summaryrefslogtreecommitdiff
blob: dd2d4b226d2506830b2bca5a6259b061c334284e (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
[20:01:20] <hwoarang> !time
[20:01:20] <willikins> hwoarang: Europe - London - Tue Jun 12 20:00 BST
[20:01:39] <hwoarang> jmbsvicetto grobian ulm dberkholz Betelgeuse a3li
[20:01:40] <hwoarang> ping ^
[20:01:46] <grobian> hwoarang: pong
[20:01:46] <hwoarang> who's here
[20:01:47] <ulm> pong
[20:02:22] <jmbsvicetto> pong
[20:02:27] <jmbsvicetto> I'm here
[20:02:45] <jmbsvicetto> I poked Betelgeuse on #-devrel a few minutes ago and
he didn't reply, so we might need to call him
[20:03:32] <hwoarang> seems like we miss quite a few people
[20:03:46] <jmbsvicetto> a3li: ping
[20:04:28] <jmbsvicetto> I'm calling Betelgeuse
[20:04:58] <a3li> jmbsvicetto: hi
[20:05:06] <grobian> jmbsvicetto: thx
[20:05:42] <jmbsvicetto> He should show up soon
[20:05:55] <hwoarang> good
[20:05:57] <grobian> good, allows me to make a pot of tea ;)
[20:05:59] <hwoarang> we still miss dberkholz
[20:06:20] <hwoarang> i'll wait 4' more minutes (till 20:10) and then move on
[20:10:47] <Betelgeuse> hello
[20:10:54] <Betelgeuse> Sorry about the delay
[20:10:59] <grobian> just in time (tm)
[20:11:05] <Betelgeuse> My 4G stick had trouble getting an address
[20:11:11] <hwoarang> !time
[20:11:12] <willikins> hwoarang: Europe - London - Tue Jun 12 20:10 BST
[20:11:18] <hwoarang> ok time to begin
[20:11:34] <hwoarang> agenda:
http://archives.gentoo.org/gentoo-project/msg_affe2989c83112c5b08dcd77226790fa.xml
[20:12:40] <hwoarang> ssuominen proposed to allow extra arguments to default
src_configure and src_install for EAPI5
[20:12:44] <hwoarang> if I understand correctly.
[20:13:17] <grobian> to "default" in src_configure and src_install
[20:13:50] <hwoarang> yes
[20:13:52] <ulm> I'm against it. phase functions don't take arguments
[20:14:07] <grobian> I've already expressed my feelings, and it hasn't changed
much
[20:14:33] <grobian> I don't like things like "default -j1"
[20:14:47] <hwoarang> if everyone has a clear head in the topic, this seems
like a straight yes/no voting
[20:14:54] <a3li> same for me (and chainsaw). it doesn't feel right to call it
'default' and then make it parametrized
[20:14:56] <grobian> without proper specification I'm against it anyway
[20:15:05] <hwoarang> I am in favor of samuli's proposal fwiw
[20:15:27] <hwoarang> grobian: what do you mean by proper specification?
[20:15:32] <hwoarang> i think it is pretty clear
[20:15:40] <hwoarang> make default understand $@
[20:15:44] <grobian> hwoarang: do default -j1 is ok to you?
[20:15:51] <hwoarang> so whatevery you pass to default it will be appended
[20:16:12] <ulm> it's not natural what would be the meaning of $@ e.g. in
src_install
[20:16:13] <grobian> what do the arguments mean in the future when src_install
and src_configur change?
[20:16:25] <grobian> ^^ what ulm said
[20:16:29] <hwoarang> grobian: i will save some bytes than having to
re-implement src_install { emake -j1 install } for example
[20:16:33] <ulm> $@ could be passed to emake, or to dodoc
[20:16:45] <Betelgeuse> I would solve the problem with additional variables
[20:16:54] <ulm> Betelgeuse: yeah
[20:16:57] <jmbsvicetto> I agree with Betelgeuse
[20:17:05] <hwoarang> ulm: i think it is obvious that the arguments would be
passed to emake calls
[20:17:05] <grobian> hwoarang: just make it explicit, don't introduce lots of
vagueness because of saving a few bytes
[20:17:07] <hwoarang> and econf 
[20:17:21] <grobian> hwoarang: then why not call it emake and econf?
