mirrors/github.com_openSUSE_osc
1
0
Fork 0
mirror of https://github.com/openSUSE/osc.git synced 2025-02-21 17:52:14 +01:00
Code

Merge pull request #1701 from dmach/fix-detecting-binaries-v2

Browse Source 
Fix typo in core.binary() that caused text files being detected as binary
This commit is contained in:
Daniel Mach 2025-02-13 14:48:31 +01:00 committed by GitHub
commit 47cd6b37c8
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
2 changed files with 15 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
osc/core.py
View File

@ -2796,7 +2796,9 @@ def binary(data: bytes):
We're using heuristics according to OBS: src/backend/BSSrcServer/filediff - look for "diff binary detection" We're using heuristics according to OBS: src/backend/BSSrcServer/filediff - look for "diff binary detection"
""" """
binary_chars = re.findall(b"[\x00-\0x07\x0e-\x1f]", data) if b"\0" in data:
return True
binary_chars = re.findall(b"[\x00-\x07\x0e-\x1f]", data)
return len(binary_chars) * 40 > len(data) return len(binary_chars) * 40 > len(data)

17
tests/test_core.py
View File

@ -159,12 +159,12 @@ class TestBinaryFile(unittest.TestCase):
path = os.path.join(self.tmpdir, "binary") path = os.path.join(self.tmpdir, "binary")
with open(path, "wb") as f: with open(path, "wb") as f:
f.write(1000 * b"a") f.write(1000 * b"a")
f.write(b"\0") f.write(b"\1")
self.assertFalse(binary_file(path)) self.assertFalse(binary_file(path))
with open(path, "wb") as f: with open(path, "wb") as f:
f.write(4096 * b"a") f.write(4096 * b"a")
f.write(b"\0") f.write(b"\1")
self.assertFalse(binary_file(path)) self.assertFalse(binary_file(path))
def test_binary(self): def test_binary(self):
@ -173,19 +173,26 @@ class TestBinaryFile(unittest.TestCase):
# sufficient control chars in first 4k # sufficient control chars in first 4k
with open(path, "wb") as f: with open(path, "wb") as f:
f.write(1000 * b"a") f.write(1000 * b"a")
f.write(26 * b"\0") f.write(26 * b"\1")
self.assertTrue(binary_file(path)) self.assertTrue(binary_file(path))
# sufficient control chars in first 4k # sufficient control chars in first 4k
with open(path, "wb") as f: with open(path, "wb") as f:
f.write(3993 * b"a") f.write(3993 * b"a")
f.write(103 * b"\0") f.write(103 * b"\1")
self.assertTrue(binary_file(path))
# a single \0 is good enough for us to say it's a binary file
with open(path, "wb") as f:
f.write(3993 * b"a")
f.write(b"\0")
f.write(999 * b"\1")
self.assertTrue(binary_file(path)) self.assertTrue(binary_file(path))
# detected as text because we're reading only first 4k characters # detected as text because we're reading only first 4k characters
with open(path, "wb") as f: with open(path, "wb") as f:
f.write(4096 * b"a") f.write(4096 * b"a")
f.write(1000 * b"\0") f.write(1000 * b"\1")
self.assertFalse(binary_file(path)) self.assertFalse(binary_file(path))