- ghc-steering-committee

Template too large, removed direct link
by Joachim Breitner 18 Sep '23

18 Sep '23

Hi, we used to have a direct link in our README to start a new template pre-filled with the template, but that only worked as long as our template text was short enough to fit into an URL, and would just give an error message now. I have removed that from the README for now. It requires proposal authors to understand enough git to know how to create a new file and copy some other file into it. Maybe not too high a hurdle. We can add the link back if we have shrunk our template to a sensible size again… these things grow like nothing else, besides maybe bureaucracy. If someone makes a heroic effort to make the template more concise without losing anything important (I think the endorsement section can go as it is unused, but that alone is not enough), that would be great! Cheers, Joachim -- Joachim Breitner mail(a)joachim-breitner.de http://www.joachim-breitner.de/

1 0

GHC steering committee status
by Simon Peyton Jones 13 Sep '23

13 Sep '23

Dear GHC steering committee Re GHC proposals I was talking to Simon M about - Making the shepherd's job easier and quicker - Providing a way for members of the committee to say (explicitly) "I don't have an opinion about this proposal" and thus recuse themselves. Explicit recusal is better than "silence means assent" because silence can, and often does, mean "I'm under water and not listening". With that in mind, I want to revive my suggestion of running an online spreadsheet <https://docs.google.com/spreadsheets/d/1e6GdwHmAjeDEUhTvP-b18MDkpTfH3SMHhFu…>to describe the state of proposals that are in our active purview. That is, I'm not proposing to show proposals that are under discussion by the community, or accepted etc -- just the ones that are in our inbox, *where the next action is ours.* That makes it much easier to see our current inbox, without looking back for Jocahim's last email and then adding deltas for all the emails since -- which I cannot do in my head. Specifically, I have *a column for each committee member's vote. * That should make it easy for the shepherd to see who is yet to express an opinion, which I always find difficult. As I say above, I am very uncomfortable with "silence means assent" for any but the smallest proposals. *I suggest instead that if there is a proposal where you feel unable to offer an opinion, due to lack of expertise, or lack of time, you can say "recuse" meaning that you explicitly want to stand down from this particular proposal.* (Somehow "abstain" carries a connotation of conflict of interest or something, but it amounts to the same thing.) If you are too under water to even recuse yourself, maybe it's time to step down. So in columns F to O we should see explicit responses from every committee member, in a timely way. Joachim doesn't like spreadsheets like this because they can easily get out of date. But it must be better than manually trawling email. *And it is up to each of us (not Joachim) to fill in our own column for proposals in our inbox.* If we don't like it, we can change it. You all have edit rights. I have populated it with data from Joachim's message of 2 Sept, but there has been some action since then, so shepherds please update it. This is not Joachim's task! Thanks Simon

5 7

Fine-Grained Unused Warnings (#42)
by Chris Dornan 09 Sep '23

09 Sep '23

Proposal: Fine-Grained Unused Warnings (#42) Author: Jakob Brünker Rendered proposal: https://github.com/JakobBruenker/ghc-proposals/blob/fine-grained-unused/pro… Discussion: https://github.com/ghc-proposals/ghc-proposals/pull/434 Recommendation: Acceptance ## Summary GHC currently warns equally about any superfluous binding that it discovers, regardless of whether the definition is unreferenced anywhere itself, or is merely mentioned by other definitions that are similarly superfluous. The proposal proposes to replace these warnings with just the warnings about definitions that are not referenced anywhere, but list within those warnings the entities it references that would otherwise not be required. The same information is being conveyed under this proposal as previously, but in a more structured manner. The proposal is careful to ensure the -Wunused-* controls (-Wunused-top-binds, etc.) are respected, so the subwarnings of transitively unused entities are only listed if their corresponding -Wunused-* control is active. As Jakob points out in the proposal, getting these warnings into some structure is partly motivated by IDE usage where the 'additional [transitive] warnings take up a lot of valuable screen real estate'. ## Example [From the proposal] In the following code, module Foo () import Data.List as L foo = L.intercalate `intercalate` is transitively unused, since it is only used by `foo`, which is unused. The warning that's produced will be Foo.hs:5:1: warning: [-Wunused-top-binds] Defined but not used: ‘foo’ as a consequence the following imports are transitively unused [-Wunused-imports]: Data.List imported at foo.hs:3:1: The warning about transitively unused bindings only appears if both the warning flag for the unused binding (in this case -Wunused-top-binds) and the the one for the transitively unused binding itself (in this case -Wunused-imports) are enabled. ## Recommendation This proposal is not at all disruptive (arguably to the extent of not requiring a GHC proposal review), is well thought out, and can improve the GHC/IDE UI so I recommend that we accept it.

