PyScript 💻
💬 | General / ╰╸pyscript-core
until I turn it back on
1
I am also on the CPython core devs server which has a very similar icon, so I always get confused 
I am also on the CPython core devs server which has a very similar icon, so I always get confused <py> icon?
I am also on the CPython core devs server which has a very similar icon, so I always get confused 
1 1
Discord Server and not Discord Channel as this is a server and within server we have channels
1
two to three dozen isn't super uncommon for general ones, though I don't know if there are programming- discord communities we could model off of two to three dozen isn't super uncommon for general ones, though I don't know if there are programming- discord communities we could model off of 
- Discussions · nodejs/node
py-repl mentions exec-id and root, but they are internal-use only and they should not be documented IMHO 1 1
1
My usual handle everywhere is just ntoll, with the orangutan 
Hello world!
) ) 
1
Yesterday & Today - 11/02/2022 
greetings from Sweden 4
2userError haha I misunderstood what you said, I thought you were suggesting that instead of us doing showError("blah") we would do raise UserError and then the withUserErrorHandler should allow the code to keep going. Buuuut.... I clearly misunderstood and tried to do the impossible? 
It must be the 3rd or 4th time in a week!!! lol
Anyway, will comment on the issue with a few considerations/tweaks
2 It must be the 3rd or 4th time in a week!!! lol
Anyway, will comment on the issue with a few considerations/tweaks 
1addToProject CI job fails suddenly: https://github.com/pyscript/pyscript/actions/runs/3379019433 
1 1
1 1 Yesterday & Today - 11/03/2022 node when running jest tests 
1<py-terminal> tag, implemented as a Plugin 
3
1 3<py-terminal> tag, implemented as a Plugin /pyscriptjs/plugins or /pyscriptjs/src/plugins), new repo (like /pyscript/plugins) or.... ? Yesterday & Today - 11/04/2022 ffmpeg under WASM/WASI, which also explains (although a bit too shortly) the differences between WASI and emscripten:
https://runno.dev/articles/ffmpeg 4
2
2 1 1
(module
(import "console" "log" (func $log (param i32)))
(func (export "logIt")
i32.const 13
call $log))
this imports console.log as $log, then creates a function logIt that calls it.
links: https://developer.mozilla.org/en-US/docs/WebAssembly/Understanding_the_text_format
wabt: https://github.com/WebAssembly/wabt #include <Python.h>) or it's a plan C library and you call it using ctypes/cffi. (I'm not sure that ctypes/cffi works for WASM but in principle they should).
For MicroPython, I have no idea, but the first step is to understand whether it has support for extension modules independently of WASMimport numpy in pyodide . I predict that reimplementing the js module would take several months, and probably we would introduce other quirks. This always happen when you try to map features of language A onto language B
♂️ , but feel free to try 1 Yesterday & Today - 11/07/2022 1to_file parameter. 1
1
['joe', 'jane', 'mary'].includes('jane') 1<py-repl> widget but with different runtimes (first MicroPython, then Pyodide) - in MicroPyScript.
3
2 But I think we need to use showWarning or maybe just create the error banner when we call handleFetchError, the test is failing because the page never stops loading - the fetch raises the error so we never call the logic to hide the loading screen.
Also since we get a list of paths in fetch Paths, would it make sense to show the warning and keep trying other paths? 
I wonder if we should be start and if we handle any of the UserError we try to toggle the loading screen off But I think we need to use showWarning or maybe just create the error banner when we call handleFetchError, the test is failing because the page never stops loading - the fetch raises the error so we never call the logic to hide the loading screen.
Also since we get a list of paths in fetch Paths, would it make sense to show the warning and keep trying other paths?
I wonder if we should be start and if we handle any of the UserError we try to toggle the loading screen off pyscript_run in support.py: it takes a param wait_for_pyscript=True which does exactly what you need
ImportError (if the failed file was a .py file) or a FileNotFoundError later. But yeah that makes a lot of sense! Although I wonder if this failing test also picked something odd?
The loading screen doesn't finish loading which we expect, but nothing is showing in the page either and no error in the console which is odd 
. Are you waiting for some answer/input from me, or you have everything you need not to be blocked?
ImportError (if the failed file was a .py file) or a FileNotFoundError later. Yesterday & Today - 11/08/2022 display() and py-config [[fetch]]) 3 1display() and py-config [[fetch]]) .
2
1
Yesterday & Today - 11/09/2022 
1
is magic 1
1 4
_createBanner in handleFetchError.
If you do --headed does the splash screen disappear from the page? Also what failure are you seeing? _createBanner in handleFetchError.
If you do --headed does the splash screen disappear from the page? Also what failure are you seeing? 
--headed it works, with --dev it doesn't 1
Yesterday & Today - 11/10/2022 1 1 2 1
1
/captcha start
1
feel like @ntoll would be fond of it? 5
2
1
1 
1 Yesterday & Today - 11/11/2022 squash&merge. I started my more-plugins branch on top of py-terminal, but now that py-terminal is merged to main, when I do git merge origin/main I get tons of conflict, because from the point of view of git the commit which merged py-terminal has nothing to do with the original branch .squash&merge. I started my more-plugins branch on top of py-terminal, but now that py-terminal is merged to main, when I do git merge origin/main I get tons of conflict, because from the point of view of git the commit which merged py-terminal has nothing to do with the original branch . git rebase --onto=main antocuni/py-terminal antocuni/more-plugins
which basically means "take all the commits which starts from py-terminal (excluded) to more-plugins (included) and rebase them on main custom-typings.d.ts and do something like:
export declare type "Pyodide" { blah}
But that didn't work, although my terrible workaround seems to be okay(ish) , I'm creating an interface like this just to support the extra methods that the Pyodide typing doesn't have
export declare interface PyProxy extends PyProxy {
connect(): void;
get(value: string): function | object;
set(id: string, proxy: PyProxyClass): void;
}
Although the get return is pretty bad 
1) those are very good points also for the get specifically even with these types we still need to write a type for the thing we are getting like Display for example.
I might have been a bit too eager to kill all warnings npm run lint thinking that this was the issues that strict mode would complain about 
1:wq :wq vim suddenly looks like a very reasonable choice 
1
2 4 ❤️ 4
1 1
1
1 
stdout of a <py-script> tag on a tag-by-tag basis?
Similar to how <py-script std-out="..."> prior to the introduction of display()? OutputManagerCtx was half-broken . For example, it didn't intercept the stdout which were emitted by C code
1 1 1 1 Yesterday & Today - 11/14/2022
If not I might poke at it at the research repo so I get my head around it 
It's a fiddly topic.) 1
<my-counter> https://github.com/pauleveritt/pyodide-components/blob/main/src/pyodide_components/counter.py#L6
2<my-counter> https://github.com/pauleveritt/pyodide-components/blob/main/src/pyodide_components/counter.py#L6 1 1 ).
I do like the registry idea. But that makes it fun, exploring Python-based components
<pyscript src="app.py"></py-script>
<pyscript worker="one" src="worker.py"></py-script> 
♂️♂️ 1pydom.document.body is something like this:
class PyNode:
...
def set_inner_html(self, val):
self.postMessage("update", self.id, "innerHTML", val)
innerHTML = property(get_inner_html, set_inner_html) Yesterday & Today - 11/15/2022 1 2 1<py-script version=“2.7” curly_braces=“true”> 1 Yesterday & Today - 11/16/2022 chat channel? 
5 Yesterday & Today - 11/17/2022 pyconfig we are iterating over each parameter and calling validateParamInConfig if the parameter isnt in the config we just ignore it, should we throw an exception instead? 2pyconfig we are iterating over each parameter and calling validateParamInConfig if the parameter isnt in the config we just ignore it, should we throw an exception instead? Plugin.configure hook
5. each plugin is responsible to read and mark all its settings in its .configure()
6. after the hook, we check whether there are still settings which have never been read. If there are, we emit an error/warning Plugin.configure hook
5. each plugin is responsible to read and mark all its settings in its .configure()
6. after the hook, we check whether there are still settings which have never been read. If there are, we emit an error/warning <py-config>
packages = ['numpy']
[plugins.py-terminal]
auto = true
[plugins.splashscreen]
autoclose = false
</py-config>
I think the process is essentially still the same as you outlined, but except list of top-level keys in py-config is fixed (and includes the 'plugins' table).
Yesterday & Today - 11/18/2022 Plugin.configure hook
5. each plugin is responsible to read and mark all its settings in its .configure()
6. after the hook, we check whether there are still settings which have never been read. If there are, we emit an error/warning )
Agreed that we should either error or emit a warning, specifically in the "i have a typo" case were we expect things to work but they would be silently ignored and probably break further down the line when we try to do somethingplugins table lets us distinguish misspelled settings from misspelled packages
Like if we parse py-config and it has a top-level key futch, is that a misspelling of [[fetch]] or trying to identify settings for a plugin called futch (or fitch)? [plugin.splashscreen] and [splashscreen], apart that the user has to type more: we can easily detect whether there is an "unexpected section" in both cases. 1[plugin.splashscreen] and [splashscreen], apart that the user has to type more: we can easily detect whether there is an "unexpected section" in both cases. packages = ['numpy']
[splashscreen]
autoclose = true
[py-terminal]
auto = true
{
"packages": [
"numpy"
],
"py-terminal": {
"auto": true
},
"splashscreen": {
"autoclose": true
}
}
VS:
packages = ['numpy']
[plugin.splashscreen]
autoclose = true
[plugin.py-terminal]
auto = true
{
"packages": [
"numpy"
],
"plugin": {
"py-terminal": {
"auto": true
},
"splashscreen": {
"autoclose": true
}
}
}A higher-level is intentional: there might be multiple. And you're plugging into that thing, not to PyScript.
I think this is Ted's view. config.keys() have been used or not, each plugin gets its own object of settings. plugins.* sections: if someone writes [plugins.py-treminal] I want to make sure to inform them about the typo 1
There are 2 sides of the coin: how do we allow plugin authors to declare their config requirements and how we allow users to set these configs in their own apps.packages = ['numpy']
[splashscreen]
autoclose = true
[py-terminal]
auto = true
{
"packages": [
"numpy"
],
"py-terminal": {
"auto": true
},
"splashscreen": {
"autoclose": true
}
}
VS:
packages = ['numpy']
[plugin.splashscreen]
autoclose = true
[plugin.py-terminal]
auto = true
{
"packages": [
"numpy"
],
"plugin": {
"py-terminal": {
"auto": true
},
"splashscreen": {
"autoclose": true
}
}
} <py-config>
plugins = ["foo", "bar", "http://..../baz.js"]
</py-config>
where foo and bar are downloaded from some kind of Official PyScript Plugin Store, and baz is downloaded from a custom URL collections.deque doesn't support indexing but only push and pop so why not just use a list..? It's all about tradeoffs right..? so why not just use a list..? It's all about tradeoffs right..? MemoryError. So he wanted to manually implement a deque using a fixed-size list, and I suggested using collection.deque which seemed perfect for his use case 1 beforeSetup hook if you call it after the setup? It's just confusing IMHO 1
beforeSetup hook if you call it after the setup? It's just confusing IMHO 1beforeAAA, beforeBBB and beforeCCC hooks instead of afterDDD since they will be fired all together anyway? beforeExecute I am certain that all beforeSetup hooks were called, because that's the only way to ensure that the setup has been done.