[20:17:30] <hwoarang> becaue you like the default bits
[20:17:34] <hwoarang> you just need some extra stuff
[20:17:35] <jmbsvicetto> instead of allowing anything to be passed to the
phase function, create some sensible variables that can be used to alter the
default behaviour
[20:17:41] <grobian> anyway, enough, this is discussion that should go on ML
[20:18:03] <grobian> now jmbsvicetto is going to get to something which gets
closer to a spec
[20:18:07] <hwoarang> yeah but we need to give some hints what to discuss
[20:18:17] <hwoarang> to from what I can tell $@ is not acceptable
[20:18:23] <hwoarang> and the majority wants new variables
[20:18:24] <grobian> hwoarang: honestly, the whole initial thread was full of
hints
[20:18:26] <hwoarang> is that correct?
[20:18:51] <jmbsvicetto> hwoarang: econf_opts / emake_opts / einstall_opts ?
[20:18:58] <hwoarang> grobian: yes but i believe we need to give a direction
[20:19:04] <hwoarang> of what *we* think it is appropriate
[20:19:06] <hwoarang> as a solution
[20:19:10] <hwoarang> jmbsvicetto: that looks ok to me
[20:19:14] <ulm> EMAKE_INSTALL_EXTRA_ARGS ;)
[20:19:14] <_AxS_> would it make sense at this point to vote (down) the
allowance of $@ (aka parameters) in src_compile/src_install ?
[20:19:41] <hwoarang> _AxS_: no because $@ does not make sense anyway at this
point
[20:19:48] <_AxS_> ah ok.
[20:19:50] <hwoarang> default does not understand $@ from what I can tell
[20:19:55] <hwoarang> so doesnt matter if you use it
[20:20:34] <hwoarang> fair enough. I will make a comment on the bug and ML
thread, that using suitable variables is preferred
[20:20:38] <hwoarang> everybody ok with this ^^
[20:20:39] <grobian> hwoarang: perhaps it would help if there was a strong
usecase that really showed the use of allowing arguments
[20:21:20] <a3li> hwoarang: personally, I guess I'd go with "don't use default
if you need to change things"; but that's better than default <stuff>
[20:21:36] <grobian> hwoarang: perhaps you/samuli is just looking for an extra
phase to be introduced
[20:22:05] <grobian> src_makeinstall + src_install
[20:22:10] <Betelgeuse> a3li: Experience with java eclasses showed that it's
bitter if all similar logic goes through eclasses
[20:22:24] <Betelgeuse> a3li: It was a big pain to introduce anything new when
all was copy pasted
[20:22:28] <jmbsvicetto> hwoarang: anyway, it's something for the next council
to decide
[20:22:33] <ulm> Betelgeuse: "bitter"?
[20:22:39] <hwoarang> grobian: i dont think so. Jus trying to use "default"
for weird build systems that DESTDIR alone may not make sense
[20:22:40] <Betelgeuse> better
[20:22:43] <Betelgeuse> sorry
[20:23:01] <_AxS_> ulm: everything is bitter when talking about java pkgs :)
[20:23:17] <a3li> Betelgeuse: so if you have a common digression from the
'default', you have an eclass to abstract it there. or do I miss the point?
[20:23:27] <grobian> hwoarang: yeah, so just override it with src_makeinstall
to emake install bla=foo, then src_install does the dodoc stuff etc.
[20:23:45] <hwoarang> grobian: and what about src_configure?
[20:23:58] <Betelgeuse> a3li: There's been variables and hooks to provide for
that
[20:24:15] <grobian> that usecase is even weaker, and I really think you
should be just typing econf again
[20:24:43] <ulm> hwoarang: src_configure calls econf only, so you say "econf
<foo>" instead of "default <foo>"
[20:24:52] <ulm> it's even shorter
[20:24:58] <a3li> and has better semantics
[20:25:01] <hwoarang> ok
[20:25:02] <ulm> and much clearer if you read it
[20:25:05] <grobian> yup
[20:25:10] <grobian> no magic implied
[20:25:14] <grobian> s/no/less/
[20:25:57] <jmbsvicetto> So are we ready to move on?
[20:26:06] <hwoarang> ok to sum up: no $@ for default, push back to ML and
give "new variables" or "new phase between src_configure and src_install"
[20:26:17] <hwoarang> and let discussion goin
[20:26:25] --> nimiux (~nimiux@gentoo/developer/nimiux) has joined
#gentoo-council
[20:27:00] <hwoarang> moving on
[20:27:09] <hwoarang> *Open floor*
[20:27:16] <hwoarang> anyone want to discuss something?
[20:27:32] -*- hwoarang looks around
[20:27:37] <Betelgeuse> I want to know why there's still no ice craem machine?
[20:28:08] <jmbsvicetto> I just want to thank you guys. It was a pleasure to
work with all of you