4 4

GHC Steering Committee Status
by Joachim Breitner 04 Sep '23

04 Sep '23

Dear Committee, time for another status email, in particular because I need an up-to-date update to get stats for Simon’s famous report at HIW on Monday. Hope to see some of you there! Let’s see what has happened since July 3 * Arnaud did round 3 of his language extension policy investigation. * we were asked to review these proposals: #597: Constraints synonyms in deriving heads, Shepherd: Moritz #604: do not let -XScopedTypeVariables imply -XTypeAbstractions, Sepherd: Joachim #608: no implicit bindings with -XPatternSignatures, Shepherd: Richard #583: HasField redesign, Shepherd: Arnaud #609: (p1; p2) for or-patterns, Shepherd: Chris #607: Amendments clarify treatment of term variables in types, Shepherd: Vlad #605: visible forall to work without ScopedTypeVariables, Shepherd: Joachim #601: Extension lifecycle, Shepherd: Simon PJ * we have a recommendation from the shepherd about: #604: do not let -XScopedTypeVariables imply -XTypeAbstractions, rec: accept #605: visible forall to work without ScopedTypeVariables, rec: accept #607: Amend #281 (visible forall) and #378 (Design of DH) to clarify treatment of term variables in types (rec: accept) #583: HasField redesign, rec: accept #601: Extension lifecycle, rec: accept * we have sent the following proposals back to revision #601: Extension lifecycle * we decided about the following proposals #596: sized literals in Show, accept #581: Namespaces-specified Imports, accept #604: do not let -XScopedTypeVariables imply -XTypeAbstractions, accept #607: Amend #281 (visible forall) and #378 (Design of DH) to clarify treatment of term variables in types (rec: accept) #605: visible forall to work without ScopedTypeVariables, accept So we currently have to act on the following 9 proposals, up one since last time: ## Waiting for committee decision #512: NoFieldSelectors as datatype annotation, Shepherd: Vlad 2022-09-03: Assigned to Baldur 2022-10-02: Reassignd to Vlad 2022-11-30: Vlad recommends rejection This needs to be picked up again, it has been lingering too long! #536: Type-level literals as a sep language extension, Shepherd: Vlad 2023-02-16 Assigned to Vlad 2023-03-06 Vlad recommends acceptance Some hesitancy in the committee to introduce this extension, the ongoing policy clarification process might guide us, but maybe we can resolve this earlier? #583: HasField redesign, rec: accept 2023-08-30 Arnaud recommends acceptance ## Waiting for shepherd recommendation #526: Applicative Comprehensions, Shepherd: Simon M 2022-10-08: Assigned to Simon M Simon, please cast your recommendation! #585: Amend Or Patterns, Shepherd: Richard 2023-04-11: Assigned to Richard There are votes on syntax going on, ran by Sebastian Graf. Maybe this should be in “needs revision” until they are through. Richard, I’ll leave this to you #609 seems to be related, not sure if we need both. #194: Updated partial type signatures, Shepherd: Eric 2023-05-25: Came back from needs revision #434: Fine-grained unused warnings, Shepherd: Chris 2023-07-02: Assigned to Chris #597: Constraints synonyms in deriving heads, Shepherd: Moritz 2023-07-25: Assigned to Moritz #608: no implicit bindings with -XPatternSignatures, Shepherd: Richard 2023-08-17: Assigned to Richard Cheers, Joachim -- Joachim Breitner mail(a)joachim-breitner.de http://www.joachim-breitner.de/

4 3

Please review #601: Extension lifecycle, Shepherd: Simon PJ
by Joachim Breitner 03 Sep '23

03 Sep '23

Dear Committee, David Thrane Christiansen suggested to categorize extensions into Experimental, Mature, Deprecated and Legacy, and add warning flag to GHC that allow users to be warned about (or shouted at for) using such extensions, if they choose so. https://github.com/ghc-proposals/ghc-proposals/pull/601 https://github.com/david-christiansen/ghc-proposals/blob/extension-lifecycl… Because of the meta-like aspect of this proposal, I’d like to assign this to Simon PJ. Please guide us to a conclusion as outlined in https://github.com/ghc-proposals/ghc-proposals#committee-process Cheers, Joachim -- Joachim Breitner mail(a)joachim-breitner.de http://www.joachim-breitner.de/

10 22

Please review #194: Updated partial type signatures, Shepherd: Eric
by Joachim Breitner 02 Sep '23

02 Sep '23

Dear Committee, Richard resubmits updated partial type signatures: https://github.com/ghc-proposals/ghc-proposals/pull/194 https://github.com/goldfirere/ghc-proposals/blob/updated-partial-type-sigs/… I’d like to ask Eric to shepherd this proposal. Please guide us to a conclusion as outlined in https://github.com/ghc-proposals/ghc-proposals#committee-process Cheers, Joachim -- Joachim Breitner mail(a)joachim-breitner.de http://www.joachim-breitner.de/

2 4

#607: Amend #281 (visible forall) and #378 (Design of DH) to clarify treatment of term variables in types (rec: accept)
by Vladislav 02 Sep '23

