@metadave if you are after something that is part of the official racket project try https://github.com/racket/typed-racket/issues?q=is%3Aopen+is%3Aissue+label%3Agood-first-bug

@spdegabrielle thank you! I’ll take a look some time today. I’d most likely need something small to start on.

@metadave any particular interests? I can try suggest some things (deleted not relevant)

Oops I forgot you original question. Did you take a look at the ‘good first bug’ link I gave above?

@metadave https://github.com/topics/racket shows issues that are labelled good first issue or good first bug in repositories that have the racket topic

MetaPict picure of the day:

Source: https://github.com/soegaard/metapict/blob/master/metapict/examples/racket-logo.rkt

@soegaard2 I would think that the dimensions were computed from something else by @mbutterick originally

@samth Do you know where to look for the original source?

You seem to have put quite some effort to obtain these values!

It’s no too bad. WebPlotDigitizer is quite handy.

And despite the name it is a native program - not a web-app.

I get the following error when drawing the second picture: length: contract violation expected: list? given: (raw-gradient (raw-color-stops (list (object:color% ...) (object:color% ...) (object:color% ...) (object:color% ...) (object:color% ...)) '(0 0.18 0.36 0.5 1.0)))

@soegaard2 I looked back in my email archives, and @mbutterick says he did it “by hand” and only SVG sources were provided. I don’t know if there was a font used to construct it originally.

@laurent.orseau You need to update. I made a change to gradient.rkt (now linear-gradient accepts a gradient as the color argument)

Okay. Thanks for checking.

I just installed metapict. I guess pgk hasn’t caught up yet.

I do notice that the WebPlotDigitizer output has many more numbers than the original SVG, which seems odd

I should have looked in the svg instead !

WebPlotDigitizer works by showing the image / diagram on screen. In the case of plots you when place a coordinate system on screen. Then you add points on top of the image along the curve you are interested. In the end, you get a txt file with numbers. The main use case is for reading x and y values on plots were the author hasn’t included the original numbers.

Just checked github - and this time I didn’t forget any files :slightly_smiling_face:

Yes. I think I could have read fewer numbers without any problems.

potentially it’s just matching discretization results very precisely which isn’t really what you want when there’s a smooth curve generating things but is unsurprising for such a tool

You can trigger pkg to refresh your package if you log in and click something there

It doesn’t matter too much - MetaPict doesn’t connect the points with straight lines - it finds a smooth curve through the points.

Done (didn’t think about that). > The results will be available after the next index refresh, which is scheduled for Thursday, August 8th, 2019 4:24:26pm (UTC)

right and presumably that curve will be very close to but not quite a circle

it works now, very nice!

Except that the gradient doesn’t do what it’s supposed to do apparently:

Fun fact: Circles drawn with Bezier curves aren’t circles.

Damn. Let me guess. You are using DrRacket and have the repl at the right?

yes

Try using the horizontal repl instead.

crosses fingers

haha, almost :smile:

let me shoot it

it also changes if I switch back and forth to another window

And here I though I had finally cracked gradients. The problem is that dc% has one coordinate system for drawing (user space) and another for gradients (pattern space). And figuring out the exact relationship is tricky.

Huh? That looks odd.

ah sounds tricky indeed

I guessed that the center of the radial gradient was wrong - but this?

it’s actually worse: If I switch to another window, it becomes all red again

Is it because the repl scrolls while drawing?

ohh, it also depends how much I scroll

in some places it looks nice

(for the top right)

I can see the gradient change as I scroll

My guess is that if you use pict->png and displays the png everything will be ok.

Maybe it is named pict->bitmap

you can also use freeze

Okay.

(yes, pict->bitmap indeed)

freeze should freeze the dc boundaries, and the gradient’s coord. references should stick

I suppose…

I guess Emacs just renders all picts to pngs where as DrRacket renders picts to the canvas-dc% directly.

Cool. Makes it less urgent to fix the problem :wink:

Maybe write a note about this somewhere?

yes.

And add freeze to all examples with gradients.

@samth can this be the banner for reddit/r/racket ? @soegaard2

I did that, but a version of the image that’s 128px high and has a dark background would be better

@spdegabrielle put the second logo in a freeze. The gradient is wrong.