- ghc-steering-committee

Please review: Constructor Synonyms / pattern synonym constructor signatures, Shephard: Chris Allen
by Joachim Breitner 20 Mar '17

20 Mar '17

Hi, this is your secretary speaking: https://github.com/ghc-proposals/ghc-proposals/pull/27 and https://github.com/ghc-proposals/ghc-proposals/pull/42 was brought before the committee. The proposals are related, so it makes sense to decide on them at once. I propose Chris as the Shepherd. (But this was a relatively arbitrary decision, and Chris, if you are unhappy with this, I can arbitrary appoint someone else.) Chris, please reach consensus as described in https://github.com/ghc-proposals/ghc-proposals#committee-process I suggest you make a recommendation about the decision, maybe point out debatable points, and assume that anyone who stays quiet agrees with you. FYI, the oher proposals under discussion at the moment are: Hex Floats Shephard: Iavor https://github.com/ghc-proposals/ghc-proposals/pull/37 Current tendency of the committee: In favor. INCOMPLETE_CONTEXTS Shephard: Roman https://github.com/ghc-proposals/ghc-proposals/pull/34 Current tendency of the committee: Against. Greetings, Joachim -- Joachim “nomeata” Breitner mail(a)joachim-breitner.de • https://www.joachim-breitner.de/ XMPP: nomeata(a)joachim-breitner.de • OpenPGP-Key: 0xF0FBF51F Debian Developer: nomeata(a)debian.org

2 4

Eval proposal (#27)
by Richard Eisenberg 15 Mar '17

15 Mar '17

I am shepherding the Eval class proposal, #29. Very, very briefly, this proposes to bring back > class Eval a where > seq :: a -> b -> b The author of the proposal cites some motivation, but a recurring theme in the discussion is that this is just the first step toward some ill-defined promised land. The author himself admits he’s not quite sure what the promised land holds, other than safety under eta-expansion, along with a restoration of parametricity. The proposal includes a new extension, on by default, -XUniversalEval, that adds Eval constraints in many places. Even with -XUniversalEval, however, this extension is not fully backward compatible, in corner cases (seq’ing in a class method; polymorphic recursion, possibly higher-rank types. I move to reject this proposal. The author (along with a few others) argues that this brings some nice theoretical properties to Haskell. I agree here. But the cost doesn’t seem to be worth the benefit, especially considering that this may be the first step toward some ill-specified goal. In short, while I might be happy enough with the language proposed here, I don’t relish the idea of getting from where we are to that language, and I don’t think the gain is worth the pain. You can see the PR here: https://github.com/ghc-proposals/ghc-proposals/pull/27 <https://github.com/ghc-proposals/ghc-proposals/pull/27> Richard

5 4

Status overview
by Joachim Breitner 10 Mar '17

10 Mar '17

Hi, as a courtesy service, here is what is currently on our plate: Hex Floats Shephard: Iavor https://github.com/ghc-proposals/ghc-proposals/pull/37 Current tendency of the committee: In favor. INCOMPLETE_CONTEXTS Shephard: Roman https://github.com/ghc-proposals/ghc-proposals/pull/34 Roman has not started to propose an action here. Roman, were you on the list when I sent https://mail.haskell.org/pipermail/ghc-steering-committee/2017-February/000… If not, Ben, can you add Roman to the list and the github organization (and maybe make me an owner)? The Eval class proposal (Shephard: Richard) was sent back to the author for further revision. Greetings, Joachim -- Joachim “nomeata” Breitner mail(a)joachim-breitner.de • https://www.joachim-breitner.de/ XMPP: nomeata(a)joachim-breitner.de • OpenPGP-Key: 0xF0FBF51F Debian Developer: nomeata(a)debian.org

3 4

Please review: Eval class, Shephard: Richard
by Joachim Breitner 01 Mar '17

01 Mar '17

Hi, this is your secretary speaking: https://github.com/ghc-proposals/ghc-proposals/pull/27 was brought before the committee. I propose Richard as the Shepherd. Richard, please reach consensus as described in https://github.com/ghc-proposals/ghc-proposals#committee-process I suggest you make a recommendation about the decision, maybe point out debatable points, and assume that anyone who stays quiet agrees with you. Greetings, Joachim -- Joachim Breitner mail(a)joachim-breitner.de http://www.joachim-breitner.de/

