otini @otini@functional.cafe

Bonjour,

Selon #FramaC, ma recherche dichotomique est correcte.

🤓

Do you know a good place to ask about fonts and typesetting, maybe specifically OpenType, maybe even more specifically OpenType Math fonts?

Planned vs budgeted — MonkeyUser

https://www.monkeyuser.com/2020/planned-vs-budgeted/

On ne va pas s'en cacher, on s'inquiète : aujourd'hui, 19 février, on a réuni environ 20% de notre budget pour 2020... Aidez-nous à relancer notre campagne de dons !

https://www.laquadrature.net/2020/02/19/la-quadrature-du-net-a-besoin-de-vos-dons/

A relatively large group of French scientists propose to add a fictional Alex to all our publications in all domains. This would make the Alex an extremely prolific scientific publisher:

http://rogueesr.fr/alex/

I think (and the people behind this idea do as well) that this initiative may be interesting to researchers all over the world. Unfortunately, the page they provide is in French. Maybe they will set up another one in English later.

#RogueESR

A trans friend of mine is looking for a roomie in Germany. Her current roomie is possibly moving out, leaving her in a dire situation. She's in an apprenticeship job that she NEEDS to keep to later acquire a certificate, so she cannot relocate too far away.

She is looking for a place within a radius of Wiesbaden/Mainz - Gießen - Hanau - Darmstadt around Frankfurt, with max 500 euros per month for rent/water/electricity/obligatory TV bill/internet.

For more info, DM me.

#art #mastoart

"Secret garden" from Pascal Campion

Further research is needed to fully understand how we managed to do such a good job.
https://xkcd.com/2268/

🎉 #PeerTube 2.1 released 🎉

This new release includes a new "internal" privacy setting for videos, hyperlinks for video timecodes in descriptions and comments, many comments improvements, a watch later icon on video thumbnails, UI improvements and many more!

Check the blog post for more information: https://joinpeertube.org/news#release-2-1-0

And the complete changelog on https://github.com/Chocobozzz/PeerTube/releases/tag/v2.1.0

Thanks to all PeerTube contributors 😘

I’m reading tutorial after tutorial on profiling in Haskell and I still can’t understand how my code can yield the cost centre stacks I get when I run it with +RTS -p.

If I understand correctly, cost centres are presented as a tree with the same structure as the actual function calls in the program. In this example, main calls fib so fib appears under main.

main Main 204 1 0.0 0.0 100.0 100.0 fib Main 205 2692537 100.0 100.0 100.0 100.0

Now in my program I have what I thought was a very simple parser (I use attoparsec) which use «primitive» parsers exposed to the library with a couple filters fmap on top, and that results in very deep cost centre trees. For instance I have :

integer :: (Read a, Num a, MonadParser m) => m a integer = read . Char8.unpack <$> decNumber <* blank <?> "decimal integer"

(decNumber is implemented from takeWhile1, I don’t remember why I didn’t use decimal directly but this is why there’s an unpack and a read on top of it, I’m pretty sure there’s a good reason) In any case, even if my reimplementing was awfully wasteful, I don’t understand how this code can yield the following cost centres

integer PDF.Object src/PDF/Object.hs:78:1-74 2296 0 2.3 2.3 2.9 2.7 runParser Data.Attoparsec.Internal.Types Data/Attoparsec/Internal/Types.hs:111:7-15 2297 8044421 0.0 0.0 0.0 0.0 fromPos Data.Attoparsec.Internal.Types Data/Attoparsec/Internal/Types.hs:46:21-27 2298 792520 0.0 0.0 0.0 0.0 dictionary PDF.Object src/PDF/Object.hs:(173,1)-(177,64) 2379 0 0.0 0.0 0.0 0.0 runParser Data.Attoparsec.Internal.Types Data/Attoparsec/Internal/Types.hs:111:7-15 2380 37745 0.0 0.0 0.0 0.0 name PDF.Object src/PDF/Object.hs:150:1-71 2378 0 0.0 0.0 0.0 0.0 operator PDF.Content.Operator src/PDF/Content/Operator.hs:(68,1)-(76,48) 2381 0 0.2 0.1 0.3 0.1 runParser Data.Attoparsec.Internal.Types Data/Attoparsec/Internal/Types.hs:111:7-15 2382 295454 0.0 0.0 0.0 0.0 fromPos Data.Attoparsec.Internal.Types Data/Attoparsec/Internal/Types.hs:46:21-27 2383 37745 0.0 0.0 0.0 0.0 regular PDF.Object src/PDF/Object.hs:75:1-79 2384 0 0.0 0.0 0.0 0.0 signature PDF.Content.Operator.Text src/PDF/Content/Operator/Text.hs:(16,1)-(32,3) 2385 0 0.0 0.0 0.0 0.0 stringObject PDF.Object src/PDF/Object.hs:(123,1)-(135,75) 2506 0 0.0 0.0 0.0 0.0 stringObject PDF.Object src/PDF/Object.hs:(123,1)-(135,75) 2300 0 0.2 0.1 0.3 0.2 runParser Data.Attoparsec.Internal.Types Data/Attoparsec/Internal/Types.hs:111:7-15 2301 1005360 0.0 0.0 0.0 0.0 fromPos Data.Attoparsec.Internal.Types Data/Attoparsec/Internal/Types.hs:46:21-27 2302 83780 0.0 0.0 0.0 0.0 name PDF.Object src/PDF/Object.hs:150:1-71 2304 0 0.1 0.1 0.1 0.1 runParser Data.Attoparsec.Internal.Types Data/Attoparsec/Internal/Types.hs:111:7-15 2305 754020 0.0 0.0 0.0 0.0 fromPos Data.Attoparsec.Internal.Types Data/Attoparsec/Internal/Types.hs:46:21-27 2306 41890 0.0 0.0 0.0 0.0

How can integer call runParser ? Ok, probably some magic from attoparsec (but still, the number of entries that yielded 😳) How can it call name and dictionary (2 completely unrelated parsers) and — worse — operator : a way more complicated parser that indirectly makes use of integer (and was defined months after integer anyway, I can guarantee : absolutely no call is made to operator from integer) ?? So what does those cost centres mean ?

Common complaint: "I can't read the backlog in IRC"

Discussions over the water cooler aren't logged either but no one seems to be asking to install a microphone in the breakroom

If you feel the need to catch up on discussions you missed (which people haven't explicitly called your attention to), your team is using real time messaging wrong. IRC is ephemeral by design, summarize discussions in email, docs, etc if they need to live longer

https://paste.x25519.net is available! An unikernel #MirageOS as https://privatebin.info/