1 |
On Thu, 09 Apr 2009 04:51:06 +0300 |
2 |
Mart Raudsepp <leio@g.o> wrote: |
3 |
> doins support for symlinks |
4 |
> ========================== |
5 |
> |
6 |
> Lacking information. Need to see if the PMS draft has anything about |
7 |
> it. The bug and summaries just talk about the support, but no |
8 |
> details. Would it be an argument to doins? |
9 |
|
10 |
The PMS draft explains it. It's simple, though: currently, doins on a |
11 |
symlink does something arbitrary. We want it to install the symlink. |
12 |
|
13 |
> unpack failing on unknown types |
14 |
> =============================== |
15 |
> |
16 |
> Uncertain about it's worthiness. Might rather have the opposite with a |
17 |
> --strict argument kind of deal. No official objection from me. |
18 |
|
19 |
What with the moving towards auto-die, the trend is towards "strict |
20 |
unless explicitly told not to be". |
21 |
|
22 |
> properties must be cached properly |
23 |
> ================================== |
24 |
> |
25 |
> No opinion, up to the package manager developers. |
26 |
> Don't see offhand why it should be an EAPI item at all. Feels like an |
27 |
> implementation detail. |
28 |
|
29 |
The cache format's tied to EAPI in a fairly unpleasant manner. It's a |
30 |
necessary evil. |
31 |
|
32 |
> DEFINED_PHASES must be supported (and cached properly) |
33 |
> ====================================================== |
34 |
> |
35 |
> No opinion, up to the package manager developers more or less. |
36 |
> Not sure why this needs to be an EAPI item. Hard to see a use case for |
37 |
> the variable being available for ebuild usage for it to be necessary |
38 |
> to be an EAPI item. |
39 |
|
40 |
DEFINED_PHASES isn't available to the ebuild. But it is in the metadata |
41 |
cache, which PMS has to cover. |
42 |
|
43 |
> By my understanding it is (also?) a required implementation detail to |
44 |
> handle pkg_pretend sanely and with minimal time hit. |
45 |
|
46 |
Correct. Without it there's a delay of ~0.1s per ebuild in the |
47 |
resolution set at pretend time, which soon adds up for certain |
48 |
reasonably common use cases. |
49 |
|
50 |
> Limit values in $USE to ones in $IUSE: |
51 |
> ====================================== |
52 |
> |
53 |
> Seems more of a QA test, but forcing it can make it be caught at |
54 |
> start. Don't see why it must be an EAPI item. Just vet the tree of |
55 |
> (future?) repoman warnings about it and make it happen for |
56 |
> all EAPIs. Impact on overlays is minimal because it is a QA error to |
57 |
> not define them and they get what they asked for. |
58 |
|
59 |
It needs to be in EAPI because we're talking about changing how the |
60 |
ebuild environment is specified. Also, repoman can't catch most |
61 |
accidental abuses of this. |
62 |
|
63 |
> --disable-dependency-tracking: |
64 |
> ============================== |
65 |
> |
66 |
> possible breakage of (custom) configure scripts that don't accept |
67 |
> unknown arguments. Would be nice to pass that for most packages, but |
68 |
> doing it always with econf seems slightly inappropriate, given the |
69 |
> above. |
70 |
|
71 |
Please provide a list of packages that use custom configure scripts, |
72 |
that currently work with econf (including all the weird things it |
73 |
already passes), that would break with this change and whose ebuilds are |
74 |
using econf. I have yet to see any examples provided. |
75 |
|
76 |
> utility commands should die by default |
77 |
> ====================================== |
78 |
> |
79 |
> Would like to see a list of those utility commands that would be made |
80 |
> to die by default. |
81 |
|
82 |
The PMS draft has one. |
83 |
|
84 |
> ban || ( foo? ( . ) . ) |
85 |
> ======================= |
86 |
> |
87 |
> It is not the business of an EAPI to start disallowing *DEPEND string |
88 |
> constructs. |
89 |
> There is no useful alternative provided yet to my knowledge and this |
90 |
> is really a QA issue, not an EAPI issue. |
91 |
|
92 |
There is no use case for the construct anyway. |
93 |
|
94 |
> Not convinced on the sub-optimal use case being the only one, either. |
95 |
> |
96 |
> Strongly objecting on the grounds that it is not something that should |
97 |
> be an EAPI issue. |
98 |
|
99 |
Behaviour of || ( foo? ( . ) ) is already an EAPI issue, and has a |
100 |
horrible section of PMS devoted to explaining its quirky behaviour. |
101 |
|
102 |
> I'm not even sure anymore where to find a list of items that is |
103 |
> current for what's on the table for EAPI-3 right now... |
104 |
|
105 |
Do a git log on the PMS draft (or look at the summary table in the |
106 |
appendix). It's fairly close to one commit per EAPI item. |
107 |
|
108 |
-- |
109 |
Ciaran McCreesh |