Else, we are just encouraging a messPromise.all([list of awaiting events]).then( dispatch "SetupComplete") (pseudo code) to achieve this end. Promise.all(foo) ensures that a single event is dispatched when they all are finished. 
1 .
When I call the afterSetup hook/event it's because I'm sure that the virtualenv setup has been completed. It's a very explicit sync point
1plugin1.afterSetup is called before plugin2.beforeSetup, and this completely breaks your use case. beforePythonIsReady event: you cannot call it in python because this is obviously happening after the interpreter is ready beforePytonIsReady then, you have no luck in a Python plugin but if your goal is to just read the state of things at that point in time, then it's still useful to have that event be passed to a Python plugin, so you can know the state at that point in time anyway. 2 1 2
6 1
2/latest in all our docs and official examples: people will just copy&paste the <script> tag and the result is that we are involuntary encouraging to publish pyscript apps which will break after 30 days. Enjoy your vacation 
/releases/2022.09.1/.... exists. Personally, I wanted to use it at some point and I couldn't find the exact url in the docs Yesterday & Today - 11/21/2022 /latest been widespread for a couple versions since versioning was first announced, maybe need to find a way to deprecate it. A warning banner at the top of the page saying you should be linking to a pinned release? Not sure exactly how that would work from a deployment standpoint, but could even be something that gets done "manually" (changed, rebuilt, uploaded to /latest resource on AWS) /latest to the current release. IMHO it's better to keep apps working+warning than broken+warningbacklog tag to our valid issues, otherwise they'll be automatically close by the bot! 
backlog tag to our valid issues, otherwise they'll be automatically close by the bot! backlog? 1
1
I don't remember that we discussed this but ok, good to know
I don't remember that we discussed this but ok, good to know 2/latest been widespread for a couple versions since versioning was first announced, maybe need to find a way to deprecate it. A warning banner at the top of the page saying you should be linking to a pinned release? Not sure exactly how that would work from a deployment standpoint, but could even be something that gets done "manually" (changed, rebuilt, uploaded to /latest resource on AWS) alpha . (Just adding that as a data point). They are driving pdb by faking user input.
I'm a bit surprised, I assumed they would use bdb which is made exactly for this
(basically: bdb implements the core functionality, and pdb adds a terminal-based UI on top of it) Yesterday & Today - 11/22/2022
I'll just get things down as bullet-points for now... this is a brain dump so it might be a bit scattered....
)
) 2/$debug/input, which seems to be implemented here:
https://github.com/microsoft/vscode-python-web-wasm/blob/78bd69365bbf53957ca16f9b1ab87fba5dfe34b3/src/common/debugFileSystem.ts#L186-L188
but then it forwards to apiClient.byteSource.read(), which is most likely the blocking call apiClient.byteSource.read() returns a promise (see: https://github.com/microsoft/vscode-python-web-wasm/blob/78bd69365bbf53957ca16f9b1ab87fba5dfe34b3/src/common/debugCharacterDeviceDriver.ts#L45) I'm not at all clear what the const disposable does to cause waiting for user input. I have no idea what the .dispose() method does, for instance (it's not something in the "standard" Promises API AFAICT)
runtime into interpreter.
Probably at some point we should do the renaming, although probably it's better to wait for the next release.
(And personally, I'd also like to make it a single entry and not a list... we support only a single interpreter anyway). runtime into interpreter.
Probably at some point we should do the renaming, although probably it's better to wait for the next release.
(And personally, I'd also like to make it a single entry and not a list... we support only a single interpreter anyway). I'm happy to tackle that once the release is out 3 1+ icon as you usually do to insert a comment ), in order a small API to learn a bit more in this iteration and add more in the following release.
1 1 1
5 2 2 1 
3 Yesterday & Today - 11/23/2022 make tests parallel throw different failures (Fabio noticed this last week (was it last week or this one?)) 1
(For those who aren't celebrating that particular holiday - Turkey and cheer for y'all too!) 4 2make tests parallel throw different failures (Fabio noticed this last week (was it last week or this one?)) make tests-parallel just run pytest -n under the hood.
-n activates the pytest-xdist plugin, which is responsible for the actual parallelization
So, it's not pytest vs make, it's pytest vs pytest+xdist.
According to my experience, it's very likely that it's a bug in our testsuite (e.g., a common mistake is to have multiple tests trying to write/read to the same file in parallel, so you have a race condition).
What is the exact error? 1++ 5 Yesterday & Today - 11/24/2022 
1
1 2 2 1 1 
1
- sorry if I'm stepping on any toes here, just excited to see this getting merged for the impending release - sorry if I'm stepping on any toes here, just excited to see this getting merged for the impending release Yesterday & Today - 11/25/2022 await somewhere), I would be surprised if it has anything to do with xdist.
If you see that the single test passes with xdist fails, it is very likely that running in parallel makes each test slightly slower, meaning that JS has more time to complete some fetch.await somewhere), I would be surprised if it has anything to do with xdist.
If you see that the single test passes with xdist fails, it is very likely that running in parallel makes each test slightly slower, meaning that JS has more time to complete some fetch. )<py-script> tag before execution and save the sources in a different div Maybe a similar thing to what we have on the fetch in py-config?plugins = ["py-plot"] and we automatically know where and what to download.
When you do pip install foo you don't have to tell pip the complete lists of files to download 
. Writing the issue right now . Writing the issue right now
I would also want the docs PR (the one that updates the examples) and the deprecation of output in for this release if possible plugins = ["py-plot"] and we automatically know where and what to download.
When you do pip install foo you don't have to tell pip the complete lists of files to download foo-plugin folder that has a bunch of python files, we would do plugins = ['foo-plugin'] and would import the whole folder? I like that
I would also want the docs PR (the one that updates the examples) and the deprecation of output in for this release if possible 
foo-plugin folder that has a bunch of python files, we would do plugins = ['foo-plugin'] and would import the whole folder? I like that 1pyscript.write and do Element().write or display right? pyscript.write and do Element().write or display right? pyscript and import pyscript are two different things is enough to block.
We should try to slowly clean the mess, not introducing new one pyscript and import pyscript are two different things is enough to block.
We should try to slowly clean the mess, not introducing new one import pyscript isn't possible at all until #961 is merged; that PR is now doing several things at once )
I need to check what’s up with the error codes I remember changing then (maybe I forgot to push ))
I need to check what’s up with the error codes I remember changing then (maybe I forgot to push ) But thanks! I'll fix the conflicts But thanks! I'll fix the conflicts 
1
Yesterday & Today - 11/28/2022 1
I’ll try to take a look tomorrow in the morning Yesterday & Today - 11/29/2022 From a Euro perspective your today is our tomorrow (and I have a small brain, so did a double take before realising what was going on). 1 From a Euro perspective your today is our tomorrow (and I have a small brain, so did a double take before realising what was going on). 1 1
For the module import, I did noticed that while on the chrome debugger everything worked fine once pyodide finished loading and even on the failing test doing 
pyscript.runtime.interpreter.FS.open("pkg/a.py") worked 1test_paths_from_packages as my guinea pig )
sys.meta_path cache invalidation; going to open a WIP PR to test against CI, but passing locally 10 times in a row 1sys.meta_path cache invalidation; going to open a WIP PR to test against CI, but passing locally 10 times in a row 
you are a wizard CI seems happy ): https://www.ted.com/talks/alan_kay_a_powerful_idea_about_ideas
runtime.run instead? interpreter unless absolutely necessary 1 2
️ Jeff mentioned that he ran the failing tests multiple times and they all passed.
CI seems happy too 1self.element.source seems to be underfined Yesterday & Today - 11/30/2022 Element long term, or if we want something better Element to be a nice way to interact with the elements in the page, do you have any ideas of alternatives?Element("foo") is nice and short 1Element("foo") is nice and short Element("foo") current APIpage.foopydom.page.foopydocument.fooapp.fooapp.elements.fooapp.elements.get('foo')
️ ️ 1jasmine @ntoll ? just curious 
1 1 3
I'm excited
1
<py-xmas> 1 1
I'm off for a few hours for german class
(as opposed to being interpreted as cpython/pyodide do)
1n to run it multiple times:
@pytest.mark.parametrize('n', range(20))
def test_paths_from_packages(self, n):
...
And then use pytest -v --tb=no for a nice summary:
1importlib.invalidate_cache
(as opposed to being interpreted as cpython/pyodide do) 1importlib.invalidate_cache loadPyodide calls runPython internally: https://github.com/pyodide/pyodide/blob/e002c19af8906fd3b9199aa4bad6fe727d0b0451/src/js/pyodide.ts#L374importlib perspective..
We should definitely add the big file one though 2 1Date object with a mock object that replays the time stamps from the native session and see how the behavior changes. Though this would be a bit complicated to get right <py-repl> to behave like before (output contextual to the cell I execute)? output is gone for py-script tags... is it also gone for repl?
output just not working at all output now only affects the values returned from expressions; it has no affect on where your stdout goesoutput now only affects the values returned from expressions; it has no affect on where your stdout goes Lemee try againprint for instance? <py-repl id="my-repl" auto-generate="true" output="my-output"> </py-repl>
<div style="background-color: #f0ab3c; height: 12rem;" id="my-output"></div>
Then putting the following in the repl should put "42" into the orange div:
>>> x = 42
>>> x
But this will put it in a <py-terminal>
>>> x = 42
>>> print(x)import this and I'd like it to output on a div of choice. import this in a div of your choice Hmmmmm
sorry Fabio
maybe it's late for me
lmk if I can help smh!stdout should go to the same location as the results of expressions, which are placed according to display() rules (next-sibling of current element)
Which seems reasonable - it means either bypassing or rethinking the stdiomultiplixer from @antocuni 's #917.
Apologies if this got too in-the-weeds too fast - I've spent a lot of time in the commit history prepping the release blog post (5300 words and counting) 1 1
wild! thanks Jeffstdout should go to the same location as the results of expressions, which are placed according to display() rules (next-sibling of current element)
Which seems reasonable - it means either bypassing or rethinking the stdiomultiplixer from @antocuni 's #917.
