unbound 1.11.0

[miccal]

Created with brew bump-formula-pr.

[miccal]

The only failure for 10.14 and 10.13 was for agda, which is to be expected.

The tests for 10.15 appear to have stalled.

[Rylan12]

The agda issue should be fine as it doesn't build from source anyway at the moment.

We do need the 10.15 checks to run, though, in order to get updated bottles. The 10.15 runners have been having some trouble recently for some reason, but rerunning them usually does it. I'll retrigger it. If it doesn't work this time, try rebasing and force-pushing.

[miccal]

Thank you @Rylan12, as always.

[SMillerDev]

The agda issue should be fine as it doesn't build from source anyway at the moment.

Is there an upstream issue for this?

[miccal]

On 10.15 and 10.13:

Error: 2 failed steps!
brew test --verbose agda
brew test --verbose csound

while on 10.14 only adga failed.

csound is a known random failure.

So ok to merge I think.

Thanks.

[miccal]

Is there an upstream issue for this?

@SMillerDev I believe it is because of the Cabal v1 install deprecation, and is therefore not an upstream issue?

[SMillerDev]

In that case someone should make a pull request to fix it.

[Rylan12]

It sounds like @Bo98 is working on the agda issue. I'm not sure if it's: 1. going well or 2. will be done soon

@SMillerDev I understand not wanting to merge things that have had CI failures, but I think holding this up isn't doing anything positive. agda won't build regardless of this PR and I don't think merging this will make fixing agda it any easier/harder. Not merging only means we are holding back a new version of unbound for users (plus the ~171 formulae that depend on it in some way)

Are there other consequences I'm missing?

[SMillerDev]

I'm sure it won't hurt the progress there. But keeping this for another time like this isn't helping anything either. I think we need to be stricter in this and as a policy either have a pull request or an issue open. Currently we keep on rediscovering issues and it's hurting progress like here.

TL;DR we need a PR or issue for agda if we're not going to fix it here.

[Rylan12]

I think that's a reasonable requirements to stick to. In this case, though, there is an issue open (#55253).

[SMillerDev]

#55253 isn't really an issue about agda being broken though. That one is meant to guide the migration.

[Rylan12]

But the reason agda is broken is that cabal v1 is deprecated, so I think it's implied that all formulae in that issue will be broken.

Maybe that's not the case. If it's not, then there definitely should be an issue or PR for it.

[Bo98]

1. going well

no

2. will be done soon

A quick temporary fix is to make it use "v1-install" directly (and no sandbox).

"v2-install" requires either the MacPorts-style hack or wrapper scripts for every binary, both of which should have some brew helpers to accomodate.

A hybrid approach of v2-installing dependencies and not using Cabal for the actual agda build is another possibility I'm investigating.

I planned to discuss Cabal a bit at the next maintainer call, but I should probably not wait until then and figure something out this week. I should have dealt with it weeks ago since it's my fault really that the test fails.

[miccal]

Agda issue reported in #58984

[miccal]

@Rylan12, @SMillerDev and @Bo98 I rebased and the only failure now is agda.

Thanks.

[BrewTestBot]

@Rylan12 has triggered a merge.

[Rylan12]

Thanks, @miccal!