A source location identification regression in GHC triggered by seven bytes

2021-03-22, post № 242

haskell, compiler-bug, #ghc, #parsing

Working on a minimal XML parser which does both build a data structure and remember its nodes’ source origin, I discarded the second part of a two-tuple by not requiring anything to be present. Yet instead of informing me that a mapping cannot be deconstructed, GHC 8.10.4 gave me an obscure error message:

% runhaskell Xml.hs

<no location info>: error: Tuple section in pattern context

Thankfully I sometimes adhere to a mindfully incremental approach and was not as bewilderingly disoriented as I could have been — after all, a there is an error message possesses little more value than solely stating an executable cannot be built.

Inspecting the probable location I had intelligence on, I managed to extract a seven long byte sequence which triggers a <no location info> compilation error:

j(0,)=0

Yet compilers are squeamish and shapeshifting entities — especially with regards to diagnostic capabilities. Whilst j(0,)=0 cannot be traced, j(,0)=0 induces the expected and appropriate error. As such, I tested the phenomenon using varying GHC versions:

% ghc --version && printf '%s' 'j(0,)=0' > /tmp/nli.hs && ghc /tmp/nli.hs 2>&1 \
cmdand cmdand>     | grep -q '<no location info>' ; echo $?
The Glorious Glasgow Haskell Compilation System, version 8.10.4
0

% ghc --version && printf '%s' 'j(0,)=0' > /tmp/nli.hs && ghc /tmp/nli.hs 2>&1 \
cmdand cmdand>     | grep -q '<no location info>' ; echo $?
The Glorious Glasgow Haskell Compilation System, version 8.8.4
1

To my surprise, I seemed to have found a GHC regression — triggered by only seven bytes.

Alas, this regression is a known one: Another <no location info> error (Tuple section in pattern context) has been opened on the 8th of March 2021, with its comments identifying it as a regression. Yet the fix outlined and performed is as much resolving an aspect of the issue as is it making a stylistic decision — myself facing the same problem of source location corseness, I am leaning further and further towards the approach of less pedantically accurate reporting allowing for a more minimal implementation.

Intriguingly Matured Graphics

2021-03-20, post № 241

imagery, Pygame, #throwback, #png, #colorful

Following digital excavation efforts at my disk’s deep directories, I stumbled upon a collection of colorful pseudo-random walk graphics. Since they are dated September of 2015 and were generated using unidiomatic slow snake script, their only property of note is a visually jolly aura; the following three possess a particularly vibrant one:

Autumn Colors

Deep Blue

Woven Violet

Read further …

Cellular circuit simulation

2021-02-20, post № 240

c++, grid-world, #binary, #logic, #gate, #ppm

Inspired by grid worlds, non-linear notation and two-dimensional esolangs, I have attempted to design a few ASCII-art languages myself, none satisfactory enough for publication. Without the toolchain to interact in a non-typewriter manner — both on the software as well as on the hardware side — paired with the need for an apt encoding to facilitate higher-order capabilities, I could not manage to create something which stands on its own feet as a proper language, as opposed to nothing more than a convoluted yet primitive processor emulator.

In the fall of 2020, when I was tasked to teach elementary binary semantics courtesy of a brand new mandatory lecture at my university — constructing half-adders from basic gates and combining them to build full adders —, I thought that exactly this bare-bones grid world might be a fruitful endeavor for constructing and combining gates with a visualization of the entropy’s movement across the circuit (one might foolishly think of a bit meandering across a wire, although this interpretation has no physical merit to it).
Within a few hours, I had managed to settle on a grid world definition together with an under 200 lines long interpreter for it. As I opted for an ASCII-CLI-look — significantly boosting development time —, I added image output facilities for this blog post (for which I swiftly designed a few pixel glyphs; only those used by the grid world), avoiding the need to take screenshots of my terminal emulator.

Source: cellular-circuit-simulation.cpp, building: Makefile
Circuits: cellular-circuit-simulation_adder.circuit, cellular-circuit-simulation_odometer.circuit, cellular-circuit-simulation_spiral.circuit

The grid world

As in most grid worlds, non-inert characters are kept to a minimum: there are two entropy sources 0 and 1, the unary negation gate ! and three binary gates &, | and ^. All other characters except the space allow entropic bits to replicate, the special jumpers < and > allow to cross a gap of three characters, rendering interleaving wires possible.

Designing a 𝟥-bit adder

Calculating 0b110 + 0b011 == 0b1001 using a 𝟥-bit adder (input bits are interleaved, less significant bits reside on the left).

Read further …

Visualizing cycles in row-major transposition encodings

2021-01-23, post № 239

mathematics, programming, c++, shell, #matrix, #encoding, #permutation, #rainbow

A matrix $A\in \mathcal{D}^{h\times w}$ of discretely representable entries $\mathcal{D}$ may be linearly layed out in memory using row-major order, concatenating successive rows into a contiguous $(h\cdot w \cdot\texttt{sizeof}\,\mathcal{D})$ -bytes long array. Such a representation, however, is disruptive to the matrix’ two-dimensional nature: whilst horizontally neighboring elements remain neighbors, vertically neighboring entries are torn apart by insertion of $w$ non-neighboring elements. As such, on matrices naturally defined operations get distorted by this encoding.
One such inherently two-dimensional operation is matrix transposition. In the realm of matrices, $\mathcal{D}^{h\times w}\neq\mathcal{D}^{w\times h}$ are for nonsquare dimensions semantically different, being mapped to one another by transposition. Projecting onto their encoding, this semantic is lost and one is left with a permutation on memory $\bullet^\top\in\mathrm{Sym}(\{1,\dots,hw\})$ .
To visualize this permutation, its cycle decomposition is computed of which each cycle is given a color of the rainbow dyeing this cycle’s corresponding two-dimensional pixels when interpreting its path on the underlying array in the semantics of the original matrix.