Apologies if this got too in-the-weeds too fast - I've spent a lot of time in the commit history prepping the release blog post (5300 words and counting) 1 2♂️ display() and rethought stdout) would break them 1display() and rethought stdout) would break them
I'll also see what I can inlcude in my talk. TY!♂️ Yesterday & Today - 12/01/2022 
2022-12-01 Y&T or 01 Dec Status Update, or something like that
1
️ Kalendis Decembribus MMXXII A.D. 2 1[[fetch]] section in the draft blog post, the snippet for
Multiple Fetch Configs
has with open('data.csv', 'rb') as fp: while it should be sensordata.csv 1♂️ so display(<br>) prints the literal characters: it's missing quotes around <br>, should be display("<br>")for loopso display(<br>) prints the literal characters: it's missing quotes around <br>, should be display("<br>")for loop♂️ const { loadPyodide } = require("pyodide");
function busyWait(ms) {
let start = new Date().valueOf();
while( (new Date()).valueOf() - start < ms)
;
}
async function main() {
let pyodide = await loadPyodide({
indexURL: "./node_modules/pyodide",
});
console.log("Pyodide version: ", pyodide.version);
try {
pyodide.FS.writeFile('bar0.py', 'x = 0');
pyodide.runPython('import bar0; print(bar0.x)');
// if you uncomment EITHER line, it works
//pyodide.runPython('import importlib; importlib.invalidate_caches()');
//busyWait(1000);
pyodide.FS.writeFile('bar1.py', 'x = 1');
pyodide.runPython('import bar1; print(bar1.x)');
}
catch(e) {
console.log(e);
}
}
main(); 1
Bug This might be a duplicate of #737, but I'm opening a new issue because I'm seeing something very weird. I'm using Pyodide 0.21.3. To Reproduce Consider the following s...
runtime.invalidate_module_path_cache() is a proper and reliable workaround ♂️♂️ 1
<!DOCTYPE html>
<html>
<body>
<?php
$x = "Hello world!";
echo $x;
?>
</body>
</html>
1 1 1 Yesterday & Today - 12/02/2022 
const { loadPyodide } = require("pyodide");
function busyWait(ms) {
let start = new Date().valueOf();
while( (new Date()).valueOf() - start < ms)
;
}
async function main() {
let pyodide = await loadPyodide({
indexURL: "./node_modules/pyodide",
});
console.log("Pyodide version: ", pyodide.version);
try {
pyodide.FS.writeFile('bar0.py', 'x = 0');
pyodide.runPython('import bar0; print(bar0.x)');
// if you uncomment EITHER line, it works
//pyodide.runPython('import importlib; importlib.invalidate_caches()');
//busyWait(1000);
pyodide.FS.writeFile('bar1.py', 'x = 1');
pyodide.runPython('import bar1; print(bar1.x)');
}
catch(e) {
console.log(e);
}
}
main(); 
6 4 1 1 Yesterday & Today - 12/05/2022 tests/integration/*.py, they don't use jest.
I agree that jest is horrible debugger line
1fetch docs 
2 
1 1 Yesterday & Today - 12/06/2022 1 2 1 ) 1
Yesterday & Today - 12/07/2022 doing this once things are loading is a good idea Yesterday & Today - 12/08/2022 
display correctly with matplotlib
1 1display with matplotlib in the meanwhile. 1 1
1 1
PyScript 2022.12.1.RC1
https://pyscript.net/snapshots/2022.12.1.RC1/
1
fetch too 1 1
1make test-integration, I get a consistent failure in TestConfig.test_runtime_config
FileNotFoundError: [Errno 2] No such file or directory: '/var/folders/c1/br776hbd0llfk_l01ckngqvm0000gp/T/tmp92ipi5ml/pyodide-build-0.20.0.tar.bz2' make test-integration, I get a consistent failure in TestConfig.test_runtime_config
FileNotFoundError: [Errno 2] No such file or directory: '/var/folders/c1/br776hbd0llfk_l01ckngqvm0000gp/T/tmp92ipi5ml/pyodide-build-0.20.0.tar.bz2' pytest cacherm -rf .pytest_cache/ inside pyscriptjs directory 1 1
2 
2panel from 13.1 -> 14.1) is up for a quick review: https://github.com/pyscript/pyscript/pull/1032panel from 13.1 -> 14.1) is up for a quick review: https://github.com/pyscript/pyscript/pull/1032 1
3 1 1 Yesterday & Today - 12/09/2022
https://notebook.basthon.fr/?from=examples/python3-pythontutor.ipynb
https://pythontutor.com
1matplotlib in pyscript vs jupyterlite and consequences of merging these two approaches. And this also touches on switching to AGG (this discussion will come up when we start talking about web-workers) : https://github.com/pyscript/pyscript/issues/983#issuecomment-1343948823
just mentioning here 
1
1 
jobs I've ever had to do... ) 1
Let's focus on PyScript 
1 1 1 
1dom manipulation/interaction API, what tools/libraries/APIs would you take inspiration from? I.e., I think playwright api to navigate and access elements is pretty nice... dom manipulation/interaction API, what tools/libraries/APIs would you take inspiration from? I.e., I think playwright api to navigate and access elements is pretty nice... dom manipulation/interaction API, what tools/libraries/APIs would you take inspiration from? I.e., I think playwright api to navigate and access elements is pretty nice... lxml in the past quite a bit. Some aspects are but most of it definitely does not feel very Pythonic. 1 @antocuni over on the Python discord. Composing an answer right now. Brett and I have been talking about PEP-11 and getting CPython (Christian's work) to tier 2. 1 on web-assembly as basically just another CPU target. OS - ...
Yesterday & Today - 12/12/2022 ) 2
2 2 2 1 1 2 3 2 1 Yesterday & Today - 12/13/2022 such a bummer 
2 1 2 2 1
https://t.co/hlVN5eUKP1
1 2
Yesterday & Today - 12/14/2022 .
For comparison, have a look at kivy, and in particular its Pong example: here, you define a 
PongBall widget with a property pos and you use this property in the UI file. When you update pos in the code, the UI is automatically updated
https://kivy.org/doc/stable/tutorials/pong.html#pongball-class.
For comparison, have a look at kivy, and in particular its Pong example: here, you define a PongBall widget with a property pos and you use this property in the UI file. When you update pos in the code, the UI is automatically updated
https://kivy.org/doc/stable/tutorials/pong.html#pongball-class 
️ ️ 2 1 1 Yesterday & Today - 12/15/2022 
Yesterday & Today - 12/16/2022 <form id="inputForm">
<label for="english">Translate English 🇬 🇧 to Pirate speak 🏴<200d>☠️:</label>
<input type="text" name="english" id="english"
placeholder="Type English here..." />
<input type="submit" value="Translate"/>
<div id="output"></div>
</form>
<py-script>
import arrr
from polyplug import plug, update, receive
@plug("#inputForm", "submit")
def handle_form(event):
"""
Take the English input from the form, turn it into Pirate talk and update
the DOM with the result.
"""
english = event.target.find("#english").value
pirate_text = arrr.translate(english)
output = event.target.find("#output")
output.innerHTML = f"<p>{pirate_text}</p>"
update("#output", output)
</py-script>
<div id="arrr"/>
We get....
1 Yesterday & Today - 12/19/2022 PyTutor plugin that basically automates the "show code" feature we have in our examples..
I'm trying to check but we didn't have any tests for that feature, did we? 1 1
1 ♂️ .
Too late, sorry for that.
Time to go sleeping 
Yesterday & Today - 12/20/2022 plugins to fetch a python plugin, they get loaded correctly because we are using the same fetch logic as py-config test_default: 5.29 seconds
test_gh_pages_0_21_3: 6.40 seconds
test_gh_pages_zipimport2: 4.30 seconds
But I have a very unstable 4g connection, so I don't know how real these numbers are test_default: 2.52s (without --no-fake-server between 3-4s)
test_gh_pages_0_21_3: 2.51 ( without --no-fake-server 3.9-4s)
test_gh_pages_zipimport2: 2-3s (without --no-fake-server 2.88s)test_default: 2.52s (without --no-fake-server between 3-4s)
test_gh_pages_0_21_3: 2.51 ( without --no-fake-server 3.9-4s)
test_gh_pages_zipimport2: 2-3s (without --no-fake-server 2.88s) --no-fake-server 
(the checks done inside test_zz_example are by far not enough, IMHO) 
(the checks done inside test_zz_example are by far not enough, IMHO) ♂️ ♂️ 
️ about the plugins stability. This is experimental work for now. I'm still figuring out a ton of things and don't want to spend more time on readapting tests then on the logic and design itself. Once we figure a nice way to actually test ALL python plugins and some other design choices around plugins we add tests. That should come before the next release. 1
Alex also helped me with the 
async import which solved a lot of problems. The javascript plugin will have to look like any Plugin and contain the same shape, but currently only afterSetup is being called.
We also can't do extend Plugin since the module has no notion of pyscript.plugin.Plugin. There's a lot of rough edges, but at least the code that fetches the plugin isn't too ugly
Hopefully by tomorrow I'll have a PR up 1 Yesterday & Today - 12/21/2022 1 print('hello world'):
https://antocuni.github.io/files/pyodide/misc/benchmark_pyodide_startup.html 1
Alex also helped me with the async import which solved a lot of problems. The javascript plugin will have to look like any Plugin and contain the same shape, but currently only afterSetup is being called.
We also can't do extend Plugin since the module has no notion of pyscript.plugin.Plugin. There's a lot of rough edges, but at least the code that fetches the plugin isn't too ugly
Hopefully by tomorrow I'll have a PR up but the pyscript.plugin.Plugin above got me confused (not sure why because we don't have that on the Python namespace anyway lolpyscriptjs/plugin.Plugin 1 2 1 Yesterday & Today - 12/22/2022 Yesterday & Today - 12/23/2022 Yesterday & Today - 12/26/2022 Yesterday & Today - 12/27/2022 Yesterday & Today - 12/28/2022 Yesterday & Today - 12/29/2022 Yesterday & Today - 12/29/2022 Yesterday & Today - 12/30/2022 Yesterday & Today - 01/02/2023 runtimes.ts? I'm renaming 'runtimes' to 'interpreter' and I wonder if perhaps it should move to main instead? Yesterday & Today - 01/03/2023 1 5
https://github.com/pyscript/pyscript/discussions/1091 1 Yesterday & Today - 01/04/2023 4 3create_proxy deprecation warning. Will investigate and get a follow-up PR in. 3name = "Bla"
instead of
[project]
name = "Bla[project] , where we move all the configs that we allow on root now? Or we are happy with the current status?
I personally feel like it'd be a positive change that would be more aligned with the standard in Python-land and also keep things more organized... but don't have a super strong opinion about it.
Thoughts? create_proxy deprecation warning. Will investigate and get a follow-up PR in. 1 Yesterday & Today - 01/05/2023 3 Yesterday & Today - 01/06/2023 Yesterday & Today - 01/09/2023 1 Yesterday & Today - 01/10/2023 Yesterday & Today - 01/11/2023 )
https://www.gitbar.it/ 4 1)
https://www.gitbar.it/ 
1 Yesterday & Today - 01/12/2023 2 Yesterday & Today - 01/13/2023 Yesterday & Today - 01/16/2023 2
1 1
5 Yesterday & Today - 01/17/2023 1 1 1
Looks like all the proposals were accepted! I'm shocked!! Or did I miss someone?
1♂️ Yesterday & Today - 01/18/2023 1 1 1 1 Yesterday & Today - 01/19/2023 1 1 1
must be so late
must be so late Yesterday & Today - 01/20/2023 
f'<div id={el_id!r}>'
Instead of:
f'<div id="{el_id}">'
I've opened up https://github.com/pyscript/pyscript/pull/1121 to simply ignore this recommendation mostly because I think the second versions is more explicit.
if we all agree with this change, we need to merge it soon so we can merge other PRs.
