Gentoo Archives: gentoo-dev

From: Alexander Tsoy <alexander@××××.me>
To: gentoo-dev@l.g.o
Subject: Re: [gentoo-dev] RFC: using .xz for doc/man/info compression
Date: Mon, 12 May 2014 10:47:46
Message-Id: 20140512144736.491a9668@work.puleglot
In Reply to: Re: [gentoo-dev] RFC: using .xz for doc/man/info compression by Gordon Pettey
1 ÷ Sun, 11 May 2014 18:26:32 -0500
2 Gordon Pettey <petteyg359@×××××.com> ÐÉÛÅÔ:
3
4 > A lot of small files (e.g. AUTHORS, ChangeLog
5 >
6 > FWIW: On my system, I have 59M of bz2 files in /usr/share/man and
7 > /usr/share/doc. A short script to decompress those and recompress with xz
8 > -6e reduced that to 36M.
9
10 Very strange o_O
11
12 Here is my test results. xz options: "--lzma2=preset=6e,dict=4MiB".
13 Larger dictionary size does not improve compression ratio, I get
14 even worse results with just "-6e" or "-9e". man-bz2 is a full copy of
15 my /usr/share/man, man-xz is a recompressed one.
16
17 Size comparison:
18
19 $ du -s man-bz2/ man-xz/
20 82032 man-bz2/
21 82308 man-xz/
22
23
24 Decompression speed:
25
26 $ time find man-bz2/ -type f -name "*.bz2" -exec bzcat '{}' > /dev/null \;
27
28 real 0m35.110s
29 user 0m14.509s
30 sys 0m15.227s
31 $ time find man-bz2/ -type f -name "*.bz2" -exec bzcat '{}' > /dev/null \;
32
33 real 0m35.407s
34 user 0m14.432s
35 sys 0m15.186s
36 $ time find man-xz/ -type f -name "*.xz" -exec xzcat '{}' > /dev/null \;
37
38 real 0m46.571s
39 user 0m17.077s
40 sys 0m23.906s
41 $ time find man-xz/ -type f -name "*.xz" -exec xzcat '{}' > /dev/null \;
42
43 real 0m46.137s
44 user 0m17.276s
45 sys 0m23.426s
46
47
48 As you can see, xz is actually worse in speed and compression ratio.
49
50 --
51 Alexander Tsoy

Replies

Subject Author
Re: [gentoo-dev] RFC: using .xz for doc/man/info compression Alexander Tsoy <alexander@××××.me>
Re: [gentoo-dev] RFC: using .xz for doc/man/info compression Tom Wijsman <TomWij@g.o>
Re: [gentoo-dev] RFC: using .xz for doc/man/info compression Gordon Pettey <petteyg359@×××××.com>
Re: [gentoo-dev] RFC: using .xz for doc/man/info compression Andrew Savchenko <bircoph@×××××.com>