Initial transposition cycles

Above listed are all visualizations for $1\leq h\leq 8,1\leq w\leq 16$ , shuffled. Whilst some behave extremely regularly — for example square matrices’ transposition permutations decompose into transpositions —, others are wildly intricate. Each of them adheres to a rotational symmetry; the top left and bottom right are fixed points.

Read further …

Factoids #2

2020-12-26, post № 238

mathematics, #bijection, #calculus, #Lipschitz, #naturals

VII) Cardinality coercion: $\mathbb{R}\hookleftarrow\{\mathbb{N}\to\mathbb{N}\}\twoheadrightarrow\mathbb{R}$

Claim. There exist both $\iota:\{\mathbb{N}\to\mathbb{N}\}\hookrightarrow\mathbb{R}$ together with $\pi:\{\mathbb{N}\to\mathbb{N}\}\twoheadrightarrow\mathbb{R}$ .

Proof. Iota. Define $\iota:\{\mathbb{N}\to\mathbb{N}\}\hookrightarrow\mathbb{R}$ via

$\begin{aligned} \iota(a):=\quad&\sum_{j=1}^\infty 2^{\left(1-\sum_{i=1}^{j}(1+a(i))\right)}\cdot(2^{a(j)}-1)\\ =\quad&\left(0.\underbrace{1111\dots11}_{\times a(1)}0\underbrace{111\dots11111}_{\times a(2)}0\dots\right)_2 \end{aligned}$

and observe any sequence’s reconstructibility by dyadic expansion.

Pi. Define $\pi:\{\mathbb{N}\to\mathbb{N}\}\twoheadrightarrow\mathbb{R}$ via

$\begin{aligned} \pi(a):=\quad&a(1)+\sum_{j=2}^\infty 2^{1-j}\cdot\delta_{a(j)\in 2\mathbb{Z}}\\ =\quad&a(1)+\left(0.\delta_{a(2)\in 2\mathbb{Z}}\delta_{a(3)\in 2\mathbb{Z}}\delta_{a(4)\in 2\mathbb{Z}}\dots\right)_2 \end{aligned}$

and observe any real’s constructibility by dyadic expansion.

Thus, $\{\mathbb{N}\to\mathbb{N}\}\cong\mathbb{R}$ in $\mathbf{(SET)}$ is shown.

Read further …

A Month of Pining

2020-11-28, post № 237

hardware review, #opinion, #diary, #freedom

Precisely one month ago, I opened a long awaited parcel shipped directly from Hong Kong: my very own Pinebook Pro had arrived. Eager to further my grasp on both software and hardware freedom, I unwrapped my mail and began to incorporate pine64’s flagship laptop into my workflow. Accompanying this journey, I wrote a diary which I share in this post.

Whilst a Pinebook may not boast with specs of performance, novel input methods or included licenses, it is part of a niche computing sector which values and upholds principal human rights even in the alien realm of microprocessor-based technology. Following my steadily growing discomfort regarding my computing equipment over the past months, I hoped for the Pinebook Pro to introduce me to a world of productive, in part minimalist and most importantly truly free computing experience. After a month of usage, I can confidently proclaim it to have delivered on this front more satisfactory than any other of my multitude of computing devices I have amassed over the years.

Especially in current times, it is calming to interact with an intricately mingled system of physical machinery and abstract commandments in which I can at least to a certain extent trust, lifting an ever-present veil of fear cloaking most boxes of bits. A Pinebook comes with software switches to turn off the built-in camera, microphone and WiFi card as well as Manjaro KDE Plasma running out of the box. Whilst I would prefer not having a camera or microphone built into my device at all and have a real physical switch to turn both off, as well as a more minimalist desktop environment, I am capable to entrust it my presence infront of it; in contrast to e. g. a proprietary phone lying on a tabletop nearby.

I am unaffiliated with pine64^[1] and purchased my Pinebook Pro myself. I wrote the entirety of this blog post including the diary seen below using my Pinebook^[2].

A detailed account of my experience with my Pinebook is given by a diary I wrote over the course of November 2020:

Read further …

Halloween MMXX

2020-10-31, post № 236

imagery, poem, #halloween, #spooky

The sea is full of dangers,

size being only one.

Alluring depths and gruesome beasts

have intertwined in majesty.

The humble one left all alone

sought guidance in a spark of light.

Extra assets: halloween-mmxx.c, halloween-mmxx_settings.h, halloween-mmxx_all-images.tar.gz

Crashing GCC with 63 Bytes

2020-10-15, post № 235

C++, error, #GCC, #g++, #internal compiler error

Enthusiastically following C++20’s new compile-time capabilities as well as awaiting more powerful ones being promised in upcoming standards, yet being stuck on C++17^[1], I try to take full advantage of the limited C++17-constexpr features. However, without static storage promotion, a constexpr std::string is all but a pipe dream, thereby necessitating the use of a wrapped static pointer; std::string_view.

One does not program against one’s compiler; one programs against the abstract machine defined by the standard.

Read further …

Posts: 290-283, 282-275, 274-267, 266-259, 258-251, 250-243, 242-235, 234-227, 226-219, 218-211, 210-203, 202-195, 194-187, 186-179, 178-171, 170-163, 162-155, 154-147, 146-139, 138-131, 130-123, 122-115, 114-107, 106-99, 98-91, 90-83, 82-75, 74-67, 66-59, 58-51, 50-43, 42-35, 34-27, 26-19, 18-11, 10-3, 2-1

Jonathan Frech's blog; built 2024/08/31 22:59:44 CEST