Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
    wlach
    @wlach:matrix.org
    [m]
    BTW, there's a pyodide/sprints channel if you want to join
    mclare
    @mclare:matrix.org
    [m]
    hahaha I fought with it back in 2018 and had a long standing issue that was eventually closed
    oh cool
    can you drop the link?
    oh its on gitter
    wlach
    @wlach:matrix.org
    [m]
    pyodide/sprints on gitter.im if you're using matrix or https://gitter.im/pyodide/sprints otherwise
    mclare
    @mclare:matrix.org
    [m]
    got it, thanks
    mclare
    @mclare:matrix.org
    [m]
    hmmm when I run pnpm run dev-compiler with the latest updates, I'm getting a bunch of warnings regarding the @rollup/plugin-typscript package
    wlach
    @wlach:matrix.org
    [m]
    what do you see?
    mclare
    @mclare:matrix.org
    [m]
    probably about 20 warnings
    Plugin typescript: @rollup/plugin-typescript TS2769: No overload matches this call.
    Overload 1 of 2, '(...items: ConcatArray<Node<Data>>[]): Node<Data>[]', gave the following error.
    Property 'type' is missing in type 'ScriptNode' but required in type 'Node<Data>'.
    Overload 2 of 2, '(...items: (Node<Data> | ConcatArray<Node<Data>>)[]): Node<Data>[]', gave the following error.
    Type 'ScriptNode' is not assignable to type 'Node<Data>'.
    the dev site will still fire up, but I don't know if suppressing/fixings these warnings is something we should do, there are a bunch of different warnings
    Type 'Node<Data>' is not comparable to type 'ScriptNode'.
    wlach
    @wlach:matrix.org
    [m]
    Oh, yeah. :( I know there's some warnings that aren't yet fixed, but aren't fatal. I think I've just been ignoring them
    mclare
    @mclare:matrix.org
    [m]
    ah ok
    no worries then! I just wanted to make sure i wasnt using the wrong version of a plugin
    wlach
    @wlach:matrix.org
    [m]
    definitely would accept patches to fix them, but it's probably a bit of a rabbit hole
    mclare
    @mclare:matrix.org
    [m]
    yeah that's fair
    mclare
    @mclare:matrix.org
    [m]
    since pyodide is still building, I'm going to take another look at issue #199 for parsing myst content inside blocks since I'd worked on the panels
    I'll take a look at myst, but do you know the depth that the directives go?
    if that's a bigger issue to tackle, I'm open to any smaller suggestions for stuff I can work on
    wlach
    @wlach:matrix.org
    [m]
    definitely recursively parsing myst inside blocks is a bit of a larger thing to tackle. there's actually an issue open for it: irydium/irydium#199 -- it might be made easier by switching to MySTJS (although that's also a bit of a can of worms): irydium/irydium#254
    mclare
    @mclare:matrix.org
    [m]
    haha, well I can take a look at mystjs at least
    wlach
    @wlach:matrix.org
    [m]
    yeah it might be worth checking out-- also probably a bit of a can of worms, but there's the potential to share all sorts of work between jupyterbook/curvenote/irydium if we did adopt it
    if you're looking for something on the easier side, here's a smaller issue I've been meaning to get to:
    irydium/irydium#234
    mclare
    @mclare:matrix.org
    [m]
    cool, yeah I'd like to get at least one thing done during the sprints, so I'll start with that and if I finish, read up on mystjs.
    pyodide, still building, 90 min later
    wlach
    @wlach:matrix.org
    [m]
    I also haven't looked at it basically at all yet, but might be interesting to think about how pyscript could be made to work with irydium
    mclare
    @mclare:matrix.org
    [m]
    yeah that's an interesting question, they made it look really easy with the demos...
    mclare
    @mclare:matrix.org
    [m]
    so it could be added like one of the language plugins?
    1 reply
    ok well I'll open an issue for it in any case
    wlach
    @wlach:matrix.org
    [m]
    yup can't hurt to start a discussion around it
    mclare
    @mclare:matrix.org
    [m]
    I'm going to work on the css issue, because I think if I solve it, I might be able to tackle the pyscript issue. according to what was presented, all you really need to do is load two scripts
    so making it easier to load scripts with irydium seems to be the common theme
    wlach
    @wlach:matrix.org
    [m]
    yeah, I think it would be broadly useful to do for a lot of things
    mclare
    @mclare:matrix.org
    [m]
    trying to sort through script imports, would the goal be that the css files end up in a <style> block within the generated svelte? similar to how they are currently loaded within the vega example?
    extractCode currently returns only frontMatter, scripts, and codeCells so I don't know if you want to add another object to return, or if it should be a part of the scripts
    wlach
    @wlach:matrix.org
    [m]
    mclare: that sounds about right to me
    mclare
    @mclare:matrix.org
    [m]
    ok so I think I'll need to poke around some of the remark parsers to figure out the best way to do that
    wlach
    @wlach:matrix.org
    [m]
    mclare: Might be best to start with just handling a css block in front matter -- should mostly be able to avoid needing to mess with the parsers that way and you can prove the idea out
    mclare
    @mclare:matrix.org
    [m]
    ok sounds good
    wlach
    @wlach:matrix.org
    [m]
    mclare: how's it going?
    probably should have mentioned that I'd be happy to pair on some of this stuff 😄
    mclare
    @mclare:matrix.org
    [m]
    Hey there, I did figure out what I needed to do, but I’m kind of hampered right now because I’ve had something of a tech disaster this trip (my MacBook keyboard is hopelessly broken because it’s one of the butterfly ones, and my right half of my moon lander has become unresponsive). I have two ideas for how to do the css fix, and I guess it’s just a matter of if you want the changes to be in the parseMd.ts file (i.e. appending a <style> block to the body of any code cell where type === svelte) or if the style append should be passed in the object that returns frontMatter, scripts, and then appended to the code attribute.
    I was leaning towards the first one, since I’m assuming style blocks will only apply with top level svelte or svelte component blocks
    I’m flying back hopefully in an hour (flight got delayed)
    wlach
    @wlach:matrix.org
    [m]
    Hmm, yeah, I was thinking that the styles would be applied globally in the root component (at least for a first pass), kind of like what we're doing with bootstrap. Now that I think of it. You'll actually need to use a special svelte syntax to make this happen, fortunately it's not too bad: https://natclark.com/tutorials/svelte-global-styles/
    Have a safe flight back!
    mclare
    @mclare:matrix.org
    [m]
    Ah I was wondering about that because I thought child components in svelte didn’t inherit from parent components! Good to know
    wlach
    @wlach:matrix.org
    [m]
    Yeah style scoping is one of the big benefits of Svelte, but sometimes you want things to apply globally