1 |
Hi |
2 |
|
3 |
Francesco Riosa wrote: |
4 |
> <quote> |
5 |
> zsync is a file transfer program. It allows you to download a file from |
6 |
> a remote web server, where you have a copy of an older version of the |
7 |
> file on your computer already. zsync downloads only the new parts of the |
8 |
> file. It uses the same algorithm as rsync <http://rsync.samba.org/>. |
9 |
> |
10 |
> zsync does not require any special server software or a shell account on |
11 |
> the remote system (rsync, in comparison, requires that you have an rsh |
12 |
> or ssh account, or that the remote system runs rsyncd). Instead, it uses |
13 |
> a control file — a |.zsync| file — that describes the file to be |
14 |
> downloaded and enables zsync to work out which blocks it needs. This |
15 |
> file can be created by the admin of the web server hosting the download, |
16 |
> and placed alongside the file to download — it is generated once, then |
17 |
> any downloaders with zsync can use it. Alternatively, anyone can |
18 |
> download the file, make a .zsync and provide it to other users (this is |
19 |
> what I am doing for the moment). |
20 |
> |
21 |
> zsync is currently no more than an alpha. I have tried to make it quite |
22 |
> verbose, so it is clear what it is doing, and the checksum verification |
23 |
> and file handling are designed to minimise the risk of it losing any |
24 |
> data. It works well enough for me. |
25 |
> </quote> |
26 |
|
27 |
And where is the benefit beside mirrors don't need to have a running |
28 |
rsync? If it uses exactly the same algorithm for finding the |
29 |
differences, users won't download less. |
30 |
|
31 |
Or did I get it completely wrong? |
32 |
|
33 |
Greetings, |
34 |
|
35 |
blubb |
36 |
|
37 |
-- |
38 |
gentoo-dev@g.o mailing list |