Otherwise we can always update all the f-strings to use the !r 1 Yesterday & Today - 01/23/2023 Yesterday & Today - 01/24/2023 1 Yesterday & Today - 01/25/2023 Yesterday & Today - 01/26/2023 Yesterday & Today - 01/27/2023
heard they're terrible Yesterday & Today - 01/30/2023 Yesterday & Today - 01/31/2023 1globalThis also works in web workers, which don't have window 1globalThis also behaves differently inside iframes 1 Yesterday & Today - 02/01/2023 Perhaps that's the problem, we believe the "perfect repeatable science boxes" PR and forget the "real" world is messy. As the old joke says, "In theory there is no difference between theory and practice, but in practice there is." 1 Yesterday & Today - 02/02/2023 1
1 1 Yesterday & Today - 02/03/2023 1
1 1 2 (IIRC Damian explain this to me a couple of years ago... they have severe resource constraints)
1 (IIRC Damian explain this to me a couple of years ago... they have severe resource constraints) Yesterday & Today - 02/06/2023 Yesterday & Today - 02/06/2023 
output attribute for REPLs) is finally ready for review. Sorry it took an age, obviously no rush to get it integrated
https://github.com/pyscript/pyscript/pull/1106 1
what was her feedback for us?
what was her feedback for us? 1 I'd be happy to help out with it I'll wait for the pyodide fork and then fork off that fork Yesterday & Today - 02/07/2023 AttributeError: 'TestExamples' object has no attribute '_js_errors'
when I try to run it with pytest -vv tests/integration/ --log-cli-level=warning
but! the funny thing is, if I try to run with make test
everything works perfectly
anyone reproducing it or seeing similar? AttributeError: 'TestExamples' object has no attribute '_js_errors'
when I try to run it with pytest -vv tests/integration/ --log-cli-level=warning
but! the funny thing is, if I try to run with make test
everything works perfectly
anyone reproducing it or seeing similar? 4c30359 (i.e., the latest commit in main as of now) and they seem to work fine weirdly written
I forgot to install pytest-playwright and that was the problem
opening a PR with docs on how to install the test env
(still not sure how it was running normally with make test...) Yesterday & Today - 02/08/2023 ♂️ 2 Yesterday & Today - 02/09/2023 Interpreter class: we need to make it "webworker-friendly", which means (at least):
1. kill interpreter.globals
2. kill all the usages of pyimport
3. make interpreter.run async again 
1 Yesterday & Today - 02/10/2023 1
I'm looking at the diff in schema and will start update it Thanks for the pydantic class!
Also interestingly it seems that the new version returns a completely different output 
Yesterday & Today - 02/13/2023
And for the upgrade I've been doing the same, trial and error and trying to figure out what was blowing up haha
And for the upgrade I've been doing the same, trial and error and trying to figure out what was blowing up haha Yesterday & Today - 02/14/2023 Yesterday & Today - 02/15/2023 Yesterday & Today - 02/16/2023 
asyncio.gather etc and don't have to reimplement these ourselves.
If we had a syncifying decorator we can always do dispatch with await and then call the function synchronously outside:
@syncify
async def do_the_thing(selector, new_value):
await main_thread.querySelector(".blah").setattr("innerHTML", new_value)
await main_thread.fetch("a thing")
def g():
do_the_thing(".blah1", "blah_value")
do_the_thing(".blah2", "blah_value")
The definition of syncify would look like:
def syncify(func):
def wrapper(*args):
orig_loop = asyncio.get_event_loop()
asyncio.set_event_loop(MainThreadDispatchLoop())
try:
return asyncio.run(func(*args))
finally:
asyncio.set_event_loop(orig_loop)
return wrapper . What is the problem that you are trying to solve with this?. What is the problem that you are trying to solve with this? 1 1 1plt.show().
We should fix it
plt.show() calls display(plt).
Or maybe we could add a new "pyscript backend" to matplotlib, so that .show() calls display() automatically 1.setattr() syntax. Maybe it's possible to achieve something like this:
from js import document # this is "sync" be default
async def myfunc():
doc = Lazy(document)
foo = doc.getElementById("foo")
foo.innerHTML = "hello"
await foo
# or maybe await doc; I guess they probably would be equivalent.setattr() syntax. Maybe it's possible to achieve something like this:
from js import document # this is "sync" be default
async def myfunc():
doc = Lazy(document)
foo = doc.getElementById("foo")
foo.innerHTML = "hello"
await foo
# or maybe await doc; I guess they probably would be equivalent Lazy(some_js_object) returns an object which records all the operations done on it (and on the results of these ops) and then sends them to the main thread when you do await 1plt.show() calls display(plt).
Or maybe we could add a new "pyscript backend" to matplotlib, so that .show() calls display() automatically Lazy(some_js_object) returns an object which records all the operations done on it (and on the results of these ops) and then sends them to the main thread when you do await await foo was more expressive. I think it's not immediate to a noob user what they are waiting for... Maybe being more explicit and awaiting a sync method is more expressive even if a level deeper? Yesterday & Today - 02/17/2023 .setattr() syntax. Maybe it's possible to achieve something like this:
from js import document # this is "sync" be default
async def myfunc():
doc = Lazy(document)
foo = doc.getElementById("foo")
foo.innerHTML = "hello"
await foo
# or maybe await doc; I guess they probably would be equivalent foo = await doc.getElementById("foo")
await doc.someFunc(foo)
2. Your proposal, perfectly possible to implement afaict
foo = doc.getElementById("foo")
foo.someAttr = "blah"
await foo # returns a proxy for foo,
# also assigns someAttr as a side effect
3. I'm not sure what this would do:
foo = doc.getElementById("foo")
await doc,someFunc(foo)
await foo
Does it execute document.getElementById twice? Once? If only once, how does that work? Do the objects foo and doc.someFunc(foo) know they are tied together in some way? await it twice.
E.g., if you start with Lazy(document), you can do all operations that you want, and they would be executed only when you use await. Once you have done that, you cannot call await on it (or on any other intermediate object) again.
That said, I'm not sure it's a good idea, I'm mostly thinking aloud Yesterday & Today - 02/20/2023 Yesterday & Today - 02/21/2023 Yesterday & Today - 02/22/2023 Yesterday & Today - 02/23/2023 Yesterday & Today - 02/24/2023 Yesterday & Today - 02/24/2023 1 Yesterday & Today - 02/27/2023 Yesterday & Today - 02/27/2023 SyncLink proxies and PyProxys split-interpreter has a PR open, and it should be ready I guess -- https://github.com/pyscript/pyscript/pull/1218 Yesterday & Today - 02/28/2023 
CONTRIBUTING file:
Contributors are invited to be maintainers of the project by demonstrating good decision making in their contributions, a commitment to the goals of the project, and consistent adherence to the code of conduct. New maintainers are invited by a 3/4 vote of the existing maintainers.
So, let's vote! I'm obviously +1 7 Yesterday & Today - 03/01/2023 Yesterday & Today - 03/01/2023 api.ts, pyodide.ts, and pyodide.umd.ts so we export a list of three configs.
https://github.com/pyodide/pyodide/blob/main/src/js/rollup.config.mjs
Yesterday & Today - 03/02/2023 2 Yesterday & Today - 03/03/2023 Yesterday & Today - 03/03/2023 pys-onClick should be removed not in the next version of pyscript, but the one the comes next
because we're adding the deprecation warnings right now? (they weren't there in the 2022.12.1 version)
correct @FabioRosado ? 
1 Yesterday & Today - 03/06/2023 
...
I have a branch of pyscript in which pyodide actually runs in a web worker
It's a complete hack and it's based on the excellent synclink branch which @madhurt and @hood started in Lisbon, but it's complete enough that test_pyscript_hello passes.
It's very likely that this branch will never be merged and that I'll have to re-do it on top of the synclink branch once it's done, but it's cool to see that it actually works.
https://github.com/pyscript/pyscript/tree/antocuni/play-with-workers 3 1 1 Yesterday & Today - 03/07/2023 Yesterday & Today - 03/08/2023 <py-script> tags dynamically (or, equivalently, that if you do they are not executed).
Originally I introduced that test because it was easy to implement and mimics the behaviour of <script>, but I'm not 100% of what are good use cases for it
<py-script> tags dynamically (or, equivalently, that if you do they are not executed).
Originally I introduced that test because it was easy to implement and mimics the behaviour of <script>, but I'm not 100% of what are good use cases for it <py-script> tags in the DOM.
I see two scenarios:
1. I want to execute python code from JS: currently you can add a <py-script> tag, but we can imagine providing an API to just call a python function or a snippet of python code. Something like pyscript.runPython("....").
2. I want to execute python code from python: in that case, just call a function and/or import a module.
Am I missing anything? 
What's New section isn't finished, obviously, will need to put a button on that tonight Yesterday & Today - 03/09/2023 py-repl on the synclink branch since they are flaky and have lots of timing issues.
owing to that, I randomly tried a test called test_run_clears_previous_output
according it the title, there should be only 1 cell, since we clear the previous one and run it again
but using the --dev, it seems like there are 2 cells
This can also be seen in other related tests like test_python_exception_after_previous_output and test_hide_previous_error_after_successful_run
Let me know if this is expected aka I am interpreting this incorrectly OR if the test itself is incorrect?
Yesterday & Today - 03/10/2023 
<script> tag to the DOM?
I tried this:
let mod = await import("https://cdn.jsdelivr.net/pyodide/v0.22.1/full/pyodide.js");
However, the resulting mod seems to be empty:
> mod
Module {Symbol(Symbol.toStringTag): 'Module'} await import() the mod is empty, but loadPyodide is available in the global namespace. Is this expected? 
3 2 2
https://pyscript.net/snapshots/2023.03.1.RC2/pyscript.js) and the all were working fine.
Ship it ? 3
3 3 2
3
2 1
1 1<script> tag to the DOM?
