haskell-infra / hackage-trustees

Issue tracker for Hackage maintainance and trustee operations
https://hackage.haskell.org/packages/trustees/
42 stars 7 forks source link

A number of HVR's packages need NMU's #297

Closed emilypi closed 3 years ago

emilypi commented 3 years ago

Hi all, the following packages need to be bumped to unblock a large number of packages. We've been trying to reach HVR for over 3 months and it seems life has taken over and he has been somewhat responsive (e.g. with cryptohash-sha256), but not been as responsive as we need in order to move things along.

We need to coordinate some NMU's. The following packages need help:

There may be more on to this list, but I'd like to get started planning and tracking for these ASAP per the NMU policy.


Progress:

Maintenance

Needed Revisions/Releases:

angerman commented 3 years ago
package ghc-9.0.1 hackage latest as of today
ghc 9.0.1 8.10.2
base 4.15.0.0 4.14.1.0
ghc-boot-th 9.0.1 8.10.2
ghc-boot 9.0.1 8.8.3
ghc-prim 0.7.0 0.6.1
ghci 9.0.1 8.10.2
hpc 0.6.1.0 0.6.0.3
integer-gmp 1.1 1.0.3.0
libiserv 9.0.1 -
template-haskell 2.17.0.0 2.16.0.0
bergmark commented 3 years ago

I saw that Oleg asked to take over maintainership of uuid*, so i'd say we should let that play out. Perhaps hackage admins can expedite it? Would it be possible to find maintainers for the regex packages as well?

I have been toying with the proposal of revising/dropping NMUs as a trustee thing as it requires a lot more effort than other things we do. i.e. it would at least feel like less work for me to personally take over maintainership and do the uploads.

emilypi commented 3 years ago

Great news on uuid, uuid-types, and hashable, thanks @phadej! HVR poked his head out this morning to mention that he updated the bounds on text-short, so that one is solved for now. We'll continue discussing plans to co-maintain in private. This leaves us with the regex libraries. Personally i'm fine if @bergmark or some combination of @bergmark + others here take on co-maintenance. I'm happy to help move things along as well.

sjakobi commented 3 years ago

This leaves us with the regex libraries. Personally i'm fine if @bergmark or some combination of @bergmark + others here take on co-maintenance.

@andreasabel offered to help as a co-maintainer in https://github.com/haskell-hvr/regex-base/issues/9.

emilypi commented 3 years ago

Ah! right, yes @andreasabel as well.

Bodigrim commented 3 years ago

Couple more:

@bergmark could you possibly make revisions for these?

emilypi commented 3 years ago

Good news:

HVR popped up and made @phadej, @andreasabel and I maintainers of regex-*, and @phadej in particular a maintainer of uuid and uuid-types. We'll have things unblocked shortly, and then we can address any lingering revisions.

andreasabel commented 3 years ago

I get to work on regex-*; @hvr invited me into the team.

bergmark commented 3 years ago

Great stuff. I think we can close this then?

andreasabel commented 3 years ago

Er, I also need upload permissions on hackage for the regex-* family.

andreasabel commented 3 years ago

@bergmark Please keep this open until the hackage permissions have also been granted.

bergmark commented 3 years ago

@Bodigrim sorry i missed your request for hslogger & cassava! I've revised hslogger. For cassava, i may be a bit outdated on how cabal-install deals with flags now but I'm not able to get it compiling, getting

error: **INVARIANT BROKEN** Detected invalid combination of `text-short` and `bytestring` versions. Please verify the `pre-bytestring-0.10-4` flag-logic in the .cabal file wasn't elided.

It works with stack, but I'm afraid adding the revision could cause compilation issues for cabal-install.

Bodigrim commented 3 years ago

For cassava, i may be a bit outdated on how cabal-install deals with flags now but I'm not able to get it compiling

I think your package index is outdated and does not include recently revised text-short-0.1.3.

  if flag(bytestring--LT-0_10_4)
    build-depends: bytestring <  0.10.4
                 , bytestring-builder >= 0.10.8 && < 0.11
  else
    build-depends: bytestring >= 0.10.4
                 , text-short == 0.1.*

I presume you've been running something like cabal build --constraint 'bytestring >= 0.11' --allow-newer 'cassava:bytestring'. If there is no text-short available to satisfy bytestring-0.11 (which was the case before text-short-0.1.3 got revised), in order to find a build plan cabal is forced to enable bytestring--LT-0_10_4 - with catastrophic consequences.

@bergmark could you please cabal update and give it one more go?

bergmark commented 3 years ago

That wasn't it, still not exactly sure what is going on but i managed to solve it by using local versions of regex-base and regex-posix with patched bytestring upper bounds! So it will be fine once they are updated as well. http://hackage.haskell.org/package/cassava-0.5.2.0/revisions/

emilypi commented 3 years ago

@bergmark good to hear.

Just an update: @Bodigrim and I have been granted co-maintainership of Cabal, cabal-install, zlib, and tar as of today. An RC is out for zlib, and we'll have one out shortly for regex-* as needed, once HVR approves of the PR prep.

andreasabel commented 3 years ago

I released regex-base-0.94.0.1. Getting to work on regex-posix.

andreasabel commented 3 years ago

regex-posix got its revision: https://hackage.haskell.org/package/regex-posix-0.96.0.0/revisions/

emilypi commented 3 years ago

If we don't see a review from @phadej or @hvr on the existing zlib-0.6.2.3-rc in the next 2-3 days, I'll just do a release if everyone's okay with that.

andreasabel commented 3 years ago

regex-tdfa has also been revised

emilypi commented 3 years ago

zlib-0.6.2.3 is now live!

We did it! Thank you for the hustle everyone. Hopefully we are now in a place where this doesn't happen again because support maintainers are now in place for the relevant packages. Thank you to everyone who helped us get things moving again!

@bergmark now we can close the issue 😄

bergmark commented 3 years ago

great!

phadej commented 3 years ago

Can cassava's base bound be relaxed as well?

bergmark commented 3 years ago

@phadej i did http://hackage.haskell.org/package/cassava-0.5.2.0/revisions/ previously, did i miss something?

phadej commented 3 years ago

I don't see base in your revision. only bytestring and QuickCheck.

phadej commented 3 years ago

I made revision for cassava-0.5.2.0 allowing base-4.15: https://hackage.haskell.org/package/cassava-0.5.2.0/revisions/