From 7e058072049ba96f74160b3c0e8ad16610de2af7 Mon Sep 17 00:00:00 2001
From: Martin Durant <martindurant@users.noreply.github.com>
Date: Fri, 22 Sep 2023 11:41:21 -0400
Subject: [PATCH] Allow size= in open() (#797)

---
 s3fs/core.py            |  6 +++++-
 s3fs/tests/test_s3fs.py | 12 ++++++++++++
 2 files changed, 17 insertions(+), 1 deletion(-)

diff --git a/s3fs/core.py b/s3fs/core.py
index 3d1a61af..54c75201 100644
--- a/s3fs/core.py
+++ b/s3fs/core.py
@@ -604,6 +604,7 @@ def _open(
         fill_cache=None,
         cache_type=None,
         autocommit=True,
+        size=None,
         requester_pays=None,
         cache_options=None,
         **kwargs,
@@ -680,6 +681,7 @@ def _open(
             autocommit=autocommit,
             requester_pays=requester_pays,
             cache_options=cache_options,
+            size=size,
         )
 
     async def _lsdir(
@@ -2056,6 +2058,7 @@ def __init__(
         cache_type="readahead",
         requester_pays=False,
         cache_options=None,
+        size=None,
     ):
         bucket, key, path_version_id = s3.split_path(path)
         if not key:
@@ -2094,6 +2097,7 @@ def __init__(
             autocommit=autocommit,
             cache_type=cache_type,
             cache_options=cache_options,
+            size=size,
         )
         self.s3 = self.fs  # compatibility
 
@@ -2133,7 +2137,7 @@ def __init__(
             # Reflect head
             self.s3_additional_kwargs.update(head)
 
-        if "r" in mode and "ETag" in self.details:
+        if "r" in mode and size is None and "ETag" in self.details:
             self.req_kw["IfMatch"] = self.details["ETag"]
 
     def _call_s3(self, method, *kwarglist, **kwargs):
diff --git a/s3fs/tests/test_s3fs.py b/s3fs/tests/test_s3fs.py
index 7fb400cb..8675589b 100644
--- a/s3fs/tests/test_s3fs.py
+++ b/s3fs/tests/test_s3fs.py
@@ -166,6 +166,18 @@ def test_simple(s3):
         assert out == data
 
 
+def test_with_size(s3):
+    data = b"a" * (10 * 2**20)
+
+    with s3.open(a, "wb") as f:
+        f.write(data)
+
+    with s3.open(a, "rb", size=100) as f:
+        assert f.size == 100
+        out = f.read()
+        assert len(out) == 100
+
+
 @pytest.mark.parametrize("default_cache_type", ["none", "bytes", "mmap", "readahead"])
 def test_default_cache_type(s3, default_cache_type):
     data = b"a" * (10 * 2**20)