I tried this:
let mod = await import("https://cdn.jsdelivr.net/pyodide/v0.22.1/full/pyodide.js");
However, the resulting mod seems to be empty:
> mod
Module {Symbol(Symbol.toStringTag): 'Module'} mod = await import("https://cdn.jsdelivr.net/pyodide/v0.22.1/full/pyodide.mjs"); Yesterday & Today - 03/13/2023 Yesterday & Today - 03/14/2023 
1 Yesterday & Today - 03/15/2023 synclink branch, a lot of examples fail due to not being able to create an instance of py-tutor plugin. Will push some more changes though. And will collaborate with @hood again today.... Yesterday & Today - 03/16/2023 Yesterday & Today - 03/17/2023 Yesterday & Today - 03/17/2023 Yesterday & Today - 03/20/2023 Yesterday & Today - 03/20/2023 [[fetch]] adds a table (dictionary) to the array “fetch”; similarly with [[interpreters]]. https://toml.io/en/v1.0.0#array-of-tables 1Uncaught (in promise) TypeError: singleFile.endsWith is not a function
and when I use the equivalent code in toml, the plugin is just successfully imported singleFile.src.endsWith should work in that case, if I understand what's happening there 1 1runButton id and it doesn't need a label element around that doesn't necessarily point at that button neither ... thoughts? P.S. within ShadowDOM IDs can be unique (in there) but because we're not placing stuff within SD but also we're appending clones multiple times, I suggest we just move the label to the button and call it a day
5 2.js file to orchestrate that so that we have one command to watch and serve examples, as npm run dev does now? For the time being I am just using npm run server which is a python -m http.server and a npm run esbuild-dev which watches files ... beside having dependencies simplified, it takes ~50ms to build either minified or not with source maps 
.js file to orchestrate that so that we have one command to watch and serve examples, as npm run dev does now? For the time being I am just using npm run server which is a python -m http.server and a npm run esbuild-dev which watches files ... beside having dependencies simplified, it takes ~50ms to build either minified or not with source maps 
.js file to orchestrate that so that we have one command to watch and serve examples, as npm run dev does now? For the time being I am just using npm run server which is a python -m http.server and a npm run esbuild-dev which watches files ... beside having dependencies simplified, it takes ~50ms to build either minified or not with source maps
cp or any common *nix operations within npm scripts and call it a day ... to which I'd be OK as I don't think Windows users would ever use Makefile anyway, or better this doesn't seem to be a concern until now, but WSL for Windows would still work perfectly fine.cp or any common *nix operations within npm scripts and call it a day ... to which I'd be OK as I don't think Windows users would ever use Makefile anyway, or better this doesn't seem to be a concern until now, but WSL for Windows would still work perfectly fine. npm run esbuild or npm run esbuild-dev to measure locally differences ... I think just the fact we'll need one dependency instead of dozens is a pros, but maybe not everyone is experiencing the same speed improvement (and CI is not using it right now so I don't know about that neither)npm run esbuild or npm run esbuild-dev to measure locally differences ... I think just the fact we'll need one dependency instead of dozens is a pros, but maybe not everyone is experiencing the same speed improvement (and CI is not using it right now so I don't know about that neither) 1. That's awesome Yesterday & Today - 03/21/2023 
1
1
Yesterday & Today - 03/22/2023 
1 2 1 Yesterday & Today - 03/23/2023
And as soon as we have workers, we will be able to use it also for input()pdb.set_trace() 1) 3 1pyscript-write group so you both have wr rights once you accept :). 2 1 Yesterday & Today - 03/24/2023 pyscript-write group so you both have wr rights once you accept :). ... trying to figure out what's going on - solved 1@ts-ignore, I think we should disable that eslint rule. It doesn't make sense to complain about it (if we decide that it's fine to explicitly ignore typescript types for the next line, it means that we know what we are doing and eslint should just agree with us) 1beforePyReplExec and afterPyReplExec plugin hooks should be optional for js plugins to implement or not https://github.com/pyscript/pyscript/pull/1310
1any? 1
beforePyReplExec and afterPyReplExec plugin hooks should be optional for js plugins to implement or not https://github.com/pyscript/pyscript/pull/1310 1 1 src related to custom elements? 'cause I've just read about other src problems and I am not sure any ot these two is relatedsrc related to custom elements? 'cause I've just read about other src problems and I am not sure any ot these two is related ... thanks for supporting it, I never meant to take over you, just wanted to bring in some possible improvement around it
1 1 1
1 1 2 1Yesterday & Today thread? Yesterday & Today - 03/28/2023 Yesterday & Today thread? ♂️♂️ 1 Yesterday & Today - 03/29/2023 https://github.com/pyscript/pyscript/discussions/1326 
1 1pyscript/__init__.py into multiple files: one thing which I would really like is to have a pyscript._internal (or a similar name) for all the things which are needed by pyscript itself but should not be called by the users (e.g. the new _run_python). If we put them inside _internal, we can probably even remove the leading underscore from the individual functions 1<rant>
I hate squash&merge. It makes it very hard to develop a PR on top of another one, because when you merge it git creates a completely unrelated commit and you are no longer able to just git merge main
</rant> 1RemoteInterpreter <rant>
I hate squash&merge. It makes it very hard to develop a PR on top of another one, because when you merge it git creates a completely unrelated commit and you are no longer able to just git merge main
</rant> reply on the wrong message? Else it doesn't make sense export type worker_initialize worker_initialize lives in a completely separate file which is not even seen by the main bundle of pyscript
https://github.com/pyscript/pyscript/pull/1333/* eslint-disable @typescript-eslint/no-unused-vars */
const xPyRepl = customElements.define('py-repl', PyRepl); customElements.get('py-repl') would indeed return that if defined, nothing otherwise, but most importantly we have the PyRepl class right there in the scope so that xPyRepl should go indeed. Yesterday & Today - 03/30/2023 
3 1 1 1 1 grid is taken for a module to reuse in more than a project ... so I ended up with greedily but I don't like it too much neither grid is taken for a module to reuse in more than a project ... so I ended up with greedily but I don't like it too much neither 
1itsybitsyteenyweenyyellowpolkadot-grid is also free on npm 
2
<script type="module"> 1 31 Mar 2023: Yesterday & Today (catch you lovely folk next week)
4 4<script> to bootstrap discussion ... if nothing else to add we have bugs coming due current situation https://github.com/pyscript/pyscript/discussions/1326#discussioncomment-5492789
1<script> to bootstrap discussion ... if nothing else to add we have bugs coming due current situation https://github.com/pyscript/pyscript/discussions/1326#discussioncomment-5492789 . 03 Apr 2023: Yesterday & Today . 
main branch, seems like I cannot build pyscript
Specifically, I get this:
file:///Users/madhur/Desktop/Anaconda/pyscript/pyscriptjs/esbuild.mjs:5
import { cp, lstat, readdir } from 'fs/promises';
^^
SyntaxError: The requested module 'fs/promises' does not provide an export named 'cp'
at ModuleJob._instantiate (internal/modules/esm/module_job.js:124:21)
at async ModuleJob.run (internal/modules/esm/module_job.js:179:5)
at async Loader.import (internal/modules/esm/loader.js:178:24)
at async Object.loadESM (internal/process/esm_loader.js:68:5) 1v18.11.0 for node and 8.19.2 for npm
And now it's stuck at idealTree:pyscriptjs: sill idealTree buildDeps when I do npm install node_modules folders and then you do an npm ci 1node_modules folders and then you do an npm ci npm i (install) the ci uses the package-lock.json file to ensure everyone uses actually the very same modules and very same modules versions ... or at least that's the idea, while npm i might grab patches without asking ... worth checking out if next version of some dev/dependency might break our builds (not today though) 1<script type="py"> instead of <py-script> (worth it to me, but not for everyone to follow?), among all the details and improvements already shared, there is this particular reason that alone should make everyone think we did a mistake to use a layout element to define code data blocks ... maybe it helps as summary with demo you can test, images as results, and why that happens: https://github.com/pyscript/pyscript/discussions/1326#discussioncomment-5509027
1 04 Apr 2023: Yesterday & Today 04 Apr 2023: Yesterday & Today 
1 1
View raw logs I see the following.
I think it means that for some reason github did not even start my job, not that the tests themselves are stuck. Am I right?
pyscript.js and worker.js 1Test Reporter stuff which was added by @hood recently it's awesome .
Let me post a screenshot here so that everybody is aware that it exists. It shows you a nice summary of all passed/failed tests, and you can click on failed ones to see the output.
Live example:
https://github.com/pyscript/pyscript/pull/1363/checks?check_run_id=12514717064
05 Apr 2023: Yesterday & Today 
--dev, I discovered an unrelated problem with CORS headers.
On main, the problem manifests itself only with --dev, but as soon as we merge #1333 (webworkers) we will have the same problems also on main.
I have opened an issue:
https://github.com/pyscript/pyscript/issues/1365 https://github.com/pyscript/pyscript/pull/1292#issuecomment-1488157783 ... I've also checked for other issues and it looks like we really want to increase that timeout as we don't get to know in which hardware CI is running (AFAIK)
https://github.com/pyscript/pyscript/pull/1292#issuecomment-1488157783 ... I've also checked for other issues and it looks like we really want to increase that timeout as we don't get to know in which hardware CI is running (AFAIK) wait_for_pyscript for each test
2. a timeout argument to wait_for_pyscript so that we can change it on a test-by-test basis
3. an increased timeout for the problematic tests
Hopefully this should improve the situation: we will hopefully be able to set a smaller default timeout (because the current 30s is way too much for the vast majority of tests) and a higher one only for the tests which actually need it 4 06 Apr 2023: Yesterday & Today 1 07 Apr 2023: Yesterday & Today 
10 Apr 2023: Yesterday & Today 11 Apr 2023: Yesterday & Today 12 Apr 2023: Yesterday & Today 1 1 13 Apr 2023: Yesterday & Today 1 14 Apr 2023: Yesterday & Today 3 2 17 Apr 2023: Yesterday & Today 2 18 Apr 2023: Yesterday & Today 1
19 Apr 2023: Yesterday & Today <py-script> tag literally exactly what I would expect and explained when I've suggested to use <script type="py"> instead ... but of course we can ask everyone to use different chars to represent HTML and we could escape those chars on our side ... no other ideas beside using inverted tags for HTML/SVG content ... example:
def _repr_svg_(self):
return (
f'>svg height="{self.r*2}" width="{self.r*2}"<' +
f'>circle cx="{self.r}" cy="{self.r}" r="{self.r}" fill="red" /<>/svg<'
)
... we could fix that when parsing <py-script> tags then .. (this is a horrible idea anyway). 1<py-script> tag literally exactly what I would expect and explained when I've suggested to use <script type="py"> instead ... but of course we can ask everyone to use different chars to represent HTML and we could escape those chars on our side ... no other ideas beside using inverted tags for HTML/SVG content ... example:
def _repr_svg_(self):
return (
f'>svg height="{self.r*2}" width="{self.r*2}"<' +
f'>circle cx="{self.r}" cy="{self.r}" r="{self.r}" fill="red" /<>/svg<'
)
... we could fix that when parsing <py-script> tags then .. (this is a horrible idea anyway). <!tag ... /!> which should not bother the parser as that's an invalid tag name ... <[tag ... /]> might do too as long as those chars don't interfere with both the HTML parser and Python ... the alternative is to use <script> and stop this "madness" sooner than later 20 Apr 2023: Yesterday & Today attachShadow({mode: 'open'}) automatically augment the custom element with a shadowRoot property making our own shadow property redundant -> https://developer.mozilla.org/en-US/docs/Web/API/Element/attachShadow#parameters ... as you're all busy at the PyCon, do you mind if I suggest a "chore" PR that removes such redundant property from our custom elements? make test and I think that's friction for contributors, specially since we ask if all tests passed locally whenever a PR is filed ... should we tackle this after the conf? I find this issue quite problematic make test and I think that's friction for contributors, specially since we ask if all tests passed locally whenever a PR is filed ... should we tackle this after the conf? I find this issue quite problematic but I am happy to share my specs ... it's a pretty beefy laptop last gen too: Lenovo ThinkPad Z13 top specs (Ryzen 7, 32GB RAM, 1TB nvme and so on). but I am happy to share my specs ... it's a pretty beefy laptop last gen too: Lenovo ThinkPad Z13 top specs (Ryzen 7, 32GB RAM, 1TB nvme and so on). 
but I am happy to share my specs ... it's a pretty beefy laptop last gen too: Lenovo ThinkPad Z13 top specs (Ryzen 7, 32GB RAM, 1TB nvme and so on).
