Compiler Team 2018-10-23 Retrospective

	A	B	C	D	E	F	H	I	J	K	L	M	N	O	P	Q
1	😍 — Things that are awesome!						😱 — Things that desperately need improvement
2	staying on tops of regressions basically works	8					rfcbot can idle for weeks	1
3	positive friendly communication	3	x				triage meetings always over time, sometimes hung up on 1 thing	2
4	work planning	1					hard to answer "what to do next" or "how can I help"	3	x	x	x	x	x	x	x	x
5	working groups as pathway to regular contribution	4	x				hard to find folks to fix regressions, particularly in some areas (e.g. LLVM)	2			x
6	having clear goals (when we do)	1					hard to do high-level design + planning and not just triage	4	x	x	x
7	delegating and mentoring to newcomers	7	x	x		x	no central place to get "big picture" of what's going on	5	x	x	x
8	public meetings on discord/zulip	2	x				leaders stretched thin, need to scale up	1	x
9	error message work	1	x	x	x		technical debt in compiler, heterogeneous structure	2	x	x	x
10	new crater (horray pietroalbini)	1					best practice workflow (--keep-stage etc) not well-documented or tested	1	x		x
11	zulip	1					perf regressions: we track, but don't always react unless someone complains	1
12	members are responsive on zulip/discord	1	x				more focus on features relevant to roadmap	1	x
13	implementation gets done	3					more focus on QoL improvements	1	x
14	consensus usually reached quickly	2	x				More regular reviewing of assigned PRs could be helpful	1
15	code has debt, but not overwhelming	1					communication with IDE folks	1	x	x
16	internals threads for public debate	1					Issues tagged T-compiler could use more info to get newbies started	1
17	public roadmap	1	x	x	x	x	More intermediate tasks for "newcomers"	1	x
18	blog posts on specific dev topics	1	x				sometimes chaotic, held together by sheer force of will vs organization	1
19							modules with low bus factor	2	x	x				x
20							more documentation, rustc book not easy enough to find	2	x
21							more standalone crates that can be independently built/tested	1	x			x
22							faster build times (for building `rustc` itself)	1			x		x	x		x
23							centralize on one chat platform	1	x
24							bors time is long	1	x	x		x		x
25							sometimes hard to know who to ask to review	1	x
26							Adding retrospectives in order to improve the iterations,	1
27							having clear deadlines on issues when there is a rush	1	x
28
29
30	Priorities
31	fast compilation perf	72%
32	RLS	60%
33	cleanup, debt	50%
34	trait system improvements	40%
35	const generics	36%
36	bors time	36%
37	parallelized compilation	31%
38	NLL improvements	27%
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