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
|
\summary{2016}{6}{12}
Agenda: \agoref{gentoo-project}{50dbe189dd2641d5730f08944e7fa7ce}
\agendaitem{Discussion on mgorny's items}
\index{INSTALL\_MASK}\index{USE\_EXPAND}\index{LINGUAS}\index{L10N}
\index{package!sys-devel/gcc}\index{use flag!multislot}
Reference: \agoref{gentoo-dev}{68a870c0519fb1cb7152db38fc9d4935}
Michał Gorny proposed several unrelated items in a lengthy e-mail:
\begin{enumerate}
\item a "file installation mask" GLEP,
\item for portage to actually adhere to USE\_EXPAND semantics,
\item\label{mgms} to abolish the sys-devel/gcc multislot use flag entirely,
\item\label{mgln} the migration from LINGUAS USE\_EXPAND to L10N
\item discussion of the gui use flag.
\end{enumerate}
He later withdrew all these proposals. Items \ref{mgms} and \ref{mgln} were
discussed anyway.
Regarding \ref{mgms}, \bug{584610}, blueness as toolchain member explained that
now the new implementation of the multislot use flag in sys-devel/gcc
controls a blocker. If unset, previous gcc versions are blocked and
consequently unmerged on update (leading to potential ABI breakage in
C++). If set, minor version bumps accumulate on update.
This led to a discussion why 1) the blocker exists, why 2) the slot
was refined from major version bumps (4.8, 4.9) to point releases
(4.8.1, 4.8.2, ...), and why this behaviour was not reserved to an
overlay for toolchain specialists. For none of these questions valid
technical arguments were known.
After discussion blueness proposed to e-mail vapier, request
clarifications on the changes and ask him if he wants to force
USE=multislot universally, since this seems to be the easier choice.
Regarding \ref{mgln}, see
\begin{itemize}
\item \agoref{gentoo-dev}{a08ea09c2c8e534fd9bc1146703c66ff}
\item \agoref{gentoo-dev}{41e09d1ddc8b30abb9f9d21d205b7b82}
\end{itemize}
ulm explained that LINGUAS should become a normal environment variable
(since it has a special meaning to gettext) and the current USE\_EXPAND
functionality should be transferred to a new variable L10N. No council
decision is required; the new variable will be introduced in a news
item.
\agendaitem{Bugs with council involvement}
Dilfridge was reminded to complete the missing summaries; blueness
announced that he just finished two summaries.
No action was taken regarding games.eclass and the games team.
\bug{565566} was mentioned but no action was taken.
\agendaitem{Subsequent non-public meeting}
A non-public discussion of bug 592598 followed.
|
GitWeb