But if the the tests timeouts, just make a PR to increase them 1 21 Apr 2023: Yesterday & Today 24 Apr 2023: Yesterday & Today 25 Apr 2023: Yesterday & Today
But if the the tests timeouts, just make a PR to increase them 
26 Apr 2023: Yesterday & Today 27 Apr 2023: Yesterday & Today
In the callback that creates the MutationObserver, are we looping once at page load? We have two loops together and I wonder if it will cause a slight delay if the index page is huge?
(Mind you, if its huge and has a bunch of packages the load will be higher anyway haha)
In the callback that creates the MutationObserver, are we looping once at page load? We have two loops together and I wonder if it will cause a slight delay if the index page is huge?
(Mind you, if its huge and has a bunch of packages the load will be higher anyway haha) connectedCallback there, and that triggers any time the node is moved around the DOM. I've mentioned that at the end of the description of my PR.connectedCallback there, and that triggers any time the node is moved around the DOM. I've mentioned that at the end of the description of my PR. I think we still dont dismiss reviews when we push new commits, so feel free to merge once CI is green 28 Apr 2023: Yesterday & Today 01 May 2023: Yesterday & Today 01 May 2023: Yesterday & Today 02 May 2023: Yesterday & Today 
03 May 2023: Yesterday & Today
If you need a repro, it's here: https://4897a01d-e56f-47fa-814f-70c7c1e11730.pyscriptapps.com/5412a8f5-0787-42bb-92a2-9a582e7bc7f2/latest/
If you need a repro, it's here: https://4897a01d-e56f-47fa-814f-70c7c1e11730.pyscriptapps.com/5412a8f5-0787-42bb-92a2-9a582e7bc7f2/latest/ shadowRoot and so on ... all doable on the surface but it requires some refactoring 1 04 May 2023: Yesterday & Today PyWidget since we don't seem to use it anywhere?
Any thoughts? CC: @Fabio @Jeff Glass ? ️ ?
3 1PyWidget also renders the function createWidget as useless since we don't use it anywhere. Do we potentially see it being used? Or perhaps we can remove this too? Basically, we can delete the full file i.e. pyscriptjs/src/components/pywidget.ts
1<py-register-widget> in the wild. 1PyWidget since we don't seem to use it anywhere?
Any thoughts? CC: @Fabio @Jeff Glass ? 05 May 2023: Yesterday & Today 08 May 2023: Yesterday & Today 09 May 2023: Yesterday & Today 10 May 2023: Yesterday & Today 10 May 2023: Yesterday & Today 
11 May 2023: Yesterday & Today 
tests/integration/test_zz_examples.py::TestExamples::test_panel_kmeans is kinda broken and fails with ValueError: brush parameter was not found in list of parameters of class Selection 
test_panel_kmeans 12 May 2023: Yesterday & Today 15 May 2023: Yesterday & Today 1 1 16 May 2023: Yesterday & Today main.ts 17 May 2023: Yesterday & Today main 
1
18 May 2023: Yesterday & Today 19 May 2023: Yesterday & Today 22 May 2023: Yesterday & Today 22 May 2023: Yesterday & Today 23 May 2023: Yesterday & Today @when - we talked about being able to remove listeners. I thought what I heard was that the api would be calling remove_when() (or something) on the function, ie:
@when('click', 'button')
def foo():
print("fOO!")
... # later
foo.remove_when(...)
Leaving aside the specifics of the API for that function - is this what we want? 1
24 May 2023: Yesterday & Today pyscript --sw-proxy ./project-folder and we place a service worker (or we augment the present one) with all the logic ... something a bit "magic" like vite or other web dev oriented CLI tools do, nothing too articulated as a thought
25 May 2023: Yesterday & Today 
1
1 26 May 2023: Yesterday & Today 29 May 2023: Yesterday & Today 30 May 2023: Yesterday & Today 1 31 May 2023: Yesterday & Today 
1 01 Jun 2023: Yesterday & Today 02 Jun 2023: Yesterday & Today 1 1 05 Jun 2023: Yesterday & Today 06 Jun 2023: Yesterday & Today 06 Jun 2023: Yesterday & Today 07 Jun 2023: Yesterday & Today 08 Jun 2023: Yesterday & Today 
pyscript.core: if you check out pyscript-next and run npm run build, it fails with:
Error: ENOENT: no such file or directory, stat 'cjs'
I manually run mkdir cjs and now it builds fine .gitignore in that folder doesn't work ... I need to bring it into the parent folder ... which is annoying ... if we could have a whole pyscript-core repository instead it'd be much easier to setup that folder but not sure what others think about thispyscript.core: if you check out pyscript-next and run npm run build, it fails with:
Error: ENOENT: no such file or directory, stat 'cjs'
I manually run mkdir cjs and now it builds fine 
[next] as prefix ... the rest is the same<py-config> (or equivalent), or it's required to use an external config file? py-script plugin to do so, otherwise yespy-script plugin to do so, otherwise yes 1 09 Jun 2023: Yesterday & Today 12 Jun 2023: Yesterday & Today 13 Jun 2023: Yesterday & Today 14 Jun 2023: Yesterday & Today 14 Jun 2023: Yesterday & Today 15 Jun 2023: Yesterday & Today 15 Jun 2023: Yesterday & Today 16 Jun 2023: Yesterday & Today 16 Jun 2023: Yesterday & Today 19 Jun 2023: Yesterday & Today 20 Jun 2023: Yesterday & Today 20 Jun 2023: Yesterday & Today 2023.01.01 (it was 2023.03.1) https://github.com/pyscript/pyscript/pull/1546 update Merged, thanks @FabioRosado ! 
congrats everyone!!)
Shall we go ahead and finally sort the pending issue around examples, plugins and other resources?
After discussions in https://github.com/pyscript/pyscript/discussions/1442 and the many chats during the core /pyscript fun meetings, I think we are landing at a place where we want to:
pyscript org called plugins where we collect all the pyscript plugins maintainer by the community (that include but is not limited to the ones we create and maintain). It's important to understand here that this does not mean that ALL plugins go there. For instance, we may want to keep a few core plugins that are shipped by default into core
21 Jun 2023: Yesterday & Today 22 Jun 2023: Yesterday & Today 23 Jun 2023: Yesterday & Today @pyscript/core documentation https://github.com/pyscript/pyscript/blob/a3bf2881e42592d4360eff132118446af426e5eb/pyscript.core/docs/README.md 
@pyscript/core documentation https://github.com/pyscript/pyscript/blob/a3bf2881e42592d4360eff132118446af426e5eb/pyscript.core/docs/README.md 1 26 Jun 2023: Yesterday & Today 27 Jun 2023: Yesterday & Today 
Announcement: PyScript NEXT (new core) Tech-Preview Release!!
After invaluable community feedback, many experiments in code, and careful cooperation with valued collaborators, we are thrilled to announce the tech preview of the new version of PyScript. This work marks a significant milestone in our journey towards PyScript's vision and what our community has been asking for.
Here are the highlights and key features of this groundbreaking technical Milestone, the Tech Preview Release:
Massive Size Reduction: We have optimized PyScript to offer a minimal download size. The core library is only ~10kb compressed and all the rest is downloaded only on demand. Combined with the very small footprint of MicroPython, you can have a fully working app in ~250kb, resulting in an astounding reduction of 98% compared to the previous version (~13.5Mb = ~11MB Pyodide + ~2.5MB Pyscript Classic). This makes PyScript more lightweight and faster to download with all the supported interpreters, ensuring a seamless user experience to the user.
Versatile Execution Options: With PyScript, you have the flexibility to choose the execution environment that best suits your needs. Whether you prefer executing your code on the main thread or leveraging the power of web workers for parallel processing, in a sync or async way, PyScript has got you covered.
Blazing Fast MicroPython Interpreter: The new PyScript core now incorporates a cutting-edge MicroPython interpreter, which delivers unparalleled speed and performance. Experience lightning-fast execution of your Python code in the browser environment. 28 Jun 2023: Yesterday & Today 29 Jun 2023: Yesterday & Today 30 Jun 2023: Yesterday & Today 30 Jun 2023: Yesterday & Today 03 Jul 2023: Yesterday & Today 04 Jul 2023: Yesterday & Today 05 Jul 2023: Yesterday & Today 06 Jul 2023: Yesterday & Today 06 Jul 2023: Yesterday & Today 07 Jul 2023: Yesterday & Today 07 Jul 2023: Yesterday & Today 10 Jul 2023: Yesterday & Today 11 Jul 2023: Yesterday & Today 11 Jul 2023: Yesterday & Today 12 Jul 2023: Yesterday & Today 12 Jul 2023: Yesterday & Today 
wasm-snip .a files 13 Jul 2023: Yesterday & Today 
14 Jul 2023: Yesterday & Today 14 Jul 2023: Yesterday & Today polyscript thing, not sure). So would be good to have an extra pair of eyes on it 17 Jul 2023: Yesterday & Today 18 Jul 2023: Yesterday & Today 19 Jul 2023: Yesterday & Today 19 Jul 2023: Yesterday & Today 20 Jul 2023: Yesterday & Today 20 Jul 2023: Yesterday & Today 21 Jul 2023: Yesterday & Today 24 Jul 2023: Yesterday & Today 24 Jul 2023: Yesterday & Today 25 Jul 2023: Yesterday & Today 
1 26 Jul 2023: Yesterday & Today 26 Jul 2023: Yesterday & Today 27 Jul 2023: Yesterday & Today 
2 28 Jul 2023: Yesterday & Today 28 Jul 2023: Yesterday & Today 31 Jul 2023: Yesterday & Today 31 Jul 2023: Yesterday & Today 01 Aug 2023: Yesterday & Today @pyscript/core on npm 02 Aug 2023: Yesterday & Today 03 Aug 2023: Yesterday & Today 
04 Aug 2023: Yesterday & Today 07 Aug 2023: Yesterday & Today 08 Aug 2023: Yesterday & Today 09 Aug 2023: Yesterday & Today 10 Aug 2023: Yesterday & Today 11 Aug 2023: Yesterday & Today 
1
14 Aug 2023: Yesterday & Today 15 Aug 2023: Yesterday & Today 16 Aug 2023: Yesterday & Today 16 Aug 2023: Yesterday & Today 17 Aug 2023: Yesterday & Today 18 Aug 2023: Yesterday & Today 18 Aug 2023: Yesterday & Today 21 Aug 2023: Yesterday & Today 21 Aug 2023: Yesterday & Today 22 Aug 2023: Yesterday & Today 23 Aug 2023: Yesterday & Today 24 Aug 2023: Yesterday & Today 24 Aug 2023: Yesterday & Today 25 Aug 2023: Yesterday & Today 
1 28 Aug 2023: Yesterday & Today 29 Aug 2023: Yesterday & Today 30 Aug 2023: Yesterday & Today 31 Aug 2023: Yesterday & Today 31 Aug 2023: Yesterday & Today time.sleep in workers 01 Sep 2023: Yesterday & Today )
So, better to focus on wrapping up any other things related to the release. 04 Sep 2023: Yesterday & Today 05 Sep 2023: Yesterday & Today newdocs/README.md file.
