Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Release v2.0 #2063

Closed
MatthewDaggitt opened this issue Aug 17, 2023 · 41 comments
Closed

Release v2.0 #2063

MatthewDaggitt opened this issue Aug 17, 2023 · 41 comments
Labels
release
Milestone
v2.0

Comments

@MatthewDaggitt
Copy link
Contributor

Okay, so I realise I've been somewhat absent recently. My current job ends on September 12th and my next one doesn't start until late November, so I plan to get v2.0 out the door during that period of time. I promise I will wade through the back log of PRs that have built up 👍
@MatthewDaggitt MatthewDaggitt added this to the v2.0 milestone Aug 17, 2023
@jamesmckinna
Copy link
Contributor

Congratulations on the new job! 🎉

@jamesmckinna
Copy link
Contributor

jamesmckinna commented Sep 13, 2023 •
edited
Loading

As to managing the backlog, as both a contributor to making the pile grow, as well as more recently, helping to whittle it down (at least a little bit), I was interested to read that our colleagues on the main Agda issue tracker are correspondingly concerned about their own policies to manage their PR backlog.

It might be worth a separate issue, but rather than generate more clutter for the time being, I'm linking it here while our efforts are directed towards the v2.0 backlog... once that's out, suggest we (re)start the conversation about how to manage the way forward.

@MatthewDaggitt
Copy link
Contributor Author

Okay, with the merge of #1895 I'm going to get serious about making a release.

As of now, all new PRs will get a v2.1 milestone unless they explicitly address an issue that needs to be fixed before v2.0.

Existing open PRs will be triaged depending on how easy they are to merge.

When would interested people be available for a meeting next week to discuss outstanding PRs? If you are, please state your availability in GMT below:

Matthew

  • Monday 2nd - 7am - 4pm
  • Tuesday 3rd - 7am - 11am, 2pm-4pm
  • Wednesday 4th - 7am - 4pm
  • Thursday 5th - 7am - 4pm
  • Friday 6th - 7am - 11am

@Taneb
Copy link
Member

Taneb commented Sep 26, 2023 •
edited
Loading

My availabilities next week:

  • Monday 6-7am, 7:30-9am, 10:30am-late
  • Tuesday 6-7am, 8am-12pm, 2pm-late
  • Wednesday 6-7am, 7:30am-late
  • Thursday 6-7am, 7:30am-12pm, 2pm-late
  • Friday 6-7am, 7:30am-12pm, 1:30pm-late

Note that we're in summer time! The UK is one hour ahead of GMT, much of Europe is two hours ahead.

@jamesmckinna
Copy link
Contributor

jamesmckinna commented Sep 26, 2023 •
edited
Loading

This is great news!

As for a meeting next week, I'm a bit all over the place at the moment, as evidenced by the manic PR/issue/commentary explosion. Will try to identify times to meet soon!

I've been working on 'one last' (sic) branch for the AlmostCancellative issues (the corresponding ones for Monotonic can wait, but it might be good to try to knock those off as well), so I hope we can merge that/them. But: up for discussion ;-)

Ahead of such discussion, I've been making 'bids' against my own open issues and PRs (and some others) by tagging them with v2.0 (wish-to-fix-now), v2.1 (can wait) and v3.0 (I'd like it one day). Suggest others do same?

@JacquesCarette
Copy link
Contributor

Next week is tricky for me: the good part, I'll be in Europe, so time zones less of an issue; less good from an availability perspective: I'll be in Dagstuhl. So while it is quite likely that I could get out for an hour, knowing when that'll be is essentially impossible to know now.

So: schedule it, and I'll see.

@jamesmckinna
Copy link
Contributor

I'll be largely free-but-frazzled next week, so let me know when works best for you and @Taneb , and let's see how we do...

@MatthewDaggitt
Copy link
Contributor Author

Okay, let's go for Wednesday 7:30am GMT then. Whoever is around is welcome to join.

@MatthewDaggitt
Copy link
Contributor Author

Meeting link here: https://meet.google.com/trh-fpbj-kfi

@jamesmckinna
Copy link
Contributor

Not really a separate issue, but to note here: before release, we should nitpick the CHANGELOG to satisfy ourselves it makes global as well as local sense (orphaned/duplicated comments; list of deprecations/additions getting out of alphabetical/file contents order, spellcheck, fix-whitespace, ...)

