1 |
2008/5/30 Hemmann, Volker Armin <volker.armin.hemmann@××××××××××××.de>: |
2 |
|
3 |
> On Freitag, 30. Mai 2008, Duncan wrote: |
4 |
> |
5 |
> > I've not done paludis due to its lack of binary package support. Even |
6 |
> > tho I'm running only a single computer |
7 |
> |
8 |
> there is another reason not to use paludis: |
9 |
> |
10 |
> you can't go back. |
11 |
> |
12 |
> At least not easily. |
13 |
> |
14 |
> With pkgcore you can switch between pkgcore and portage 'on the fly'. |
15 |
> emerge |
16 |
> app a, pmerge app b, emerge app c. |
17 |
> |
18 |
> The config files are not touched. |
19 |
|
20 |
|
21 |
with paludis the portage config files are not touched. paludis has its own |
22 |
config dir where to install its personal config files. i use portage |
23 |
regulary to compile with ABI=x86 which paludis has some problems with and |
24 |
sometimes goes a little mad. |
25 |
|
26 |
> |
27 |
> Paludis on the other hand can only described with 'vendor lock in' |
28 |
> and 'gratuitous incompatibilty'. And don't forget that it is slow. |
29 |
> |
30 |
|
31 |
on what base you say it's slow?! i'm using it because on my system is tons |
32 |
of times faster; i really assure you that is faster in discovering deps and |
33 |
it finds them in the right way at least. i'm now rebuilding a new clean |
34 |
system and with portage i've got problems building it since portage quite |
35 |
some time has not selected the right dependency and thus builds fail. with |
36 |
paludis for now this hasn't yet happened. to not speak of the continue |
37 |
option: with portage after a package compile/install fail portage stops and |
38 |
doesn't continue. after you do a --resume skipfirst most of the times the |
39 |
build will completely fail because of deps not met and then there would be |
40 |
no way to resume the actual build. paludis has the continue-on-errors option |
41 |
with cases like if package is independent of the one failed (useful in new |
42 |
sytem builds), if requirements not met (useful in world updates) and also it |
43 |
gives a list of packages. you could edit the list and set S instead of P |
44 |
whenever you'd like to skip a specific package. |
45 |
|
46 |
|
47 |
That it also requires a shitload of dependencies and installs more crap than |
48 |
> portage and pkgcore combined doesn't make it better. |
49 |
> |
50 |
|
51 |
on this i can agree with you. i'd just want to show one thig, though: on the |
52 |
new system i've installed the weekly stages from funtoo, rebuild the system |
53 |
with emerge -e system, rebuild again the system with emerge -e system and my |
54 |
personal use flags and then build xorg-server with portage. after that, |
55 |
since i wanted kde4-svn i installed paludis and, miracle, the only real dep |
56 |
needed for me there was boost. it seems that now the deps aren't so huge |
57 |
anymore. anyway that it takes a big deal of time every update is right (2,5 |
58 |
hours with inram compilation). |
59 |
|
60 |
At a last point: don't forget WHO is behind paludis - some of the most |
61 |
> abusive |
62 |
> persons gentoo has ever seen. The same people responsible for most |
63 |
> problems. |
64 |
> |
65 |
> Abusive, agressive, searching for stuff that is not covered by rules, |
66 |
> behave |
67 |
> like a rabid ape until everything is covered by rules, suffocating gentoo |
68 |
> and |
69 |
> then turn into rule nazis and game the system. Yes, this people are behind |
70 |
> paludis - and 'exherbo'. |
71 |
|
72 |
|
73 |
a software is not to be judged by its creators, but on what it does. as i've |
74 |
said paludis has some flaws (is written in c++, it takes ages to update, it |
75 |
doesn't behave very well with x86 abi on multilib profiles, lacks binpkg |
76 |
support) but also has some better issues (a big deal of hooks that improve |
77 |
it very much, it can handle very external overlays and update them without |
78 |
passing through another package and this is i think the best thing of |
79 |
paludis, is can handle per package or per overlay, keyword and unmask |
80 |
configs, it can use eix through the eix hook, it can show new packages in |
81 |
the tree after update and show a report of your system, is fast on first run |
82 |
world update - on 400+ package world the time of resolving deps and |
83 |
everything is less than 30 secs, while portage on first run goes for some |
84 |
minutes, it has reconcilio that is faster and cleaner than revedep-rebuild |
85 |
and it handles the new scm build system which is really nice. the problem |
86 |
with that build system is its test. portage won't adopt it for the moment |
87 |
because it still has to be tested but that a package can suggestother |
88 |
packages and the user would be able to accept suggestions or not is a big |
89 |
step towards usability. |
90 |
|
91 |
-- |
92 |
dott. ing. beso |