06 Sep 2023: Yesterday & Today 
IMPORTANT: Release Plan
As we are working to get things aligned and release, I've been thinking how do that in the most graceful, open and communicative way. Here's what I was thinking:
Ssince a lot has changed and there's a lot going on, I'd like to propose we release tomorrow or, the latest, Friday early (morning European time) but hold to announce until we are good with documentation and messaging (which might happen this week or may spill into next week). This way we keep the stressful part of the release out of the way (or at least we split them) and deliver without trying to rush to get all together. Thoughts?pydom under pyweb as a namespace package (idea here is that other 3rd party packages can augment the namespace package. (At least that's the idea that Antonio and I were entertaining...)
Honestly, I feel like the way pydom is currently organized is not great... Ideas?
Another question: We currently don't have @when and HTML. This PR somehow introduce them back but.... where do you think they should belong? pyscript module? pyweb?
07 Sep 2023: Yesterday & Today 08 Sep 2023: Yesterday & Today pydom and testing the release... the more a couple of things felt off (specially the naming and namespace of pydom). I've pairined with @tedpatrick and @Martin yesterday and today and we capture out thoughts in this discussion https://github.com/pyscript/pyscript/discussions/1687
Please check it out and let's discuss before we release. (I know it's a bit of last minute but I'd rather take a few more days than make a big mess )
11 Sep 2023: Yesterday & Today 12 Sep 2023: Yesterday & Today 13 Sep 2023: Yesterday & Today import pydom become from pyscript import dom? 14 Sep 2023: Yesterday & Today 
2 1 15 Sep 2023: Yesterday & Today 
Classic tag to issues the PyScript repo that relate purely to bugs or implementation details in PyScript Classic. These would be good candidates for re-examination or closure once we release. 18 Sep 2023: Yesterday & Today 18 Sep 2023: Yesterday & Today
https://blog.pyodide.org/posts/0.24-release/PyScript ready 19 Sep 2023: Yesterday & Today 19 Sep 2023: Yesterday & Today
I have created another calender which contains the PyScript public events:
2 1 20 Sep 2023: Yesterday & Today 20 Sep 2023: Yesterday & Today 2 21 Sep 2023: Yesterday & Today 21 Sep 2023: Yesterday & Today env attribute is not available in PyScript, yes? Or rather, is always set to the same value across all tags by this line?https://github.com/pyscript/pyscript/blob/287d0fa1afb2d395f9382147054c9fd7656a8e43/pyscript.core/src/core.js#L144 env is the most unique identifier we have and py-script is our env in polyscript, yes ... it's not exposed though, but no example or use case required otherwise ... we could expose it but I am not sure what's the purpose so that maybe we can do better?env should be out of the discussion in PyScript .... if you specify a different env it means you want a new pyodide bootstrap and we don't have any mechanism to map a <py-config> element to that desire needed ... please keep env out of the explanation as that's really more like a polyscript internal feature, exposed by need, rather than by usefulness, thanks! 
1
1 22 Sep 2023: Yesterday & Today 25 Sep 2023: Yesterday & Today 25 Sep 2023: Yesterday & Today <mpy-config> and <py-config> work as expected ... don't these? I wasn't aware of any revert there ... https://github.com/pyscript/pyscript/pull/1753 
CTRL-C, which Python is supposed to intercept and raise KeyboardInterrupt.
E.g., try to go to my console demo and press CTRL-C: you can see that KeyboardInterrupt is not raised:
https://antocuni.pyscriptapps.com/pyterminal-poc/latest/?console/ CTRL-C, which Python is supposed to intercept and raise KeyboardInterrupt.
E.g., try to go to my console demo and press CTRL-C: you can see that KeyboardInterrupt is not raised:
https://antocuni.pyscriptapps.com/pyterminal-poc/latest/?console/ ^C though which seems ... OK? is not that we can exit the terminal ... or should we? ^C though which seems ... OK? is not that we can exit the terminal ... or should we? 
files configuration option:
https://github.com/pyscript/pyscript/discussions/1755
1coincident internals to judge whether this can be implemented without changes, so I guess it's better to wait for @hood input 
chat or announcements as well when you get details, in case there are community members that can come 
1chat or announcements as well when you get details, in case there are community members that can come I have to admit I've been mostly occupied by (Anaconda) Slack - and this is my concerted effort to kill-off-Slack and do all the things on discord. Thanks for the heads up.
4 1
coincident internals to judge whether this can be implemented without changes, so I guess it's better to wait for @hood input Atomics.wait and blocks all activity. This has to be changed instead to have a timeout and periodically check for a keyboard interrupt. 1 26 Sep 2023: Yesterday & Today 
KeyboardInterrupt 2 3 27 Sep 2023: Yesterday & Today 
https://github.com/pyscript/pyscript/pull/1760.
This uses make test-integration, which runs tests sequentially.
In theory, we also have test-integration-parallel but it seems to be very flaky: many tests r...
test_report workflow not running 28 Sep 2023: Yesterday & Today 
♂️ I have a quick question:
Is there any support in the latest RC1 to program event handlers (e.g. on_click) with Python functions using MicroPython ?
https://pyscript.com/@joshualowe1002/config-with-micropython/latest
https://pyscript.com/@joshualowe1002/config-with-micropython/latest 29 Sep 2023: Yesterday & Today 29 Sep 2023: Yesterday & Today 02 Oct 2023: Yesterday & Today 02 Oct 2023: Yesterday & Today polyscript
https://joeyh.name/blog/entry/Haskell_webassembly_in_the_browser/
"Programming for the 0.0009%" 1 1polyscript
https://joeyh.name/blog/entry/Haskell_webassembly_in_the_browser/
"Programming for the 0.0009%" .wasm files but we don't have yet a use case for that, so I won't spend time on it 03 Oct 2023: Yesterday & Today pyscript-next ) to fix a few things in the display function (cc/ @antocuni @Jeff Glass ): https://github.com/pyscript/pyscript/pull/1784 3 1 
I'd like to organise a Zoom call where we can "get our story straight". I.e. we have the "stuff" in place to make this work, we just need to be able to explain ourselves in a "how to do plugins" sort of a way. I want to figure out what this story is... please reply with availability and comments in the 
2 04 Oct 2023: Yesterday & Today Makefile in the PyScript repo... that's another "classic" that needs updating. 
05 Oct 2023: Yesterday & Today Makefile 
1 06 Oct 2023: Yesterday & Today From the look of the SO thread, I guess this is all part of the asker's learning journey..? Also, I might be missing something, so I'm always open to suggestions. 1 09 Oct 2023: Yesterday & Today 
1 10 Oct 2023: Yesterday & Today 2 1 11 Oct 2023: Yesterday & Today 
1 1
<script type="pyc"> ... I see 
1 1 12 Oct 2023: Yesterday & Today 13 Oct 2023: Yesterday & Today 
2 1 16 Oct 2023: Yesterday & Today 17 Oct 2023: Yesterday & Today 17 Oct 2023: Yesterday & Today 
from string import Template throws in MicroPython ... please add this to the list of questions whenever you have some time .. you really missed out on today demos and no, nobody recorded, Hood was in the house too from string import Template throws in MicroPython ... please add this to the list of questions whenever you have some time .. you really missed out on today demos and no, nobody recorded, Hood was in the house too Template https://github.com/JeffersGlass/mp-templates 
src of an mpy tag https://github.com/pyscript/pyscript/issues/1804
18 Oct 2023: Yesterday & Today 19 Oct 2023: Yesterday & Today 20 Oct 2023: Yesterday & Today 20 Oct 2023: Yesterday & Today 23 Oct 2023: Yesterday & Today 24 Oct 2023: Yesterday & Today py-terminal (https://github.com/pyscript/pyscript/pull/1801) we should do a new release of PyScript next, and assuming after a few days of exercising it "in the wild" promote that version to final - et voila, we have finally released PyScript next. Thoughts and feedback in a please - and perhaps something to discuss on today's community call..?py-terminal (https://github.com/pyscript/pyscript/pull/1801) we should do a new release of PyScript next, and assuming after a few days of exercising it "in the wild" promote that version to final - et voila, we have finally released PyScript next. Thoughts and feedback in a please - and perhaps something to discuss on today's community call..? <py-terminal> Custom Element stuff works in the terminal ... is this OK or should I actually implement the terminal attribute instead? To me the current terminal works and it doesn't require to "change the world" around polyscript neither plus it's explicit when it runs from a worker and when it doesn't ... should we move forward with it? 2 25 Oct 2023: Yesterday & Today 26 Oct 2023: Yesterday & Today 27 Oct 2023: Yesterday & Today 30 Oct 2023: Yesterday & Today 31 Oct 2023: Yesterday & Today 
3
1
1 1 1
make setup should just install playwright, right..? make setup installation for playwright because "reasons" so the quickest solution is probably to cd pyscript.core && npx playwright install && npm run build 01 Nov 2023: Yesterday & Today 4 3 1
3
1 1
1
1 02 Nov 2023: Yesterday & Today 03 Nov 2023: Yesterday & Today 
autogenerate from classic) that we could add as well, but I think we should try to nail down the core behaviors before we layer more features in. 06 Nov 2023: Yesterday & Today pydom section in the docs has been added and just need to be deployed. +1 on the plugin section, which I think we already have some content in the many READMEs around ( @Andrea Giammarchi would you mind taking care of this tomorrow? ) but the docs are decouple from the release itself so we can also correct after.
pydom section in the docs has been added and just need to be deployed. +1 on the plugin section, which I think we already have some content in the many READMEs around ( @Andrea Giammarchi would you mind taking care of this tomorrow? ) but the docs are decouple from the release itself so we can also correct after.
07 Nov 2023: Yesterday & Today ) so may not be able to attend the full call this morning. Can you guys drive?
README.md still says to link to /latest... we can update that to recommend the most recent actual version as soon as we release @ntoll docs/changelog.md ? 08 Nov 2023: Yesterday & Today 08 Nov 2023: Yesterday & Today 
4 2 1 1 1/latest . That PR wasn't merged before the release....
I don't think it's a big deal but it is a missed opportunity when as we agreed
1 1
1 is very distracting since it looks like a fart emoji? is very distracting since it looks like a fart emoji? 3 1 09 Nov 2023: Yesterday & Today 10 Nov 2023: Yesterday & Today 
/snapshots and non-RC under /releases. They should all go under /releases 1 1 1 13 Nov 2023: Yesterday & Today 14 Nov 2023: Yesterday & Today ) and add anything I'm missing? Andrea, can you also check the actual hooks? I did a quick copy of some of the Polyscript hooks but didn't fully verify the diffs...
https://github.com/pyscript/docs/pull/48
) and add anything I'm missing? Andrea, can you also check the actual hooks? I did a quick copy of some of the Polyscript hooks but didn't fully verify the diffs...
https://github.com/pyscript/docs/pull/48 import { hooks } from "<path_to_pyscript>/core.js"; come from? I looked at existing docs also did a github search and there are no such example that shows me how I get path_to_pyscript. Also it's unclear to me, is it only part of javascript modules? I would prefer something as close to working code as possible. 15 Nov 2023: Yesterday & Today import { hooks } from "<path_to_pyscript>/core.js"; come from? I looked at existing docs also did a github search and there are no such example that shows me how I get path_to_pyscript. Also it's unclear to me, is it only part of javascript modules? I would prefer something as close to working code as possible. <script type="module" src="//cdn/@pyscript/core">
in your page, you can just refactor that as:
<script type="module">
import { hooks } from "//cdn/@pyscript/core";
</script>
that's it?import { hooks } from "<path_to_pyscript>/core.js"; come from? I looked at existing docs also did a github search and there are no such example that shows me how I get path_to_pyscript. Also it's unclear to me, is it only part of javascript modules? I would prefer something as close to working code as possible. 16 Nov 2023: Yesterday & Today 17 Nov 2023: Yesterday & Today 17 Nov 2023: Yesterday & Today Sorry.