@Taneb
Copy link
Member

Taneb commented Oct 4, 2023

https://time.is/730_4_Oct_2023_in_UTC

@MatthewDaggitt
Copy link
Contributor Author

Not really a separate issue, but to note here: before release, we should nitpick the CHANGELOG to satisfy ourselves it makes global as well as local sense (orphaned/duplicated comments; list of deprecations/additions getting out of alphabetical/file contents order, spellcheck, fix-whitespace, ...)

Yup, that's always necessary before a release! I usually do it as the last possible thing though, otherwise you get horrendous merge conflicts with open PRs.

@jamesmckinna jamesmckinna mentioned this issue Oct 4, 2023
Open
@MatthewDaggitt
Copy link
Contributor Author

MatthewDaggitt commented Oct 12, 2023 •
edited
Loading

There is another meeting today at https://time.is/1200_12_Oct_2023_in_UTC. Same room link posted above. @laMudri would you be available to join briefly to discuss #2054 and another related issue?

@laMudri
Copy link
Contributor

laMudri commented Oct 12, 2023

Should be able to, yeah. 👍

@MatthewDaggitt
Copy link
Contributor Author

Still 20 open issues and PR for v2.0 to go.

@jamesmckinna
Copy link
Contributor

Have all the decisions we made this week and last been carried across to the milestone, specifically those triage decisions to move things to v2.1/v3.0, eg IIRC

  • $\mapsto$ v2.1
  • $\mapsto$ v3.0

That still leaves a chunk of work to do, but maybe the end is then in sight?

@MatthewDaggitt
Copy link
Contributor Author

#1579 ? (it would be nice to have got this done in time, but ... related to #1436

This has to be bumped to 3.0 as again it's breaking. But you're right, we don't have the energy to do it now.

#2098 (unless you are committed to getting it done now?)

This can't be bumped to v2.1 as it's going to be massive breaking change. So yes, relatively committed, and the StrictTotalOrder refactor was in preparation for it.

#759 this seems to be a recurrent one, mostly for Matthew's shopping list which perhaps needs updating to show the current state of this issue and the associated PRs?

So this one, I'm going through deprecating each individual definition as well, as they've been scattered across many different modules. It also helps to check that we've actually got everything in the new hierarchy.

The rest are bumped as you suggested.

@MatthewDaggitt
Copy link
Contributor Author

Down to 12 open issues/PRs....

@MatthewDaggitt
Copy link
Contributor Author

Every open v2.0 issue now has a PR open fixing it. If everyone could do a quick review of one or two of the PRs each, then we can hopefully get a candidate release out by the end of the week.

@MatthewDaggitt
Copy link
Contributor Author

Release candidate 1 created and announced. Waiting for the flood of issues in 3... 2... 1...

Other than that, thank you to everyone who has helped get it over line!

@jamesmckinna
Copy link
Contributor

Incoming spellcheck for CHANGELOG...

@shhyou
Copy link
Contributor

shhyou commented Oct 20, 2023 •
edited
Loading

Is it possible to resolve #1743 so users can open README[/＊].agda files? It could be changing how README[/＊].agda modules are packaged in the release tarball or adding instructions to installation guide, whichever is more convenient.

@MatthewDaggitt
Copy link
Contributor Author

Is it possible to resolve #1743 so users can open README[/＊].agda files? It could be changing how README[/＊].agda modules are packaged in the release tarball or adding instructions to installation guide, whichever is more convenient.

Okay, will do our best.

@MatthewDaggitt
Copy link
Contributor Author

Would some subset of people be free for a meeting on Tuesday or Thursday this week? To discuss issues that have popped up with v2.0 and I guess moving forwards!

@jamesmckinna
Copy link
Contributor

Either works for me.

@JacquesCarette
Copy link
Contributor

Both can work for me.

@jamesmckinna
Copy link
Contributor

Apologies, sort of, for having added a couple of purely bureaucratic PRs #2205 , #2206 to the pile.
But hopefully they are/will be easy to review/merge..

@MatthewDaggitt
Copy link
Contributor Author

Okay every open issue with v2.0 on it has a corresponding open PR. So if people could take a look and try and review them we can make the next RC by the start of next week hopefully.

