1 |
On Tue, 16 May 2006 09:16:18 -0700 Brian Harring <ferringb@×××××.com> |
2 |
wrote: |
3 |
| 1) changes to the eapi=0 ebuild standard; renaming of vars |
4 |
| (PORTAGE_* -> PALUDIS_* namely) |
5 |
|
6 |
What eapi=0 standard? We emulate Portage internals where it's found to |
7 |
be necessary, and don't otherwise. |
8 |
|
9 |
| dropping of all local vars during phase changes |
10 |
|
11 |
Again, we emulate Portage misfeatures only where it's found to be |
12 |
necessary. |
13 |
|
14 |
| effective dropping of config phase |
15 |
|
16 |
Uh, no. Config's on the 0.4 list. We're just prioritising things that |
17 |
we actually need. |
18 |
|
19 |
| Mind you that's from a quick read through |
20 |
| of your ebuild env reimplementation, stated the issues already and |
21 |
| they still remain- incomplete support for the standard is one thing, |
22 |
| changing it is another (y'all are doing the latter). |
23 |
|
24 |
What standard? Are you trying to say that Paludis should emulate |
25 |
Portage's bugs, because the standard is "exactly what Portage does"? |
26 |
|
27 |
| 2) N profile inheritance- the parents change (N entries instead of |
28 |
| one) needs to be protected so that specific profile is only usable by |
29 |
| a package manager (whether portage, pkgcore or paludis) that actually |
30 |
| does N parent inheritance. This is why N profile inheritance has |
31 |
| never been added to portage (it's honestly a one line change in |
32 |
| portage)- the change must not be introduced without protection, |
33 |
| else the user's system set can become drastically reduced. It's not |
34 |
| an easily addressable problem (all solutions sans a new profile |
35 |
| directory leave open the potential for users to get bit), ignoring it |
36 |
| is a no go. |
37 |
|
38 |
Uh, no. Any user who isn't using a package manager capable of multiple |
39 |
inherits shouldn't use a multiple inherits profile. There's plenty of |
40 |
precedent of intro |
41 |
|
42 |
| Yes, you're after demoing capabilities- problem is that |
43 |
| it's exposing potential horkage in a production tree, wrong place to |
44 |
| be demoing. |
45 |
|
46 |
Heh. "potential horkage". Yes, if a user sets their profile |
47 |
incorrectly, things break. That's the case currently too. |
48 |
|
49 |
| 3) vdb CONTENTS change of dev/fif to misc. It's dumb, but that |
50 |
| change renders vdb entries incompatible- iow, proper usage/conversion |
51 |
| to paludis requires a new installation (or translation script, both |
52 |
| to/from portage). |
53 |
|
54 |
Had you bothered to read the documentation, you'd know that we don't |
55 |
claim nor desire VDB compatibility with Portage, and don't encourage |
56 |
installs onto the same ROOT. |
57 |
|
58 |
We also don't emulate Portage's broken handling of merging directories |
59 |
onto symlinks, which means that uninstalling via emerge sometimes |
60 |
leaves behind cruft. |
61 |
|
62 |
| So... short version, introduction of the profile allows for curious |
63 |
| users to get bit in the ass by intentional dropping of compatibility |
64 |
| (profile level changes are one thing, changing the ebuild standard is |
65 |
| another). In light of that, why should it be demoed in the tree |
66 |
| where the only use of it is to bootstrap a new installation? Just |
67 |
| overlay it, y'all should be maintaining an overlay fixing ebuild |
68 |
| incompatibilities anyways. |
69 |
|
70 |
Because Paludis is now a viable alternative to Portage and can be used |
71 |
to install and maintain a real system. We already support enough of the |
72 |
"ebuild standard" and emulate enough of Portage's bugs to install |
73 |
system + X + KDE + Gnome + a whole load of random stuff that people |
74 |
actually use. |
75 |
|
76 |
| > That's my proposal. The benefits I like to think are obvious. The |
77 |
| > drawbacks are, as far as I can see, in tree size, which should be |
78 |
| > minimal. |
79 |
| |
80 |
| Benefits of demo'ing for a minority (300 devs) must be balanced |
81 |
| against risk (adding profiles that portage can eat itself on, the |
82 |
| default virtual change doesn't address it |
83 |
|
84 |
Uh, a user changing to any incorrect profile will screw up their |
85 |
system. Ever tried using an amd64 profile on x86? |
86 |
|
87 |
| eapi incompatibility |
88 |
|
89 |
You mean "not emulating some of Portage's sillier bugs that very few |
90 |
things rely upon anyway", right? |
91 |
|
92 |
| vdb changes )- wrong place to be deploying incompatibilites that paludis |
93 |
| introduces is into the production tree without appropriate |
94 |
| containment/protection. |
95 |
|
96 |
Uh, VDB isn't part of the tree. Nor does it introduce any breakages for |
97 |
existing users, except those who do something especially dumb. Even if |
98 |
a user *does* change their profile to a Paludis profile, it won't cause |
99 |
Portage to explode in such a way that switching profiles back won't fix |
100 |
it. |
101 |
|
102 |
| The gain of the profile is that you can do a few new tricks for folks |
103 |
| doing boostrapping experiments- why not just introduce an ebuild that |
104 |
| sets up the new profile in a temp overlay? |
105 |
|
106 |
No, the gain of the profile is that it prepares the way for people to |
107 |
move onto a package manager that doesn't suck. |
108 |
|
109 |
Now, if you were to object to, say, -scm versions in the main tree, |
110 |
whose very existence causes Portage to crap itself messily and die, you |
111 |
might have a point. But adding a profile only screws things up for |
112 |
users who manually switch to it, and it only screws them up temporarily |
113 |
and fairly cleanly. |
114 |
|
115 |
-- |
116 |
Ciaran McCreesh |
117 |
Mail : ciaran dot mccreesh at blueyonder.co.uk |
118 |
|
119 |
|
120 |
-- |
121 |
gentoo-dev@g.o mailing list |