02 Sep '23

Dear Committee, Jakob Brünker has proposed #607, an amendment to proposals #281 (visible forall) and #378 (Design of DH). See the diff here: <https://github.com/ghc-proposals/ghc-proposals/pull/607/files> The amendment concerns type checking of term-level variables appearing in types, for example: f :: forall (a :: Bool) -> ... -- visible forall, `f` expects a type-level Bool g :: Bool -> ... -- ordinary arrow, `g` expects a term-level Bool g x = f x What should happen to `x`, bound as a term variable but used as a type? #378 "Design for Dependent Types" already has an answer to this question: treat `x` as a skolem. But there are two problems 1. #378 does not explain this design decision in sufficient detail. This led Simon to question its inclusion: "Does it have any value, really? Why not just reject?" 2. #281 follows this design in one section, saying "treated as a fresh skolem constant"; and rejects it in another section, saying "any uses of terms in types are ill-typed". There is a contradiction. The amendment addresses both issues as follows: 1. A new example is added to #378, explaining how treating `x` as a skolem is useful, but only if `foreach` is available (a retained quantifier that allows dependent pattern matching) 2. Contradiction in #281 is resolved in favor of rejecting any uses of terms in types as ill-typed, saving this feature for a future proposal. I recommend to accept. Please share your thoughts either here or directly on GitHub. Vlad

7 7

#605: visible forall to work without ScopedTypeVariables, rec: accept
by Joachim Breitner 02 Sep '23

02 Sep '23

Dear committee, Vladislav Zavialov in https://github.com/ghc-proposals/ghc-proposals/pull/605 proposes a small fix to #281, namely that for f :: forall x -> Show x => x -> String f (type t) = show @t to work, ExplicitNamespaces and RequiredTypeArguments is sufficient, and we should not also require ScopedTypeVariables, which we are moving away from. I’ll shepherd this myself, and I suggest acceptance. Cheers, Joachim -- Joachim Breitner mail(a)joachim-breitner.de http://www.joachim-breitner.de/

5 5

Language Extension Policy – Round 3
by Arnaud Spiwack 30 Aug '23

30 Aug '23

Dear all, For this round, I'd like to build on the results of last round [1]. One thing that strikes me in the results, bearing in mind that 6 out of 10 members answered the question directly, and that none of this is without dissent anyway, is that there seems to be a general feeling that extension should not stay forever. What I'm reading is that most respondents seem to believe that when an extension becomes part of GHC20xx, it should be removed (with a deprecation period, I'd imagine): if you want the feature, use GHC20xx. Now, before I explore this question further, I should acknowledge that this doesn't directly help answering the questions that I set to resolve when this all began. Namely, under what principles do we judge extensions (in particular whether a proposal should be one of several extensions). But I think it's likely to help. I'm starting with this conversation because it's relatively concrete, probably less controversial than other questions, and as such, I'm hoping that it'll help us discuss the more difficult questions with a little more understanding of what we collectively believe. Let's explore what it would mean for an extension to be removed. - The extension wouldn't appear in `$ ghc --show-options` (currently `$ ghc --show-options | grep '\-X'` has 268 lines, almost every extension accounts for 2 lines) - The programmer can't turn the extension on or off directly - The documentation, in the user manual, of the corresponding feature would be moved from the “Language extension” section to somewhere else (maybe a new section “Language features”? Not sure what is natural here) - Error messages about the feature not being activated would stop suggesting you use the extension, but instead that you use one of the appropriate GHC20xx. - For some extension, we could provide a warning (off by default) to allow those who want to avoid the corresponding to disallow it in their codebase. But certainly not all extensions would be turned into warnings: I don't anticipate anybody will want to specifically avoid -XBinaryLiteral for instance. - Am I forgetting something? It doesn't mean that the extension must disappear in the implementation of GHC: this is an implementation detail (maybe it'll be beneficial for some extensions to be removed and other not, I don't feel capable of anticipating this). But the extension would not be visible or accessible to the user. Now the question is: ignoring the question of whether we have time to retroactively do this, does this look like this would be part of your ideal vision for GHC. If not, please argue. [1] https://docs.google.com/spreadsheets/d/1cRTJ2go5opXjIp-kojR_eOA-RWNBq2jzmec… -- Arnaud Spiwack Director, Research at https://moduscreate.com and https://tweag.io.

8 12

Please review #607: Amendments clarify treatment of term variables in types, Shepherd: Vlad
by Joachim Breitner 19 Aug '23

19 Aug '23

Dear Committee, Jakob Brünker spotted some inconsistencies in 281 (visible forall) and #378 (Design of DH): https://github.com/ghc-proposals/ghc-proposals/pull/607 I’d like to assign this to Vlad. Please guide us to a conclusion as outlined in https://github.com/ghc-proposals/ghc-proposals#committee-process Cheers, Joachim -- Joachim Breitner mail(a)joachim-breitner.de http://www.joachim-breitner.de/

1 0