1 |
On 26-11-2018 21:13:53 +0000, Andrey Utkin wrote: |
2 |
> On Wed, Nov 21, 2018 at 11:45:54AM +0100, Fabian Groffen wrote: |
3 |
> > We agree it is hackish, and we agree we can do without. You simply |
4 |
> > exaggerate the problem, IMO, which mostly isn't there, because it works |
5 |
> > fine today. It can also be solved today using shell tools. |
6 |
> |
7 |
> I am sad that you don't see it as a productivity impediment that the |
8 |
> user is required to know the custom tooling to do even such a trivial |
9 |
> non-standard action as manual extraction. |
10 |
|
11 |
Huh? tar -jxf doesn't do the trick for you? |
12 |
|
13 |
> Maybe I will make myself look bad by admitting this, but I'm not meeting |
14 |
> your expectations. I use Gentoo for ~11 years, and for about one year I |
15 |
> am using my private binpkgs distributed to all my machines (i.e. I have |
16 |
> read binary package guide fair number of times, but I stopped rereading |
17 |
> it when I satisfied my needs). When in need, I still reached to trusty |
18 |
> tar, and I did not even know what are the names of special tools (a |
19 |
> toolchain?) qtbz2 and qxpak. |
20 |
> |
21 |
> Just few days ago I messed with binpkgs for investigation purpose. I |
22 |
> just wanted to extract few to somewhere (definitely not into system |
23 |
> root), and read a core dump with GDB asking it to use those extracted |
24 |
> files for debug symbols. |
25 |
> |
26 |
> Of course I used `tar xaf`, because what I know is that it's honest tbz2 |
27 |
> just with metadata appended. |
28 |
> |
29 |
> # tar xaf boost-1.65.0.tbz2 |
30 |
> |
31 |
> bzip2: (stdin): trailing garbage after EOF ignored |
32 |
> |
33 |
> Exit code is 0. |
34 |
> But the notice is annoying (on subconscious level), because Silence Is |
35 |
> Golden - "when a program has nothing interesting or surprising to say, |
36 |
> it should shut up". |
37 |
|
38 |
You seem to contradict yourself. You didn't know the tools, yet you say |
39 |
you needed to, to unpack the files. But you show here you just unpacked |
40 |
the files without said knowledge. |
41 |
|
42 |
> > % head -c `grep -abo 'XPAKPACK' $EPREFIX/usr/portage/packages/sys-apps/sed-4.5.tbz2 | sed 's/:.*$//'` $EPREFIX/usr/portage/packages/sys-apps/sed-4.5.tbz2 | tar -jxf - |
43 |
> > |
44 |
> > results in no warnings/errors from bzip about trailing garbage, possible |
45 |
> > thanks to the spec being smart enough about this. |
46 |
> |
47 |
> Thanks, this is a very concise **custom tool** to handle current binpkg |
48 |
> format. |
49 |
|
50 |
As is tar followed by tar. The obvious advantage of the latter is that |
51 |
you don't get a warning which could trigger you into thinking something |
52 |
is wrong. So, in my opinion, that is a better way of doing it compared |
53 |
to the current way. |
54 |
|
55 |
> > Not having to do this, when under stress and pressure to restore a |
56 |
> > system to get it back into production, is a plus. Though, in that |
57 |
> > scenario the trailing garbage warning wouldn't have been that bad |
58 |
> > either. |
59 |
> |
60 |
> When understress and pressure, the irrelevant warning is not bad? |
61 |
> I am sure it is really bad for operator's attention. |
62 |
|
63 |
I've been using Gentoo binpkgs for a long while, I think something like |
64 |
~14 years ago when I used them extensively. Perhaps I'm an exception, |
65 |
but back then I knew already there was an extra bit attached to the |
66 |
tars, as were all my collegues around me back then. The fact it comes |
67 |
up now (as a surprise?) maybe means the knowledge has gone. So good |
68 |
thing we're replacing it with something easier to infer from inspecting |
69 |
it. |
70 |
|
71 |
Fabian |
72 |
|
73 |
|
74 |
-- |
75 |
Fabian Groffen |
76 |
Gentoo on a different level |