1 |
Mike Frysinger ha scritto: |
2 |
|
3 |
>On Thursday 24 March 2005 09:16 am, Francesco Riosa wrote: |
4 |
> |
5 |
> |
6 |
>>Chris Gianelloni ha scritto: |
7 |
>> |
8 |
>> |
9 |
>>>On Thu, 2005-03-24 at 14:22 +0100, Francesco Riosa wrote: |
10 |
>>> |
11 |
>>> |
12 |
>>>>In addition to "Manifest" files in portage tree there are 19000 |
13 |
>>>>"digest-${P}" files in the portage tree each of this uses 4k on ext{2,3} |
14 |
>>>>filesystem and has to be scanned each time rsync is run on the tree. |
15 |
>>>>Is possible to avoid the use of theese files? |
16 |
>>>> |
17 |
>>>> |
18 |
>>>Not if you ever want to ensure that what you're downloading hasn't been |
19 |
>>>trojaned or otherwise tampered with. |
20 |
>>> |
21 |
>>> |
22 |
>>ok, this is true only when you have FEATURES="strict" or "cvs" that it's |
23 |
>>not the default. |
24 |
>> |
25 |
>> |
26 |
> |
27 |
>no it's not |
28 |
> |
29 |
>Manifest is checked when you have FEATURES=strict, digest is checked everytime |
30 |
>you fetch/unpack files ... all those little 'src_uri md5' lines is a check |
31 |
>against the digest |
32 |
>-mike |
33 |
> |
34 |
> |
35 |
ah ok, I was confused by the fact that I can change patches and ebuild |
36 |
in my *overlays* without the need of a "ebuild P digest" |
37 |
already found "digestCheckFiles" & "digestcheck" in portage.py so |
38 |
pointers not needed anymore. |
39 |
|
40 |
thanks |
41 |
-- |
42 |
gentoo-dev@g.o mailing list |