vector Jan 2020 RC

[cartazio]

@gwils @Boarders @fumieval @RyanGlScott @chessai @hvr @andrewthad
i'm thinking https://travis-ci.org/haskell/vector/builds/633862783 (current master) should work as a minor release bump on the 0.12 series, or a reasonably stable 0.13 series,

please provide feedback on if thers any outstanding tickets or PRS you'd like me to make sure to include (ideally with GH flavored markdown checkboxes for the things you want so i can keep track and get it over with today :) )

[andrewthad]

These are both pretty straightforward PRs that I would like to see merged:

• Add function for converting Array to Vector #209 (fromArray)
• Add mapMaybeM and imapMaybeM #226 (monadic map maybe)

[fumieval]

As long as Y or Z in version X.Y.Z is bumped, I'm fine either way.

[cartazio]

Atm I think the changes in master reflect a minor bump aside from perhaps the roles annotation for storable.

…

On Tue, Jan 7, 2020 at 7:25 PM Fumiaki Kinoshita ***@***.***> wrote: As long as Y or Z in version X.Y.Z is bumped, I'm fine either way. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#266?email_source=notifications&email_token=AAABBQWAONTCWRECRYA2OTDQ4UMRHA5CNFSM4KD4LE62YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEIKYJCY#issuecomment-571835531>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAABBQSVXMQDL2OVLBIEYETQ4UMRHANCNFSM4KD4LE6Q> .

[cartazio]

@hvr and Ryan: care to way in on pvp Rec ? On Tue, Jan 7, 2020 at 9:41 PM Carter Schonwald <carter.schonwald@gmail.com> wrote:

…

Atm I think the changes in master reflect a minor bump aside from perhaps the roles annotation for storable. On Tue, Jan 7, 2020 at 7:25 PM Fumiaki Kinoshita ***@***.***> wrote: > As long as Y or Z in version X.Y.Z is bumped, I'm fine either way. > > — > You are receiving this because you authored the thread. > Reply to this email directly, view it on GitHub > <#266?email_source=notifications&email_token=AAABBQWAONTCWRECRYA2OTDQ4UMRHA5CNFSM4KD4LE62YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEIKYJCY#issuecomment-571835531>, > or unsubscribe > <https://github.com/notifications/unsubscribe-auth/AAABBQSVXMQDL2OVLBIEYETQ4UMRHANCNFSM4KD4LE6Q> > . >

[sjakobi]

It would be nice if this optimization could be included, but it shouldn't block a release:

• Compare vector sizes in eq[By] #261: Optimize eq[By] and Eq instances by comparing sizes first

[lehins]

Per request from @cartazio: issues related to current RC that need to been addressed:

• This repo is at 0.12.0.1, while hackage has 0.12.0.3 #255 - prior to making a new release a git tag for current v0.12.0.3 needs to be created (that actually matches the released version)
• Build broken with latest primitive #256 - make sure new release builds with latest version of primitive
• vector storable () instance broken in current releases. #258 - ensure No more error: division by zero when trying to print or operate on empty storable vectors #168 is added to new release
• No Unbox instances for newtypes in Data.Monoid #264 - ensure Add unboxed instances for all remaining base newtypes #194 is added to new release

[cartazio]

@lehins I think master should build now with primitive 0.7 :), i cherry picked the commit needed

[cartazio]

also i think that v0.12.0.3 is live

[cartazio]

i'm going to wait a few days to do anything more intricate till after i'm done with my fever,

[enobayram]

Per request from @cartazio: It seems like people are occasionally getting bitten by the cryptic boxed Vector + compaction issue (#220), spending a lot of time debugging (I'm one of the victims). This PR adds a hint to the ensuing error message:

• Ensure Improve "uninitialised element" error message #231 is added to the new release

[cartazio]

#268 is gonna take priority! :)

[cartazio]

i'm probably gonna have to backport the fix to previous minor/major versions too

[cartazio]

• Segfault when using enumFromTo with maxBound as upper bound #188
  and
• Segfaults and "internal errors" when vector slice overflows  #257
  and
• enumFromTo fails QC at -O2 #268

probably also should figure out how to use #229 and also make sure current master is green on full build matrix of ghc versions at O2/O1

[RyanGlScott]

I'd like to see commits 3d742ab and 4b1ba83—which bump the upper version bounds on base and ghc-prim to <4.15 and <0.7, respectively—on Hackage soon, given that GHC 8.10.1-rc1 is out (8.10.1-rc1 bundles base-4.14.0.0 and ghc-prim-0.6.1). I'm not picky about whether this is done via a Hackage revision or a new release.