1 |
On Sat, 5 Nov 2005 05:24:51 -0600 Brian Harring <ferringb@g.o> |
2 |
wrote: |
3 |
| Drop the lightweight bit and merge it into multiple delivery. You're |
4 |
| after lightweight _default_, which is fine, but the phrasing is a bit |
5 |
| screwed. |
6 |
|
7 |
Hrm, I don't see those as contradictory. There's a requirement that the |
8 |
solution does not require anything fancy, and a requirement that the |
9 |
solution does not force one particular reading method upon the user. |
10 |
There's no requirement that the solution must not be usable with |
11 |
anything which is not lightweight. |
12 |
|
13 |
| > News items are published and delivered to users as follows: |
14 |
| <snip> |
15 |
| > 6. Portage filters the news item and, if it is relevant, installs |
16 |
| > it in a special location to be read by a news item reader. Messages |
17 |
| > are also displayed to inform the user that news is available. |
18 |
| |
19 |
| Out of curiousity, how are you planning on having 101 general notices |
20 |
| reigned down on a users head during an initial install? Yes, you |
21 |
| have the atom filter, but what about general notices? |
22 |
|
23 |
People who've just done an initial install won't have lots of packages |
24 |
installed. As for general notices, if there're lots of them we have |
25 |
serious issues... |
26 |
|
27 |
| Further, how are notices going to apply to users who don't have the |
28 |
| package installed, then go and merge it? Your statement above |
29 |
| implies the irrevalent (at the time of sync) notices are ignored, |
30 |
| which breaks down under the scenario where a news item is pushed out |
31 |
| that apache configuration is going to change, I merge old style |
32 |
| apache after sync'ing the news, portage (due to your news id cache) |
33 |
| considers it deleted, and doesn't display it. |
34 |
|
35 |
Wording perhaps could be clearer... News items are added to the "don't |
36 |
copy again" list when they're copied, not when they're checked. |
37 |
|
38 |
| > News items may be signed using GPG. If this is done, a detached |
39 |
| > signature should be used. |
40 |
| |
41 |
| I'd argue for must be, personally. A bogus news item claiming to be |
42 |
| from portage devs, telling users to change their SYNC setting could |
43 |
| cause massive mayhem. |
44 |
|
45 |
Signing elsewhere isn't mandatory yet. |
46 |
|
47 |
| Still haven't address my point about |
48 |
| A) package moves combined with news entries |
49 |
|
50 |
Gotta handle those manually / with epkgmove. Someone could write a |
51 |
server-side handler for automatic updates if they want, but given that |
52 |
package moves are already a case of "do all the things on a big list", |
53 |
it's not much added complexity... |
54 |
|
55 |
| B) N packages |
56 |
| |
57 |
| Assume you mean that the bit above is a full DepSet, not a single |
58 |
| atom (if that's the case, clarify that N atoms are allowed). |
59 |
|
60 |
A single atom is probably sanest... |
61 |
|
62 |
| Should clarify USE conditionals are a no go- might be worth |
63 |
| considering the full atom syntax (despite the fact portage doesn't |
64 |
| currently support it for depends), use flags, slot, etc. |
65 |
|
66 |
I'd rather not try to guess what Portage may or may not support in the |
67 |
future. There's nothing precluding using :slot and [use] atoms if |
68 |
portage ever supports them. |
69 |
|
70 |
| > ``Display-If-Keyword:`` |
71 |
| > A keyword [#glep-22]_ name, for example ``mips``. If the user |
72 |
| > is on the arch in question, the news item should be displayed. |
73 |
| |
74 |
| N keywords == N Header entries? |
75 |
|
76 |
Yup. |
77 |
|
78 |
| No go on -core imo; it's a community/dev issue, should be visible to |
79 |
| the general public rather then hidden away in the ml we do our |
80 |
| flaming in. |
81 |
|
82 |
There *might* be legit security reasons for using -core, for example |
83 |
for nasty "upgrade required" security bugs that we can't disclose |
84 |
before a given date. Hopefully this will never happen. |
85 |
|
86 |
| Already pointed out that this won't fly looking forward, it |
87 |
| implicitly assumes a single repository. |
88 |
|
89 |
Again, we can deal with that if Portage ever gets multiple repo |
90 |
support. Until it does, I'm not trying to guess how it's going to end |
91 |
up being implemented. |
92 |
|
93 |
| Nuke flashy (any phrasing that allows for blinking crap sliding into |
94 |
| portage I instinctively dislike). |
95 |
|
96 |
Is there a technical name for the big red !!!!! messages? |
97 |
|
98 |
| > Portage must keep track of news items which have already been |
99 |
| > installed to avoid repeatedly reinstalling a deleted news item. |
100 |
| |
101 |
| Track it how, via the directory, or a secondary list? |
102 |
|
103 |
The GLEP doesn't require any particular implementation. |
104 |
|
105 |
| > News items can be removed (by removing the news file from the main |
106 |
| > tree) when they are no longer relevant, if they are made obsolete |
107 |
| > by a future news item or after a long period of time. This is the |
108 |
| > same as the method used for ``updates`` entries. |
109 |
| |
110 |
| Might want to consider a maximum period of time for news entries to |
111 |
| linger by default; perhaps a header for controlling deletion (this is |
112 |
| would require commit access for whatever does the scans obviously). |
113 |
|
114 |
We don't do this with updates or ebuilds or anything else, so I don't |
115 |
really see the point. |
116 |
|
117 |
| Stop using trivial (no it's not in reference to above, just hit the |
118 |
| right word count where I'm whinging about it). |
119 |
|
120 |
But it's such a nice word! |
121 |
|
122 |
| Points above regarding working sanely for N repos need be addressed, |
123 |
|
124 |
Which I'll do right before Portage gets N repo support. |
125 |
|
126 |
| So... reiterating jasons question, example where news items transcend |
127 |
| package specific? |
128 |
|
129 |
Easy example? There might be changes in vim7 which would warrant a news |
130 |
item covering vim, gvim, vim-core, kvim. |
131 |
|
132 |
-- |
133 |
Ciaran McCreesh : Gentoo Developer (Anti-XML, anti-newbie conspiracy) |
134 |
Mail : ciaranm at gentoo.org |
135 |
Web : http://dev.gentoo.org/~ciaranm |