4 3

Redid our documentation
by Joachim Breitner 28 Feb '17

28 Feb '17

[I sent this mail from Monday to the wrong address, second try] Hi, there was a bit confusion about our documentation, so with SPJ’s blessing I went ahead and restructured it a bit. For now, this is only on a branch and not live yet: https://github.com/ghc-proposals/ghc-proposals/tree/wip/docs-restructur ing Note that it starts with a concise timeline of a proposal, and from there has links to section answering specific questions, within the same file. What was three files before (README, process, committee) is now all in here (with the exception of the “detailed instructions” for the Github-novice, which is in a separate file). This should make it easier for everyone involved to know who has to do what when. Our process is, however, flawed: We ask authors to set labels (“Under Discussion” and “Under Committee Review”), but they do not have permissions to do so. So this does not quite work. So I suggest the following change: * What was “Under Discussion” is now simply any PR that does not have any other label. This way, when opening discussion, nothing concrete has to be done. Which is easier. (GitHub allows to list all PRs that have no label, so there is no loss in functionality here.) * When the author wants to submit the PR, he sends a mail to this mailinglist (is this set up to accept mails from non-subscribers?) and it its the task of the shephard to set the label to indicate that that the committee has accepted to review the proposal. (At this point, the shephard could for example set the `Out-of-scope` label instead.) If there are no complains I will adjust the docs-restructuring branch accordingly and then move that to master. Greetings, Joachim -- Joachim “nomeata” Breitner mail(a)joachim-breitner.de • https://www.joachim-breitner.de/ XMPP: nomeata(a)joachim-breitner.de • OpenPGP-Key: 0xF0FBF51F Debian Developer: nomeata(a)debian.org

7 17

Please review: INCOMPLETE_CONTEXTS, Shephard: Roman
by Joachim Breitner 28 Feb '17

28 Feb '17

Hi, this is your secretary speaking: https://github.com/ghc-proposals/ghc-proposals/pull/34 was brought before the committee. I propose Roman as the Shephard. Roman, please reach consensus as described in https://github.com/ghc-proposals/ghc-proposals#committee-process (or refuse to shepherd and maybe propose someone else – I hand this job out based on who has already commented I suggest you make a recommendation about the decision, maybe point out debatable points, and assume that anyone who stays quiet agrees with you. Greetings, Joachim PS: I could not assign this PR to you on Github. Ben, can you add Roman to the team on Github, remove Atze, and update the list of committee members in the README (and maybe make me, as the secretary, an owner of the GitHub organisation)? -- Joachim Breitner mail(a)joachim-breitner.de http://www.joachim-breitner.de/

1 0

Review requested: Hex floats, Shephard: Iavor
by Joachim Breitner 27 Feb '17

27 Feb '17

Hi, this is your secretary speaking: https://github.com/ghc-proposals/ghc-proposals/pull/37 was brought before the committee. I propose Iavor as the Shephard. Iavor, please reach consensus as described in https://github.com/ghc-proposals/ghc-proposals#committee-process I suggest you make a recommendation about the decision, maybe point out debatable points, and assume that anyone who stays quiet agrees with you. Greetings, Joachim -- Joachim “nomeata” Breitner mail(a)joachim-breitner.de • https://www.joachim-breitner.de/ XMPP: nomeata(a)joachim-breitner.de • OpenPGP-Key: 0xF0FBF51F Debian Developer: nomeata(a)debian.org

1 0

Backpack, complete spec
by Christopher Allen 25 Feb '17

25 Feb '17

https://github.com/ghc-proposals/ghc-proposals/pull/5#issuecomment-280942000 Could I get a quick up/down on this? I'd rather not start making exceptions to the process we laid out. If a proposal is so complicated at the proposer cannot gather a complete specification of the proposal, it should under no circumstances be checked into source control. -- Chris Allen

1 0

Prioritizing proposals for processing? Viewing dates/deadlines?
by Ryan Newton 20 Feb '17

