From 41397cee70f4933b539ab1a9232f04419ac8e2bb Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Johan=20F=C3=B6rberg?= <johan@forberg.se>
Date: Tue, 2 Jul 2024 15:35:20 +0200
Subject: [PATCH] gh-121267: Improve performance of tarfile (#121267)

Tarfile in the default write mode spends much of its time resolving UIDs
into usernames and GIDs into group names. By caching these mappings, a
significant speedup can be achieved.

In my simple benchmark[1], this extra caching speeds up tarfile by 8x.

[1] https://gist.github.com/jforberg/86af759c796199740c31547ae828aef2
---
 Lib/tarfile.py                                | 28 +++++++++++++------
 ...-07-02-15-56-42.gh-issue-121267.yFBWkh.rst |  2 ++
 2 files changed, 22 insertions(+), 8 deletions(-)
 create mode 100644 Misc/NEWS.d/next/Library/2024-07-02-15-56-42.gh-issue-121267.yFBWkh.rst

diff --git a/Lib/tarfile.py b/Lib/tarfile.py
index d5d8a469779f50b..3e16dcd519b1536 100644
--- a/Lib/tarfile.py
+++ b/Lib/tarfile.py
@@ -1658,6 +1658,9 @@ class TarFile(object):
 
     extraction_filter = None    # The default filter for extraction.
 
+    uname_cache = {}            # Cached mappings of uid -> uname, gid -> gname
+    gname_cache = {}
+
     def __init__(self, name=None, mode="r", fileobj=None, format=None,
             tarinfo=None, dereference=None, ignore_zeros=None, encoding=None,
             errors="surrogateescape", pax_headers=None, debug=None,
@@ -2105,16 +2108,25 @@ def gettarinfo(self, name=None, arcname=None, fileobj=None):
         tarinfo.mtime = statres.st_mtime
         tarinfo.type = type
         tarinfo.linkname = linkname
+
+        # Calls to pwd.getpwuid() and grp.getgrgid() tend to be expensive. To
+        # speed things up, cache the resolved usernames and group names.
         if pwd:
-            try:
-                tarinfo.uname = pwd.getpwuid(tarinfo.uid)[0]
-            except KeyError:
-                pass
+            if not tarinfo.uid in self.uname_cache:
+                try:
+                    self.uname_cache[tarinfo.uid] = pwd.getpwuid(tarinfo.uid)[0]
+                except KeyError:
+                    pass
+
+            tarinfo.uname = self.uname_cache.get(tarinfo.uid, None)
         if grp:
-            try:
-                tarinfo.gname = grp.getgrgid(tarinfo.gid)[0]
-            except KeyError:
-                pass
+            if not tarinfo.gid in self.gname_cache:
+                try:
+                    self.gname_cache[tarinfo.gid] = grp.getgrgid(tarinfo.gid)[0]
+                except KeyError:
+                    pass
+
+            tarinfo.gname = self.gname_cache.get(tarinfo.gid, None)
 
         if type in (CHRTYPE, BLKTYPE):
             if hasattr(os, "major") and hasattr(os, "minor"):
diff --git a/Misc/NEWS.d/next/Library/2024-07-02-15-56-42.gh-issue-121267.yFBWkh.rst b/Misc/NEWS.d/next/Library/2024-07-02-15-56-42.gh-issue-121267.yFBWkh.rst
new file mode 100644
index 000000000000000..ca18bf37471bad3
--- /dev/null
+++ b/Misc/NEWS.d/next/Library/2024-07-02-15-56-42.gh-issue-121267.yFBWkh.rst
@@ -0,0 +1,2 @@
+Improve the performance of tarfile when writing files, by caching user names
+and group names.