Boston, MA
Joined January 2009
I recently learned my friend makes $800k per year doing the same thing I do, with the same number of years of experience, but in the Bay Area, at Meta. I make $175k per year. Meanwhile others make $50k per year working full time at the coffee shop I go to. This earth is fucked.
1
151
neovim: awesome software, but the maintainers have terrible taste when it comes to backwards compat every release brings spurious renames and random api changes, so upgrading neovim and plugins is always a painful weekend ordeal please just leave your exposed API alone...
1
244
No one cares, but I finally blocked @dysinger. So much gun/Murica content, so little Haskell.
2
Every morning I make my bed, brush my teeth, pull up my pants, type OverloadedRecordDot NoFieldSelectors DuplicateRecordFields NamedFieldPuns OverloadedLabels OverloadedRecordUpdate into my cabal file, and restart all my tooling
1
7
There's got to be a better solution to maintaining accurate version bounds of Haskell dependencies than Hackage Trustees volunteering to act on behalf of unresponsive maintainers by diverging from the source repository by mutating the metadata directly on Hackage.
1
1
It means the maintainer has to wake up one day and backport all of the mutations into their source repo, or else be making changes to a version of the source repo that is still missing known good dependency bound adjustments.
2
Frankly the whole system just seems broken, and I don't really enjoy trying to keep it afloat anywag by occasionally submitting PRs that adjust bounds, which get closed in favor of making revisions on Hackage instead.
Good advice. I think it applies to individual contributors as well.
Tiny management tip: say "our team" instead of "my team"
1
4
I appreciate @simonpj0's dedication to specifying whether his commits contain more than one wibble
3
13
2
150
Haskell performance pro tip for (most) concurrent programs: don't do anything in the main thread. Just forkIO the real main and wait until it's done.
2
3
17
mitchellsalad retweeted
I ported a bunch of code over to @mitchellsalad's Ki concurrency library last night. Really nice! Bidirectional exception propagation is the right default, imo.
2
2
11
ki 1.0.0, a structured concurrency library for Haskell, hot out of the oven! hackage: hackage.haskell.org/package/… github: github.com/awkward-squad/ki reddit announcement: reddit.com/r/haskell/comment… discourse announcement: discourse.haskell.org/t/ki-1…
1
12
48
alright let's get the minimum element of an IntSet, oh it's a partial function
This country is an embarrassment
Now do Plessy vs Ferguson/Brown vs Board of Education.
1
Just glanced at my cabal freeze file and noticed a few lines like 'mtl ==2.2.2 || == 2.3'. Why is it putting version ranges in there? I'm trying to freeze my dependencies so I don't get build failures in six months. Any way to make cabal freeze not put '||'?
1
1
🥳
The @unisonweb team got code hosting working. It's amazing. Now you can push code + docs directly to Unison Share and it's instantly there, fully hyperlinked. Here's me popping open a def from trunk of distributed library! @aryairani @shojberg @chrislpenner @mitchellsalad
1
This morning I was clicking through some old versions of a package I maintain, trying to decide which ones to mark as deprecated, and I was surprised to see how many revisions were made by Hackage trustees. No one ever contacted me about a dependency bound issue. 1/3
1
3
They just silently mutated my cabal file, far away from what I wish was the source of truth - the git repository. I now see my attempts to document all updates in my changelog, including revisions, have been incomplete. 2/3
1
1
I suppose I can try putting "x-curation: uncurated" in there and see if any trustees even bother to respect it. But I've already been burned by a breach of protocol - no communication. I think we need a better solution to tweaking known-good dependency bounds altogether.
1