Finished tar

tar.py

@@ -1,3 +1,4 @@
+import builtins
 from flask import Response
 import functools
 import os
@@ -5,6 +6,12 @@ import os
 # 4MiB chunks
 CHUNK_SIZE = 4_194_304
 
+# ASCII value for space
+SPACE = ord(' ')
+
+# ASCII value for zero
+ZERO = ord('0')
+
 # Specification de l'entête du fichier
 # Numéro Nom      Début   Taille   Description
 #      1 name         0      100   Nom du fichier
@@ -19,45 +26,47 @@ CHUNK_SIZE = 4_194_304
 
 
 def header_chunk(filename: str, filepath: str) -> bytes:
+
+    # Returns the octal representation without the initial
+    def oct(i: int) -> str:
+        return builtins.oct(i)[2:]
+
     stat = os.stat(filepath)
-    buffer = bytearray()
+    buffer = bytearray(512)
 
     # Field 1: filename on 100 bytes
-    buffer += filename.encode('ascii')
+    buffer[0:len(filename)] = filename.encode('ascii')
-    buffer += b'\x00' * (100 - len(filename))
 
-    # Field 2: mode, on 8 bytes
+    # Field 2: mode, on 8 bytes, octal, last byte must be \x00, so we set only the first 7 bytes
-    # TODO we put 777 for test
+    buffer[100:107] = oct(stat.st_mode).rjust(7, '0').encode('ascii')
-    buffer += '0000777'.encode('ascii') + b'\x00'
 
-    # Field 3: owner, we put 1000, default user
+    # Field 3: owner, on 8 bytes, octal, last byte must be \x00, so we set only the first 7 bytes
-    buffer += '0001000'.encode('ascii') + b'\x00'
+    buffer[108:115] = oct(stat.st_uid).rjust(7, '0').encode('ascii')
 
-    # Field 4: group, we put 1000
+    # Field 4: group, on 8 bytes, octal, last byte must be \x00, so we set only the first 7 bytes
-    buffer += '0001000'.encode('ascii') + b'\x00'
+    buffer[116:123] = oct(stat.st_gid).rjust(7, '0').encode('ascii')
 
-    # Field 5: file size in bytes in ascii
+    # Field 5: file size in bytes, on 12 bytes, octal, last byte must be \x00, so we set only the first 11 bytes
-    buffer += oct(stat.st_size)[2:].rjust(11, '0').encode('ascii') + b'\x00'
+    buffer[124:135] = oct(stat.st_size).rjust(11, '0').encode('ascii')
 
-    # Field 6: last modified, zeros for now
+    # Field 6: last modified, on 12 bytes, octal, last byte must be \x00, so we set only the first 11 bytes
-    # buffer += '00000000000'.encode('ascii') + b'\x00'
+    buffer[136:147] = oct(int(stat.st_mtime)).rjust(11, '0').encode('ascii')
-    buffer += oct(1721728914)[2:].rjust(11, '0').encode('ascii') + b'\x00'
 
-    # Field 7: checksum, we put spaces and we will edit it at the end
+    # Field 7: checksum, we fill it at the end
-    buffer += '        '.encode('ascii')
 
     # Field 8: type flag, 0 because we only have regular files
-    buffer += b'0'
+    buffer[156] = ZERO
 
     # Field 9: linkname, \x00s because we only have regular files
-    buffer += b'\x00' * 100
 
     # POSIX 1003.1-1990: 255 empty bytes
-    buffer += b'\x00' * 255
 
-    # Compute the checksum
+    # Compute the checksum: we start at 248 which are the 8 fields of checksum filled with spaces (32 * 8)
-    checksum = oct(functools.reduce(lambda x, y: x + y, buffer))[2:].rjust(6, '0').encode('ascii') + b'\x00 '
+    checksum = oct(functools.reduce(lambda x, y: x + y, buffer, 256)).rjust(6, '0').encode('ascii')
-    buffer[148:156] = checksum
+    buffer[148:154] = checksum
+
+    # Don't ask me why, but the checksum must end with b'\x00 ', so we skip the \x00 and write the space
+    buffer[155] = SPACE
 
     return bytes(buffer)