[MustardTiger]

>Yet you expect other people, many of them volunteers, to take up the busywork of bounds maintenance, for free, so that your immense builds work.

I have no idea how you came up with that absurd non-sequitur.

>Yesod is not broken. It builds just fine using a sane build tool. Yesod does not become broken because you insist on using a tool that crafts arbitrary build plans.

Nothing in that is accurate at all. There is nothing arbitrary about a cabal build plan.

[dllthomas]

> > Yet you expect other people, many of them volunteers, to take up the busywork of bounds maintenance, for free, so that your immense builds work.

> I have no idea how you came up with that absurd non-sequitur.

You can dispute the accuracy, but "the PVP as written puts too much burden on maintainers" was a part of the justification at the time for removing upper bounds. See, for instance, the following from someone with no connection to FP Complete that I'm aware of: https://mail.haskell.org/pipermail/haskell-cafe/2012-August/...

"As someone who recurrently is nudging a large number of maintainers every major ghc release to bump their bounds, I favor the no upper bounds approach!"

It was not a non-sequitur, but an objection to your assertion that there was no problem with the PVP.

[hvr_]

> See, for instance, the following from someone with no connection to FP Complete that I'm aware of

You may be interested to know that Carter (the guy you're quoting) knows better today, and has even become a Hackage Trustee whose mission is more or less to uphold those very PVP upper bounds ;-)

[dllthomas]

Amusing, but orthogonal to my point, which was that the move away from parts of the PVP was motivated by real concerns, and moreover that the comment above was specifically an obvious reference to concerns that had been voiced during that time.

I am not arguing that the move away from the PVP was correct - I was uncertain at the time and I remain uncertain.

[mightybyte]

The concerns you're talking about here were primarily deficiencies in tooling, not an inherent problem with the PVP. Many of those deficiencies have been fixed (for example, --allow-newer) or there is work in progress that will fix them (cabal new-build, cabal.project files, etc).

[dllthomas]

That may very well be the case. Again, I'm not arguing about the PVP. I'm arguing that discussion of those concerns is not non sequitur.

[MustardTiger]

> See, for instance, the following from someone with no connection to FP Complete that I'm aware of: https://mail.haskell.org/pipermail/haskell-cafe/2012-August/....

Yes, people make mistakes. He no longer favors that approach, as he learned how bad it is. The difference is that FPC employees still continue to push it even when they know the problems.

[dllthomas]

You seem to miss my point. I am not arguing that the move away from upper bounds was correct - I was uncertain at the time and I remain uncertain.

My point was that the comment above was specifically an obvious reference to concerns that had been voiced during that time (and speculation as to why they might have been less relevant to you) - far from an "absurd non-sequitur."

[MustardTiger]

>My point was that the comment above was specifically an obvious reference to concerns that had been voiced during that time

I am aware of that. And again, that in no way contradicts or discredits what I said. Which was that FP complete employees consistently do this. I never said they were the only people who do it. I never said they were the first people to do it. I said they do it. I really do not understand where the confusion is coming from.

[dllthomas]

You keep acting as if the only claim you've made in this entire discussion was that FP Complete employees do not include upper bounds. It's certainly not the case that anything we've been discussing has bearing on that. But that's not an honest characterization of your comments in this thread.

You said (https://news.ycombinator.com/item?id=12058419):

> The tool does not fix real problems. It avoids a real problem, which the authors of the tool created in the first place. The problem does not exist if packages use upper version bounds.

The implicit assertion in the following comments (which I found painfully clear, but maybe you genuinely missed it - surely we read from different contexts) was: Even if the tooling only solves a problem introduced by particular practices, it can be a "real problem" if those practices address a real problem themselves.

[MustardTiger]

You are responding to a strawman, and still trying to pretend it has something to do with me. What I said is what I said, and is what I meant. What you wish I meant has nothing to do with me.

[massysett]

Since Yesod builds just fine, explain how it is "broken"?

[HaskellMuppet]

The fallacy of that argument is assuming that tool X being able to build artifact Y means that X is "sane". The only thing this says is that tool X is able to build Y, nothing more and nothing less. For instance, I could easily implement a tool which is only capable to build `yesod`, and only `yesod`. Would that be a "sane" build-tool?