cuneiform team mailing list archive

[Robert Wohlrab <robert.wohlrab@xxxxxx>]

On Friday 10 April 2009 14:02:01 Aleksandr Yakimov wrote:
> I would suggest to split source code and datafiles (rarely modified
> and we have no way to generate new). Check gz, bz2 and lzma for source
> code only and dictionaries (datafiles).
This is a good idea. I moved the datafiles folder from the cuneiform-0.6.0 
release to cuneiform-data-0.6.0 and recreated the test (removed paq8l, lzo, 
(l)rzip and did the test only in extreme mode).

20858533 cuneiform-data-0.6.0.tar.lzma
20861808 cuneiform-data-0.6.0.tar.xz
24253283 cuneiform-data-0.6.0.tar.bz2
26442122 cuneiform-data-0.6.0.tar.gz
51988480 cuneiform-data-0.6.0.tar

3450462 cuneiform-src-0.6.0.tar.lzma
3451048 cuneiform-src-0.6.0.tar.xz
4985556 cuneiform-src-0.6.0.tar.bz2
5684012 cuneiform-src-0.6.0.tar.gz
28006400 cuneiform-src-0.6.0.tar

It seems to be the same result as before, but distributors should be able now 
to create smaller packages for upgrading a cuneiform installation (most of the 
time no cuneiform-data update must be installed).

cuneiform-data-0.6.0.tar.lzma
2.42user 0.10system 0:02.71elapsed 92%CPU
cuneiform-data-0.6.0.tar.xz
2.19user 0.12system 0:02.43elapsed 95%CPU
cuneiform-data-0.6.0.tar.bz2
5.39user 0.04system 0:05.70elapsed 95%CPU
cuneiform-data-0.6.0.tar.gz
0.71user 0.00system 0:00.76elapsed 93%CPU

cuneiform-src-0.6.0.tar.lzma
0.39user 0.02system 0:00.45elapsed 92%CPU
cuneiform-src-0.6.0.tar.xz
0.44user 0.03system 0:00.50elapsed 94%CPU
cuneiform-src-0.6.0.tar.bz2
1.54user 0.04system 0:01.68elapsed 95%CPU
cuneiform-src-0.6.0.tar.gz
0.27user 0.00system 0:00.30elapsed 93%CPU
-- 
Robert Wohlrab