~ Dec 6-8, 2023
PyData Global Sprints CFP : https://forms.gle/mwfasfnXvkV1ykjo6
Conference website: https://pydata.org/global2023
Here's their sprint guide notes: https://hackmd.io/lkkx0XCcQb2tZd_wO23WYw 
3
3 3 20 Nov 2023: Yesterday & Today 21 Nov 2023: Yesterday & Today 21 Nov 2023: Yesterday & Today 22 Nov 2023: Yesterday & Today 23 Nov 2023: Yesterday & Today 24 Nov 2023: Yesterday & Today 27 Nov 2023: Yesterday & Today Sorry.
~ Dec 6-8, 2023
PyData Global Sprints CFP : https://forms.gle/mwfasfnXvkV1ykjo6
Conference website: https://pydata.org/global2023
Here's their sprint guide notes: https://hackmd.io/lkkx0XCcQb2tZd_wO23WYw 1 or 
28 Nov 2023: Yesterday & Today or 1 29 Nov 2023: Yesterday & Today 4 /latest as an interesting topic https://jeff.glass/post/whats-new-pyscript-2023-11-2/
1 1 1 30 Nov 2023: Yesterday & Today /latest as an interesting topic https://jeff.glass/post/whats-new-pyscript-2023-11-2/ 
01 Dec 2023: Yesterday & Today 01 Dec 2023: Yesterday & Today 
1env, very clean and will be helpful 1 04 Dec 2023: Yesterday & Today 05 Dec 2023: Yesterday & Today 06 Dec 2023: Yesterday & Today https://github.com/pyscript/pyscript/pull/1889
1 07 Dec 2023: Yesterday & Today 
1 1examples folder being removed from the main repo. I'd like to add a note like "To see current examples, go to _______". What should I put there these days?
I'd also add a note encouraging folks to copy and play with the examples now that it's very easy to do so... 4 3 08 Dec 2023: Yesterday & Today 11 Dec 2023: Yesterday & Today 12 Dec 2023: Yesterday & Today 13 Dec 2023: Yesterday & Today 1 1audio side of it?main.py that is really the whole code for the app. Looking forward for feedback on the pydom and pyweb.media APIs.! 1 14 Dec 2023: Yesterday & Today 15 Dec 2023: Yesterday & Today 15 Dec 2023: Yesterday & Today https://github.com/pyscript/docs/pull/56
18 Dec 2023: Yesterday & Today 
1 19 Dec 2023: Yesterday & Today 20 Dec 2023: Yesterday & Today 
21 Dec 2023: Yesterday & Today 21 Dec 2023: Yesterday & Today 22 Dec 2023: Yesterday & Today 25 Dec 2023: Yesterday & Today 25 Dec 2023: Yesterday & Today 26 Dec 2023: Yesterday & Today 27 Dec 2023: Yesterday & Today 28 Dec 2023: Yesterday & Today 28 Dec 2023: Yesterday & Today 29 Dec 2023: Yesterday & Today 1 2 1 01 Jan 2024: Yesterday & Today 01 Jan 2024: Yesterday & Today 02 Jan 2024: Yesterday & Today 03 Jan 2024: Yesterday & Today 03 Jan 2024: Yesterday & Today 04 Jan 2024: Yesterday & Today 05 Jan 2024: Yesterday & Today 08 Jan 2024: Yesterday & Today 
2 09 Jan 2024: Yesterday & Today 
document.currentScript behavior: https://github.com/pyscript/pyscript/discussions/1930
1 10 Jan 2024: Yesterday & Today 11 Jan 2024: Yesterday & Today @Fabio or perhaps @tedpatrick ??? 1 1 12 Jan 2024: Yesterday & Today 15 Jan 2024: Yesterday & Today 1
16 Jan 2024: Yesterday & Today 17 Jan 2024: Yesterday & Today 18 Jan 2024: Yesterday & Today 1 1__terminal__ landed in here as discussed with both @Jeff Glass and @ntoll https://github.com/pyscript/pyscript/pull/1947 ... if it's green and approved and merged, I am confident we could release today or tomorrow.
so multiple terminals were basically for technical docs (on the understanding that each terminal is its own environment). 1 LGTM them we can refine to multiple (independent) terminals https://github.com/pyscript/pyscript/pull/1948 /cc @ntoll @Jeff Glass
1
1 1 2
npm run server then reach http://localhost:8080/test/py-terminals.html and hack away pydom and cleaning all the current mess. Got it working (at least somehow ) with micropython (both pydom and the when decorator) but there's a ton to do still. Initial work here https://github.com/pyscript/pyscript/pull/1954 .
1create_proxy = "auto" in the TOML it would be experimental_create_proxy = "auto" ... is anyone against that? It does nothing if such flag is not enabled, it changes A LOT if it is.
1 1
1 2npx mini-coi ./ would start a server that does what npx static-handler --coi ./ does now, maybe helping further users and developers to have one module that does it all for testing purposes ... thoughts? 
thank you @Andrea Giammarchi! 1sync_main_only flag to workaround missing SharedArrayBuffer cases: https://github.com/pyscript/pyscript/pull/1967
1 1 1
1 1
PyWorker/MPyWorker? Using the worker attribute on a pyscript tag without the right CORS headers still throws up on me about SharedArrayBuffer not being available in an insecure environment.
EDIT: OH! I missed that adding the sync_main_only = True param to the config was necessary. Please disregard. 1PyWorker/MPyWorker? Using the worker attribute on a pyscript tag without the right CORS headers still throws up on me about SharedArrayBuffer not being available in an insecure environment.
EDIT: OH! I missed that adding the sync_main_only = True param to the config was necessary. Please disregard. sync_main_only was a requirement for these use cases:
await when it calls sync operations from the worker ... accordingly, if you expose utilities form the worker and you await these on the main, SAB is not needed, you add that flag to the config, and you can delegate to the worker anything you like. sync_main_only removes the ability from the worker to run DOM APIs (and others from main) synchronously ... I need to double check though if that means nothing can be used, but I think your example and message is clearhttps://cdn.jsdelivr.net/npm/@pyscript/core@0.4.11/+esm I would suggest using the https://pyscript.net/releases/2024.3.2/core.js onefetch example is spot-on but I would add the following:async attribute within the script if top-level await is desiredawait pyscript.fetch("https://reqres.in/api/users/2").text() directly, which fails if the response was not ok (not in the 200-299 range).json() which is not mentioned ... the fetch API returns proper Python dictionaries when .json() is used, instead of JS proxies ....bytearray() which returns a Python result too ... as a matter of fact, all fetch methods in PyScript namespace return Pythonic references instead of JS proxies ... I think that might be interesting/important to mention
sync_main_only removes the ability from the worker to run DOM APIs (and others from main) synchronously ... I need to double check though if that means nothing can be used, but I think your example and message is clearhttps://cdn.jsdelivr.net/npm/@pyscript/core@0.4.11/+esm I would suggest using the https://pyscript.net/releases/2024.3.2/core.js onefetch example is spot-on but I would add the following:async attribute within the script if top-level await is desiredawait pyscript.fetch("https://reqres.in/api/users/2").text() directly, which fails if the response was not ok (not in the 200-299 range).json() which is not mentioned ... the fetch API returns proper Python dictionaries when .json() is used, instead of JS proxies ....bytearray() which returns a Python result too ... as a matter of fact, all fetch methods in PyScript namespace return Pythonic references instead of JS proxies ... I think that might be interesting/important to mention 1worker attribute to work properly, as the input is mandatory/needed in here, you need to enable COOP and headers to allow SharedArrayBuffer. npx mini-coi ./ would do ... save that code as index.html in any folder you like, go into that folder and run that command ... you should be in a REPL like this one:
3 1
1 1elements.py seems to really make MicroPython unhappy (triggers a memory access out of bounds exception). Think this is probably upstream..
elements.py seems to really make MicroPython unhappy (triggers a memory access out of bounds exception). Think this is probably upstream.. ltk to the pyscript github org. 1 1elements.py seems to really make MicroPython unhappy (triggers a memory access out of bounds exception). Think this is probably upstream.. 
1 
1
.zip folder of our stdlib and produce a base64 based artifact that would unpack in the FS the whole content, as opposite of manually creating a huge string with all our current and future stdlib in there as single runPython command. However, this not only would need quite some time to be implemented, it won't solve the underlying issue that MicroPython runPython (and I believe runPythonAsync too) is incapable of handling biggfer inputs so I'd rather go to fix that and think about improvements later on. After all, our files behind the CDN get already compressed, and compressing the compression feels off to me. 1
File.new([blob])) so it's possible to create downloads from the virtual FS too ... plus other fixes landed these days ... please feel free to test it via https://cdn.jsdelivr.net/npm/@pyscript/core@0.4.31/dist/core.js and https://cdn.jsdelivr.net/npm/@pyscript/core@0.4.31/dist/core.css respective JS and CSS urls, thank you! 1 1
Note the docs has search enabled, url fixes in place, plenty more examples in code, and many pages have been completely re-written, simplified and corrected. Call out to @Andrea Giammarchi for such amazing work in this respect.
1 1py-editor is always showing an error in the new release (2024.5.1): https://github.com/pyscript/pyscript/issues/2056
We're going to cut a new release so it's all fixed for PyCon. By the time you wake up and read this, it'll probably already be out. See the announcement about the latest release in #chat etc... 1 We're going to cut a new release so it's all fixed for PyCon. By the time you wake up and read this, it'll probably already be out. pyweb.ui.elements . I'll add a couple of quick PRs to change and bring pyweb.ui.elements to pyweb.elements and avoid too deep nesting. There are also a couple of small improvements in the interface that I'll land tomorrow as well. Rest will be adding docs.
1pyweb.ui.elements . I'll add a couple of quick PRs to change and bring pyweb.ui.elements to pyweb.elements and avoid too deep nesting. There are also a couple of small improvements in the interface that I'll land tomorrow as well. Rest will be adding docs. 3 ) should be reviewers, with at least one approval needed for publishing.
Just wanted to give a heads up and check if there are any objections 6 3 1pyscript.web tests and latest media but I've fixed much more and this is the current state when SharedArrayBuffer is enabled:
https://youtu.be/jz7CX8xgRNA 1 1 1 1