Guess who's an approved uploader on #hackage now?

We've fixed a bug (https://github.com/byorgey/MonadRandom/issues/53) in the MonadRandom package (https://hackage.haskell.org/package/MonadRandom) which could have *extremely rarely* caused it to crash when choosing from a weighted distribution with certain adversarial weights. How rare, you ask? So rare that I would be flabbergasted if it has ever actually happened. Still, fixing it was the right thing to do on principle.

The fix preserves the exact same behavior in almost all cases (i.e. exactly the same sequences of random numbers will still be generated given the same starting seeds) --- except that, again, *extremely rarely* it could generate a different number than before (even in some cases where it would not have crashed).

So, should this be a minor or major version bump? Discuss!

#haskell #hackage #random #version

For Haskell packages published on Hackage, do you prefer version bounds on dependencies to be narrower or wider?

narrow ~ “We have tested that the package builds with dependencies that fall in this range”

wide ~ “Things are very likely to break if you use a version outside of this range, but we haven’t necessarily tested every version in it”

Case in point, I’m considering for agda2hs whether to bump the bound on base to only include the versions we actually test on CI, or to also include older versions that might still work but aren’t currently tested.

#Haskell #hackage #packages #agda

@feld good question!

i think we automatically package all #haskell packages from #hackage and some #stackage (18k) and all #Rpackages from #CRAN (27k)

the manually packaged python and perl libs are available for multiple versions. the supported ones are visible, older still work

check out the package sets on the left https://search.nixos.org/packages?channel=unstable&from=0&size=50&sort=relevance&type=packages&query=*

that blows up the number! we don't have ancient, unmaintained packages found in debian, but modern ones like in #AUR

#nixos #nix #packaging

@gregorni

Last week I submitted a new version of my currycarbon #haskell library to #hackage (see https://archaeo.social/@ClemensSchmid/111517617941502770). I made a mistake in the setup of my golden tests, which prevented the package from building on #stackage. @alaendle observed the issue, contacted me, explained the problem and answered my questions on GitHub. It was a very kind and professional exchange (see https://github.com/nevrome/currycarbon/issues/17).

I was astonished how well this was handled and wanted to share the experience 💗

Booyah I’m able to upload #haskell packages to #hackage

it took me a long time to notice that #hackage is just #haskell + #package "portemanteau" 🤦

Just taking a nice chunk of time to be laid back and read through some quality docs on #haskell #hackage. Specifically reading about all the handy features of System.FilePath I was missing out on.

Recently I released my old #utility I personally was using for last few years called “place-cursor-at” written in #Haskell (built on top of #X11 #Xlib). It helps to move your mouse cursor around your screens to specific approximate positions. Generally it helps to use the keyboard more and less the mouse.

It was released on #Hackage and #Stackage. And now it’s also available in #Nix and #NixOS 20.09. Give it a try:

nix-shell -p place-cursor-at --run place-cursor-at

https://github.com/unclechu/place-cursor-at

"In this post, I demonstrate that critical parts of the #Haskell package management system are vulnerable to the #DependencyConfusion supply chain attack." #security #cabal #hackage

https://frasertweedale.github.io/blog-fp/posts/2021-02-12-haskell-dependency-confusion.html

So here's a frustration: #haskell module and package names collide. Apparently one can publish a package onto #hackage called Control.Lens which seems to not be part of a package named Control. SMH

By Odin, it worked! Installing #GHC 8.8.3 for #haskell means it was possible to successfully install the #Cabal package manager, which means the #HDBC and #HDBCOdbc packages could get downloaded from #Hackage, and this time #GHCI did not complain about a missing libHSold-locale library! wOOt!

@AMDmi3 I don't know. I am not using #Nix on a daily basis. Nix is only a package manager, so once it builds and installs, it is good to go (more or less). #nixpkgs is going to be involve a lot of actual work but that is going to happen independently from the #FreeBSD ports.

I was wondering if Nix could use FreeBSD ports as a source of information on how to build software on FreeBSD (I've heard that it has such possibilities for #Hackage but I have to read more about it).

These are my released #Haskell #packages:
https://hackage.haskell.org/user/ClaudeHeilandAllen

These now work with #GHC 8.8.1:
- bitwise
- bowntz
- butterflies
- complex-generic
- gearbox
- gl-capture
- hgmp
- long-double
- memo-sqlite
- pnm
- qd
- qd-vec
- ruff
- snowglobe

These don't work with GHC 8.8.1 yet (mostly because their dependencies haven't been updated yet):
- exp-extended
- gruff
- gulcii
- hp2pretty
- prof2pretty
- rounded
- wedged

Unreleased #et is blocked on `rounded`, which is blocked on `singletons` seeming to need `Cabal-2.4.1.0` which does not compile with `ghc-8.8.1` due to `Monad.fail` being removed.

The rest of the packages in the #hackage link's listing are mostly #abandonware, let me know if you want to #takeover #maintainership.

Spent an hour fighting #cabal #sandbox and got my #haskell #bytebeat #genetic #algorithm project to work with the current versions of things on #hackage.

It was a lot easier once I realized I no longer needed #gtkglext which I had already replaced with #glfw in the main code base

Still not packaged up nicely, not sure how to distribute code that has Haskell as well as #C executables, and runtime dependencies on GCC. So far it's just a Makefile.

https://code.mathr.co.uk/bitbreeder