
+from ml_datasets.util import get_file
+
+
+def test_get_file_domain_resolution_fails():
+    with pytest.raises(URLError, match="test_non_existent_file.*Name or service not known"):
+        get_file(
+            "non_existent_file.txt",
+            "http://test_notexist.wth/test_non_existent_file.txt"
+        )
+
+
+def test_get_file_404_file_not_found():
+    with pytest.raises(HTTPError, match="test_non_existent_file.*404.*Not Found") as e:
+        get_file(
+            "non_existent_file.txt",
+            "http://google.com/test_non_existent_file.txt"
+        )
+    assert e.value.code == 404
+    # Suppress pytest.PytestUnraisableExceptionWarning:
+    #          Exception ignored while calling deallocator
+    # This questionable design quirk comes from urllib.request.urlretrieve,
+    # so we shouldn't shim around it.
+    e.value.close()
diff --git a/ml_datasets/util.py b/ml_datasets/util.py
index b62e105..1cfa797 100644
--- a/ml_datasets/util.py
+++ b/ml_datasets/util.py
@@ -37,15 +37,20 @@ def dl_progress(count, block_size, total_size):
         else:
             progbar.update(block_size)
 
-    error_msg = "URL fetch failure on {}: {} -- {}"
     if not os.path.exists(fpath):
         try:
             try:
                 urlretrieve(origin, fpath, dl_progress)
-            except URLError as e:
-                raise Exception(error_msg.format(origin, e.errno, e.reason))
+            # Enrich download exceptions with full file name
+            # HTTPError is a subclass of URLError, so it must be caught first
             except HTTPError as e:
-                raise Exception(error_msg.format(origin, e.code, e.msg))
+                error_msg = "URL fetch failure on {} : {} -- {}"
+                e.msg = error_msg.format(origin, e.code, e.msg)
+                raise
+            except URLError as e:
+                error_msg = "URL fetch failure on {} -- {}"
+                e.reason = error_msg.format(origin, e.reason)
+                raise
         except (Exception, KeyboardInterrupt):
             if os.path.exists(fpath):
                 os.remove(fpath)

From 0842b98ba49cf8c6c16bf963eb36713c1435a3ae Mon Sep 17 00:00:00 2001
From: crusaderky <crusaderky@gmail.com>
Date: Fri, 12 Dec 2025 16:51:55 +0000
Subject: [PATCH 2/2] fix regex on windows and mac

---
 ml_datasets/test/test_util.py | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/ml_datasets/test/test_util.py b/ml_datasets/test/test_util.py
index 3c96330..d37826b 100644
--- a/ml_datasets/test/test_util.py
+++ b/ml_datasets/test/test_util.py
@@ -4,7 +4,9 @@
 
 
 def test_get_file_domain_resolution_fails():
-    with pytest.raises(URLError, match="test_non_existent_file.*Name or service not known"):
+    with pytest.raises(
+        URLError, match=r"test_non_existent_file.*(not known|getaddrinfo failed)"
+    ):
         get_file(
             "non_existent_file.txt",
             "http://test_notexist.wth/test_non_existent_file.txt"