1 |
W dniu śro, 17.01.2018 o godzinie 12∶50 -0800, użytkownik Zac Medico |
2 |
napisał: |
3 |
> On 01/17/2018 07:42 AM, Alec Warner wrote: |
4 |
> > On Wed, Jan 17, 2018 at 10:25 AM, Michał Górny <mgorny@g.o |
5 |
> > <mailto:mgorny@g.o>> wrote: |
6 |
> > |
7 |
> > W dniu wto, 16.01.2018 o godzinie 11∶32 -0800, użytkownik Zac Medico |
8 |
> > napisał: |
9 |
> > > On 01/16/2018 10:39 AM, Michał Górny wrote: |
10 |
> > > > W dniu wto, 16.01.2018 o godzinie 12∶44 -0500, użytkownik Alec |
11 |
> > Warner |
12 |
> > > > napisał: |
13 |
> > > > > On Tue, Jan 16, 2018 at 11:43 AM, Michał Górny |
14 |
> > <mgorny@g.o <mailto:mgorny@g.o>> wrote: |
15 |
> > > > > |
16 |
> > > > > > Include a repo.postsync.d hook to verify the rsync checkout |
17 |
> > using |
18 |
> > > > > > gemato. Given that not all people will want to have it enabled |
19 |
> > > > > > unconditionally, no setup.py rules are included -- instead, |
20 |
> > the file |
21 |
> > > > > > would be installed conditionally by the ebuild. |
22 |
> > > > > > |
23 |
> > > > > > [v2: included link to the wiki page] |
24 |
> > > > > > --- |
25 |
> > > > > > MANIFEST.in | 2 +- |
26 |
> > > > > > misc/repo.postsync.d/00gemato | 18 ++++++++++++++++++ |
27 |
> > > > > > 2 files changed, 19 insertions(+), 1 deletion(-) |
28 |
> > > > > > create mode 100644 misc/repo.postsync.d/00gemato |
29 |
> > > > > > |
30 |
> > > > > > diff --git a/MANIFEST.in b/MANIFEST.in |
31 |
> > > > > > index 4f6cac162..edc6704e7 100644 |
32 |
> > > > > > --- a/MANIFEST.in |
33 |
> > > > > > +++ b/MANIFEST.in |
34 |
> > > > > > @@ -14,4 +14,4 @@ include cnf/make.conf.example.* |
35 |
> > > > > > include .portage_not_installed |
36 |
> > > > > > |
37 |
> > > > > > # extra scripts |
38 |
> > > > > > -include misc/* |
39 |
> > > > > > +graft misc |
40 |
> > > > > > diff --git a/misc/repo.postsync.d/00gemato |
41 |
> > b/misc/repo.postsync.d/00gemato |
42 |
> > > > > > new file mode 100644 |
43 |
> > > > > > index 000000000..f2af50925 |
44 |
> > > > > > --- /dev/null |
45 |
> > > > > > +++ b/misc/repo.postsync.d/00gemato |
46 |
> > > > > > @@ -0,0 +1,18 @@ |
47 |
> > > > > > +#!/bin/bash |
48 |
> > > > > > +# repo.postsync.d hook to verify ::gentoo checkout using gemato |
49 |
> > > > > > + |
50 |
> > > > > > +name=${1} |
51 |
> > > > > > +url=${2} |
52 |
> > > > > > +path=${3} |
53 |
> > > > > > + |
54 |
> > > > > > +# keyring installed by gentoo-keys |
55 |
> > > > > > |
56 |
> > +openpgp_key=/var/lib/gentoo/gkeys/keyrings/gentoo/release/pubring.gpg |
57 |
> > > > > > |
58 |
> > > > > |
59 |
> > > > > This seems a bit leaky to me. |
60 |
> > > > > |
61 |
> > > > > Possible to get gentoo-keys to print it? |
62 |
> > > > > |
63 |
> > > > > e.g: |
64 |
> > > > > |
65 |
> > > > > openpgp_key=$(gentoo-keys --print-key-path) |
66 |
> > > > |
67 |
> > > > But app-crypt/gentoo-keys doesn't include that executable, and |
68 |
> > it has |
69 |
> > > > no dependency on app-crypt/gkeys. I'd rather not introduce an |
70 |
> > artificial |
71 |
> > > > dependency here. |
72 |
> > > |
73 |
> > > I suppose we could using a separate ebuild to install this hook, |
74 |
> > so that |
75 |
> > > we can update it separately from portage if necessary. The hook can |
76 |
> > > still live in the portage repository (like emerge-delta-webrsync which |
77 |
> > > is also installed by a separate ebuild). |
78 |
> > |
79 |
> > I don't see a strong reason to add yet another rebuild for a single file |
80 |
> > that is going to be updated really rarely. However, if we're going to do |
81 |
> > it that way, then there's no point in putting it in Portage repository. |
82 |
> > |
83 |
> > However, this 'update it separately from portage' reminds me of repoman |
84 |
> > that frequently gets seriously outdated and/or incompatible with Portage |
85 |
> > because of independent release cycle... |
86 |
> > |
87 |
> > |
88 |
> > I'll rephrase my objection. |
89 |
> > |
90 |
> > I don't care what you do as long as Zac (the person releasing portage) |
91 |
> > agrees with whatever |
92 |
> > requirements you need. If we need 3 releases in a row because the hook |
93 |
> > is buggy, as long as |
94 |
> > Zac is happy with that I'm happy with that. |
95 |
> > |
96 |
> > What I don't want to see is surprise when the hook is cut and suddenly |
97 |
> > its buggy and we need new |
98 |
> > cuts and Zac is not around, or HEAD is broken, or some other problem. |
99 |
> > |
100 |
> > Looking at the release history, multiple cuts in O(few) days is fairly |
101 |
> > common (11/20, 11/21, 12/10, 12/15) |
102 |
> > so this seems low risk to me; but AFAIK Zac is usually driving these |
103 |
> > changes himself so its a bit more obvious |
104 |
> > what is going on. Or just allow Michał to cut his own portage releases |
105 |
> > when he needs hook updates. |
106 |
> > |
107 |
> > -A |
108 |
> |
109 |
> The thing is, this pubring.gpg path tightly couples the hook to gentoo-keys. |
110 |
> I'd feel much more comfortable about including it with portage if we |
111 |
> used something like this command to query the pubring.gpg location: |
112 |
> |
113 |
> $ gkeys list-key -C gentoo -n snapshot |
114 |
> |
115 |
|
116 |
I suppose I can live with that. However, gkeys are quite a heavy |
117 |
dependency compared to the keyring, with the latter being the only real |
118 |
requirement for gemato. That is, if that gkeys function works out |
119 |
of the box for non-root users because I recall gkeys having that |
120 |
problem. |
121 |
|
122 |
-- |
123 |
Best regards, |
124 |
Michał Górny |