« first day (1609 days earlier)      last day (88 days later) » 

04:07
@lyxal seems expensive though
$4/yr is a pretty good deal
I would encourage you to consider it :p
Well the problem with getting a domain isn't actually the cost. It's the long term commitment of it
Regardless of whether it's vyxal.net or vyx.al, it still has to be maintained
Lest there be broken links on potentially thousands of answers
That's the benefit of hosting on github pages
At the very least, it's on a domain that will likely be around for a long time even if none of us ever touch it
And I'm aware that the v2 website suffers the same problem
Because it needs to be renewed every 3 months
But that's slightly better than relying on paid services others might not have access to if we ever move on
@Ginger evidently, I already have
A while back too
And I came to the conclusion that it's not worth it
Even though I was fine with going for the more expensive but cooler option
redwolf did offer to buy it for us once though, meaning that we could probably get it subsidised by the American tax payer if he ever becomes president, thereby ensuring longevity with no real personal cost to anyone involved
 
8 hours later…
12:15
@lyxal I guess
I feel like you've used this argument a lot :p
If you've ever seen a pyth answer you'll know what I mean :p
57
A: Don't google "google"

DennisPyth, 14 13 characters L/lbnb"google Defines a named function y. This divides the length by 1 if the string is not google and by 0 otherwise. The idea is not novel, but I came up with it independently. Try it online. How it works L Define y(b): lb Compute l...

for example
there seems to used to be an online interpreter at pyth.herokuapp.com
but that's long gone due to heroku deleting a lot of free accounts
and sure that's a different reason for the website to not exist anymore
but you can see the basic idea that it just takes one thing not being maintained anymore to cause a lot of answers to have broken links
If there's ever going to be a custom url, it's going to have to be under circumstances where I know it's 99.9% guaranteed to be something kept around
12:40
@Ginger where's the grammar files for the webiste stored?
src/common/scripts/languages/
iirc
@lyxal such as Rydwolf buying it? :p
@Ginger good
making a tiny change
where an unclosed string isn't highlighted as a string
even though it should
right
(that's on me for forgetting EOF shouldn't stop highlighting a string)
so here's a thing I considered: archiving only Vyxal and a shim that wraps the API to make it uniform, so we can always have the latest theseus for every version
because that highlighting is now broken for 3.4.5... which I guess nobody's using but whatever
it's about the principle!
12:45
That leaves less wriggle room for how the JS API works
I can always go back and patch all the shims
or we can keep the shim in theseus to make it easier
it could also be slightly problematic when vyncode for vyxal 3 is a thing
because it'll need to have live bit count
and that might need to be embedded into the website rather than the js
ough
I guess
holy hell
also, there could be circumstances where I need to do funky things with urls which breaks in previous versions
for example, it could be that, for some reason, a mistake in the codepage goes unnoticed for 2 years
that is supported by the interpreter already
12:49
and a URL parameter needs to be added to indicate that the new codepage should be used
we can just bump the permalink format version and plug in a converter
@Ginger I mean to say, something might come up 2 years later where the latest website needs to be updated, but can't be used with previous versions
@Ginger oh shoot it actually did
...I can't believe I've been working on this since December of last year
so are we officially rolling it out with 3.4.6? since that's the first version with it in the archive
I think so
3.5 is probably months away tbh
hooray!
should probably mention that in the release notes :p
12:53
Copy button does'nt seem to work
thamsn't one
ah right
...wait wh
> Uncaught (in promise) TypeError: 'clipboard-write' (value of 'name' member of PermissionDescriptor) is not a valid value for enumeration PermissionName.
I thought I fixed that
dammit
what browser are you using
firefox
@Ginger is it great to see us guys again because you're running short on test subjects?
you're BOXES! with LEGS!
> The clipboard-read and clipboard-write permissions are not supported (and not planned to be supported) by Firefox or Safari.
thank you fireforks, very cool
fixing now
12:58
is there an output character limit?
?
oh right that reminds me
Output seems to stop printing things after a certain point
the worker sendMessage issue has reappeared I think
a program that prints in an infinite loop never shows any output (for me on chromium)
I have a solution to this, but you're really not gonna like it
What is it?
redo the JS API so execute returns some kind of object with a function that can be called to step the interpreter
idk if that's even possible
but the worker needs some way to pause when output appears so the browser can send it back to the page
13:06
Just bundle the output into one single returned message
Don't bother with live output lol
:(
that's pretty much what I'm doing right now
@Ginger maybe something like the debugger could be used
Fairly certain there's some ability to step through there
I'm trying to nail down the exact problem
I think it's that postMessage enqueues a message-sending task in the worker's event loop, but the loop is stalled in Vyxal.execute
so the task never runs
we need some way to yield control to the event loop so it can run the tasks
(also you should probably mention theseus in the release notes for 3.4.6 :p)
Maybe
But that'll be for tomorrow
I have an 8am exam tomorrow so I need to head off for the night
one thing
13:11
What is it?
unfortunately, doing what I described will probably not be possible without redesigning the entire interpreter
Keep it simple for now
because execution is implemented as traversing the AST, and that doesn't happen in discrete steps
yeah I'm gonna leave it as it is
In that case, o/ (you'll see me around earlier than usual)
\o
all issues on the site repo have been closed, so now yall can report bugs through that

« first day (1609 days earlier)      last day (88 days later) »