Kevin Turner @keturn@toot.cat

infotainment high-five apophenia markup language support system

I always second-guess the spelling of #florespondence

#florespondece from Wednesday the 12th.

(I'm working on a pull request I started at the #PyCascades sprint yesterday, that's what reminded me of this.)

This repo doesn't live here, I'm just making PRs: <https://gist.github.com/keturn/9ea3aa7e005cb7cfa5b0bb236c5526e8>

In the GitHub workflow (which is ubiquitous these days), you submit contributions by forking the repo, which shows up prominently on your profile and namespace.

But I'm not maintaining a long-lived branch and this URL isn't one anyone should use to reference the upstream project.

I've made a little script that replaces the default branch in your forked repo with a sign that says "Pull-Request Only Fork."

Okay Pythonistas. Pythians. Portlanders and other Cascadians. I'm not at #PyCascades today but I'm signed up for Monday's sprints.

wanted: a browser/window manager plug-in that detects if I am mindlessly flipping between windows in a novelty-seeking cycle, and kicks me off for a while.

(I've had some success with things like Leechblock, but this would be more checking behavior than strictly filtering by destination.)

to temper this a bit: just discovered Darktable using 4 CPUs at a time when I think it should have been idle for the last many hours.

so maybe some bugs to file after that reading. In the meantime, something to keep an eye on.

Relatedly: tried out Darktable today (for editing digital photos shot in raw mode).

Seems pretty good! Flexible, snappy, and stable so far. But goodness there are a lot of ways to do very similar things:

What's the difference between "exposure correction" and adjusting brightness? How does that compare to tone mapping, or "global tonemap"? Where does the "base curve" come from? Which things should I use in combination, and when should I just pick one because they're alternative approaches to the same task?

I've got some reading to do.

https://www.darktable.org/

#florespondence

Unicode emoji proposals continue to be entertaining reading.

> The expression “Don’t feed the troll” could be represented as TROLL + POULTRY LEG + COMBINING ENCLOSING CIRCLE BACKSLASH. In the internet image on the right this expression is depicted as someone handing a roast chicken to a troll, enclosed in a prohibition sign.

— from L2/19-232 by West & Everson, which is a cromulent submission overall, though this suggested sequence is pushing it.

https://www.unicode.org/cgi-bin/GetMatchingDocs.pl?L2/19-232

#unicode #emoji

The reason I've been working out how to scroll by column is, of course, so I can make landscape-mode layouts for reading web fiction, like https://gitlab.com/keturn/multicolumn-ao3-usercss

`position: sticky`

Neat CSS rule where the element starts off positioned as normal, until you try to scroll it away, at which point it sticks to the edge.

https://developer.mozilla.org/en-US/docs/Learn/CSS/CSS_layout/Positioning#position_sticky

I was about to write "`scroll-snap-type: mandatory` is not sufficient on its own," and I went to double-check.

I *do* have scroll-snap working so Firefox's default handling of the right arrow key gets to the next column.

But not left. It must always round up or something, so 1 + 0.1 = 1.1 → 2, but trying to go backwards 2 - 0.1 = 1.9 → 2.

I've had some success combining this with Scroll Snap, which is a handy thing for this use case:

https://developer.mozilla.org/en-US/docs/Web/CSS/CSS_Scroll_Snap

where my arithmetic gives me a close-enough value and then scroll-snap will snap to. But it would be nice to not need to both add the keybinding *and* manipulate the document's CSS.

Know what's great for long-form text on these widescreen monitors? Multi-column layout!

And CSS has columns these days.

So just as long as there's an easy way to advance to the next column or page…

https://column-scroll.glitch.me/

haven't quite got it. A column isn't exactly an element in the DOM, so I'm having trouble accurately querying its width, in order to know how much to scroll.

I was tipped off to this when looking at the source of Colorful, which I noticed was mathing up the color codes from RGB values in these low-color modes, instead of having some 8-value lookup table.

https://pypi.org/project/colorful/

In the days of 8-color displays (or 16, where the the second half was just brighter versions of the 8), I never saw a pattern in the order the colors came in for things like ANSI X3.64 codes. I assumed they had their small set of colors but the order was arbitrary.

But 8 is how high you can count in three bits: one bit each R, G, and B. The direction may differ between implementations; looks like ANSI X3.64 has Red as the lowest and CGA has Blue as the lowest. But the order of the colors just comes from counting up in binary.

0 0 0 - black
0 0 b - blue
0 g 0 - green
0 g b - cyan
r 0 0 - red
r 0 b - magenta
r g 0 - ~yellow~ brown?

They made a special case for brown, where the green bit only counts for half, because they thought it made a better palette than a dull yellow: https://en.wikipedia.org/wiki/Color_Graphics_Adapter#With_an_RGBI_monitor

Never would have guessed they went to extra expense for that CGA brown!

Using jq to parse output from snapd to generate commands to run through sudo is a fine activity for a winter evening, right?

https://gitlab.com/snippets/1927120

I bet I was doing something else entirely before I stumbled across those old installations…