1 |
Am Sonntag, den 12.04.2009, 20:59 +0100 schrieb Ciaran McCreesh: |
2 |
> I've got the EAPI 3 branch for PMS more or less ready: |
3 |
> |
4 |
> http://github.com/ciaranm/pms/tree/eapi-3 |
5 |
> |
6 |
> The provisional included feature list is everything that was ready |
7 |
> before the deadline. |
8 |
Thanks a lot for your work. |
9 |
Sorry, I was quiet busy the last week with real life. |
10 |
|
11 |
> |
12 |
> Before the next Council meeting (ideally a week before, so we've got |
13 |
> time to get the questions worked out), I'd appreciate it if every |
14 |
> Council member could go through each item on the list below, check its |
15 |
> description in PMS (Appendix E has a handy list of links to the |
16 |
> relevant text, or look for the boxed labels in the margin) and |
17 |
> provisionally mark it as one of: |
18 |
> |
19 |
> * critical: EAPI 3 needs to be held up until this feature is in Portage. |
20 |
> |
21 |
> * yes: This feature should be in EAPI 3, but can be dropped if it turns |
22 |
> out it's going to take ages to get into Portage. |
23 |
> |
24 |
> * no: This feature shouldn't be in EAPI 3. |
25 |
> |
26 |
> * whatever: You have no strong opinion on whether this feature should |
27 |
> be in EAPI 3. |
28 |
> |
29 |
> * query: You have questions about this feature, or you think parts of |
30 |
> it need discussion, or you've found a mistake in the PMS draft. |
31 |
> |
32 |
> I'll try to address any queries as they come so people can update their |
33 |
> opinions. |
34 |
> |
35 |
> I'd also appreciate any review of wording features from anyone else. |
36 |
> There are probably still a few holes. |
37 |
> |
38 |
> Hopefully we can get a final list decided upon and provisionally |
39 |
> approved by the next Council meeting, and then as soon as Portage is |
40 |
> ready to go we can merge everything into PMS proper and get a signed |
41 |
> approval tag as we did for EAPI 2. |
42 |
> |
43 |
> Here's the list: |
44 |
> |
45 |
> * PKG-PRETEND |
46 |
critical. |
47 |
|
48 |
> * SLOT-OPERATOR-DEPS |
49 |
critical. |
50 |
|
51 |
> * USE-DEP-DEFAULTS |
52 |
critical. |
53 |
|
54 |
> * DEFINED-PHASES |
55 |
critical. |
56 |
|
57 |
> * PROPERTIES |
58 |
critical. |
59 |
|
60 |
> * SRC-INSTALL |
61 |
critical. |
62 |
|
63 |
> * CONTROLLABLE-COMPRESS |
64 |
no. |
65 |
|
66 |
> * DODOC |
67 |
critical. |
68 |
|
69 |
> * DOINS |
70 |
critical. |
71 |
|
72 |
> * ANY-USE |
73 |
yes. |
74 |
|
75 |
> * BANNED-COMMANDS |
76 |
yes. |
77 |
|
78 |
> * DOEXAMPLE |
79 |
whatever. |
80 |
... with dodoc getting recursive behaviour I guess this is not really |
81 |
critical. If portage penalizes it's users by compressing those examples, |
82 |
that's their problem. |
83 |
|
84 |
> * DOINCLUDE |
85 |
yes. |
86 |
|
87 |
> * UNPACK-EXTENSIONS |
88 |
yes. |
89 |
|
90 |
> * ECONF-OPTIONS |
91 |
yes. |
92 |
|
93 |
> * PKG-INFO |
94 |
critical. |
95 |
|
96 |
> * PROFILE-IUSE-INJECTION |
97 |
yes, but *_IMPLICIT has to be discussed. |
98 |
|
99 |
> * AA |
100 |
yes. |
101 |
|
102 |
> * KV |
103 |
yes. |
104 |
|
105 |
> * REPLACE-VERSION-VARS |
106 |
critical. |
107 |
|
108 |
> * S-WORKDIR-FALLBACK |
109 |
yes. |
110 |
|
111 |
> * UNPACK-IF-COMPRESSED |
112 |
yes. |
113 |
|
114 |
> * RDEPEND-DEPEND |
115 |
critical. |
116 |
|
117 |
> * DIE-ON-FAILURE |
118 |
yes. |
119 |
|
120 |
> * NONFATAL |
121 |
yes. |
122 |
|
123 |
> |
124 |
> Some answers to existing queries that I can remember: |
125 |
> |
126 |
> * DEFINED-PHASES and PROPERTIES have to be EAPI linked because of the |
127 |
> metadata cache. |
128 |
> |
129 |
> * ANY-USE is already special cased in the package manager and |
130 |
> specification. Everywhere that's using it is doing it wrong. It's |
131 |
> possible to rewrite the dep to mean same thing using an expanded |
132 |
> form, although it'll still be impossible to use it correctly if you |
133 |
> do that. |
134 |
> |
135 |
> * ECONF-OPTIONS is very unlikely to break custom configure scripts. We |
136 |
> already pass various weird things through econf, so any configure |
137 |
> script that dies on unrecognised options is probably going to break |
138 |
> anyway. No-one's managed to name a configure script that would be |
139 |
> broken by this change that isn't already not using econf anyway. |
140 |
... or using econf even if it shouldn't. |