@jamesmckinna
Copy link
Contributor

jamesmckinna commented Nov 25, 2023 •
edited
Loading

Regarding the release: what becomes of the the v1.* deprecation warnings everywhere in the library?
Should we be systematically removing them now, given the massive step-change to v2.0?
What's the policy generally about such things?

@MatthewDaggitt
Copy link
Contributor Author

So as part of v2.0 I removed all deprecated content before v1.2.

I don't think we want to remove all v1.* deprecated content as some of it was deprecated relatively recently... but I have no formal policy. I chose v1.2 as it got rid of a lot of empty deprecated modules.

@MatthewDaggitt
Copy link
Contributor Author

Announced release candidate 2.

@jamesmckinna
Copy link
Contributor

jamesmckinna commented Nov 27, 2023 •
edited
Loading

Sigh... our first new issue #2216 ... :-( But many thanks for the speedy response in resolving it!

@jamesmckinna jamesmckinna mentioned this issue Dec 5, 2023
Merged
@jamesmckinna
Copy link
Contributor

jamesmckinna commented Dec 11, 2023 •
edited
Loading

Should be good to go now!
UPDATED But maybe missing yesterday's window meant you were too snowed under to make the final release/announcement?

@MatthewDaggitt
Copy link
Contributor Author

Okay announced the release. I am however struggling to follow the instructions to generate the documentation.

In particular:

    cp .github/tooling/* .
    runhaskell GenerateEverything.hs
    ./index.sh
    agda -i. -idoc -isrc --html index.agda
    mv html v$VERSION
    git checkout gh-pages
    git add v$VERSION/*.html v$VERSION/*.css
    git commit -m "[ release ] doc for version $VERSION"
    git push

on the runhaskell step I'm getting

GenerateEverything.hs:19:1: error:
    Could not find module ‘System.FilePath.Find’
    Perhaps you meant
      System.FilePath.Windows (from filepath-1.4.2.2)
      System.FilePath.Windows
      System.FilePath (from filepath-1.4.2.2)
    Use -v (or `:set -v` in ghci) to see a list of the files searched for.
   |
19 | import System.FilePath.Find
   | ^^^^^^^^^^^^^^^^^^^^^^^^^^^

I've run cabal install in the directory but still no luck. @jamesmckinna did you come across this issue last time?

@Taneb
Copy link
Member

Taneb commented Dec 12, 2023

That module looks like it should come from the filemanip package.

@MatthewDaggitt
Copy link
Contributor Author

Yup, but shouldn't cabal install get it for me? I'm not even sure how to go about installing random packages on my system from Hackage, outside of cabal and stack...

@Taneb
Copy link
Member

Taneb commented Dec 12, 2023

Can you do cabal run GenerateEverything instead? That should handle dependencies properly

@Taneb
Copy link
Member

Taneb commented Dec 12, 2023

I have the suspicion that what your cabal install is doing is building GenerateEverything and AllNonAsciiChars, and then installing the executables, keeping the library dependencies hidden from GHC

@jamesmckinna
Copy link
Contributor

jamesmckinna commented Dec 12, 2023 •
edited
Loading

No, I don't seem to have come across this before.

Indeed, I recently upgraded on one machine finally to 2.6.4, and IIRC, that installed filemanip for me...? (And I don't recall having had these problems with the old release versions I'd rebuilt a while back... although I find it hard to keep such things at the front of my memory alas, so thanks, if nothing else, for the jog ... ;-))

Obvious addition: run cabal update and then ask for the latest cabal as well. Again, IIRC, upgrading my ghc with ghcup also may have helped with all of the above, esp. wrt identifying 'stable'/'recommended' versions of packages.

All of the above aside, I am a total n00b when it comes to these things... but I'll (try to) check when I'm back home tonight.

@jamesmckinna
Copy link
Contributor

jamesmckinna commented Dec 12, 2023 •
edited
Loading

Also: #2226 I'm not currently at a machine where I can easily fix this...
... UPDATED I've added this to the merge queue for master; not clear how to do this for v2.0-release?

@MatthewDaggitt
Copy link
Contributor Author

Okay, great. I've managed to upload the documentation. Thanks for your help @Taneb!

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
release
Projects
None yet
Milestone
v2.0
Development

No branches or pull requests
6 participants
@JacquesCarette @Taneb @shhyou @MatthewDaggitt @laMudri @jamesmckinna