20 Feb '17

Dear steering committee, How do we see what the deadline is for a given proposal discussion? I know we've discussed process a fair bit, but I seem to quickly forget these things if they're not super simple. I can go here <https://github.com/ghc-proposals/ghc-proposals/pulls?q=is%3Apr+is%3Aopen+so…>, and see proposals, oldest first. (The PR numbers effectively date them as well.) The labels are super useful too (Under discussion). Richard's original proposal is attached below. SPJ later suggested "2 weeks"->"4 weeks" for the consideration period. I think that corresponds to the "Pending Committee review" label. But right now, *no proposals have that label*. For example, how to judge the stage of PR number 36 <https://github.com/ghc-proposals/ghc-proposals/pull/36>?: - has the "Under discussion" label - was posted Jan 15 (1 month old) - has been commented on by some committee members - doesn't have a corresponding email thread on "ghc-steering-committee" - doesn't have a shepherd, which I *think *would appear as an "Assignee" on the PR/issue, right? For better or worse I live a very deadline driven life ;-). So it would be nice to see which proposals have a decision deadline and act there first. As a small technical point, how do we record the DATE when the "Pending committee review" label is assigned, which is what starts the clock ticking right? Thanks, -Ryan P.S. It seems like the decisions about timing and process haven't filtered through to the front-page / top level README <https://github.com/ghc-proposals/ghc-proposals>. Which says only: * "Proposals are ultimately evaluated by the GHC Steering Committee <https://github.com/ghc-proposals/ghc-proposals/blob/master/steering-committ…> based upon a number of criteria and in light of community feedback."* On Fri, Nov 18, 2016 at 8:56 AM, Richard Eisenberg <rae(a)cs.brynmawr.edu> wrote: > ..... > I propose the following: > > - Establish an official ghc-committee(a)haskell.org mailing list that > reaches just us. It will not be open to the public to join, though I have > no problem making archives public. > - When the author of a proposal (or anyone else, really, if the author > wanders off) deems the proposal ready for a decision, that author emails > the list telling us so. > - We use an organic process to decide on one individual in the committee > who will oversee the discussion on the proposal. If organic doesn’t work, > our chair(s) assign the proposal to a member. It is expected that > membership on the committee means that we will volunteer to handle > proposals as appropriate. The committee member running this discussion > process is hereby titled the Shepherd of the proposal. (NB: This is > slightly different than my understanding of a Shepherd in Rust, who is > assigned earlier in the process.) > - Neither the shepherd nor the committee is *not* responsible for reading > GitHub (or other) commentary. The proposal will be considered on its own. > If the author wishes the committee to consider any commentary, that > commentary should be incorporated into the proposal. > - Once a decision is requested, the shepherd has two weeks (in holiday > times or near the ICFP deadline, 3) to generate consensus. If consensus is > elusive, then we vote, with the Simons retaining veto power. > - If we say no: the shepherd updates the proposal (not just the > commentary) with the reasons for rejection. The proposer is welcome to > revise and try again, but the document should retain this original > rejection information. > - If we say yes: A Trac ticket is created, referring back to the proposal > and commentary. (The shepherd is responsible for making sure this happens.) > At this point, the proposal process is technically complete. I believe it > is outside of our purview to implement, oversee implementation, attract > implementors, etc. Naturally, we will want to do this as individuals, but I > believe it’s not in our remit. > >

3 2

Defaulting rules
by Simon Peyton Jones 10 Feb '17

10 Feb '17

Dear GHC steering committee Advice needed. * Some time ago amindfv submitted https://ghc.haskell.org/trac/ghc/ticket/12923. He offered a patch 8 weeks ago. * There wasn't much discussion about the proposed change, nor did anyone ask him to undertake the GHC proposals process * It's a small change, but it is a user-facing one * The design isn't obvious to me; see comment:12. * We could change it now and change it again later; it's a corner case * It matters to him and his customers. So, should we let it in? It's about the defaulting mechanism. With an unsolved constraint (C alpha) where alpha is otherwise unconstrained, GHC will try a series of types to fill in alpha. We extended the rules with ExtendedDefaultRules<https://downloads.haskell.org/~ghc/latest/docs/html/users_guide/ghci.html#t…>. He wants a further extension. I hate slipping things in late, actually after the deadline; but the fault is partly ours for not proactively warning people with open patches. Opinions? Simon

1 1