Mercurial > hg > cc > cirrus_work

comparison lib/python/unpackz.py @ 239:992f59d21832

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
working, but last count/offset not being written
author Henry S. Thompson <ht@inf.ed.ac.uk>
date Sat, 28 Sep 2024 15:19:05 +0100
parents
children 51bd09d4289e
comparison
equal deleted inserted replaced
238:a4538bdfa92a 239:992f59d21832
1 #!/usr/bin/env python3
2 '''See https://stackoverflow.com/a/37042747/2595465'''
3 import sys
4 import isal.isal_zlib
5 offset = 0
6 obuf_len = 0
7 nbuf = lastbuf = False
8 BUFSIZE = 1048576
9 outfile = None
10 if sys.argv[1] == '-o':
11 sys.argv.pop(1)
12 if len(sys.argv)==3:
13 outfile = open(sys.argv.pop(1),'wb')
14 else:
15 print('need an outfile', file=sys.stderr)
16 exit(1)
17 with open(sys.argv[1],'rb') as f:
18 z = isal.isal_zlib.decompressobj(31)
19 count = 0
20 while True:
21 if z.unused_data == b"":
22 #print('n', obuf_len, file=sys.stderr)
23 if nbuf:
24 obuf_len += BUFSIZE # still no EOS after a full buffer processed
25 buf = f.read(BUFSIZE)
26 nbuf = True
27 lastbuf = ((truesize:=len(buf)) < BUFSIZE) # will only succeed if now at EOF
28 if buf == b"":
29 if outfile is None:
30 print(count, offset)
31 else:
32 outfile.write(b'\000%d\000%d\000'%(count, offset))
33 break
34 else:
35 buf_len = len(buf)
36 #print('b', obuf_len, buf_len, len(z.unused_data), len(buf)-len(z.unused_data),
37 # nbuf, lastbuf, file=sys.stderr)
38 count = (obuf_len if (buf_len == truesize) else 0) + \
39 (len(buf)-len(z.unused_data))
40 if outfile is None:
41 print(count, offset)
42 else:
43 outfile.write(b'\000%d\000%d\000'%(count, offset))
44 offset += count
45 count = 0
46 buf = z.unused_data
47 obuf_len = len(buf)
48 nbuf = False
49 z = isal.isal_zlib.decompressobj(31)
50 got = z.decompress(buf)
51 if outfile is not None:
52 outfile.write(got)
53 if count!=0:
54 print("Unused data: count=%s offset=%s ?"%(count, offset),
55 file=sys.stderr)
56 if outfile is not None:
57 outfile.close()