@Dennis?

@Dennis could you please pull stacked?

@wat I'm extremely busy right now and will be for a while. I'll try to take a look tomorrow.

@Dennis OK, sorry

sorry to hear that. good luck with whatever :D

@ConorO'Brien Pulled. (I do have a script for that.)

to pull languages?

small-ish bug: cache hits persist between pulls. e.g. code that errors in a previous version will still "error" after a language is pulled

That's not a bug, more of a won't-fix situation. The cache doesn't even know which language you're running, and once TIO reaches v2, you'll be able to run several languages at once.

well, that sucks for code ran in previous versions, since they'll never function correctly on TIO thereafter, afaict

@ConorO'Brien If you run it again with output cache disabled, it will overwrite the cached result.

@Dennis I think that tio should somehow cater for this. for the end user it's indeed a WTF moment. cache entries should be marked by languages. multiple languages case does not preclude it, you will just need to invalidate all caches for particular languages.

Otherwise you are undermining the usefulness of the cache feature, I never know if the result I'm getting is stale or not (the cached message does not help at all here) so I'll be unchecking the cache by default the first thing I do when I come to tio, just to make sure that I'm not getting a bogus result. Your certainly do not want that.

@wat tio.run/nexus/…

@Dennis yay :)

@AndrewSavinykh That seems like a lot of trouble and unnecessary cache purging for something that won't be an issue 99% of the time. I'll think about it.

thank you

@Dennis Can I please get a pull on RProgN?

@ATaco Done.

@AndrewSavinykh There has to be a better way to do this. The frontend currently runs the code again if the response is invalid (e.g., if the requested language did not exist when the result was cached). How about if it does the same if STDOUT was empty and the code didn't exceed the time limit? That would at least catch all syntax errors.

Thanks!~

@Dennis sounds very palliative, I'm not in favour of this. It's not just about errors, it's also about wrong results (because of the language).

I totally understand that it is not easy to implement, and that it is probably not a common use case, but I see it lead to exactly the perception of the cache I described above. If you think that it's so rare that it won't affect the perception significantly, then I'll have to believe you. But you are betting on it ;)

@Mego Done.

@Mego What changed?

Cool. So, it's even harder to type now?

Does there even exist a keyboard layout with those builtin?

Actually is hard ;-;

It would certainly be possible to create one. A keyboard has about 45 keys, using them unshifted, with Shift, AltGr, AltGr+Shift, Ctrl+Shift, and Ctrl+AltGr, gives you way over 256 characters, while still retaining Ctrl for editing commands. You could also assign some prefix/dead/accent keys... The only keyboard layout I use uses just unshifted, Shift, AltGr, AltGr+Shift, and six dead keys, to enter all APL glyphs and all valid Dyalog APL names (pan-European), plus a few useful characters.

How do you type ALT-codes?

I've never figured it out.

@Pavel Hold left Alt while typing on the numpad (with NumLock on)

Oh, I need the numpad

I've always just tried it with the upper row of numbers >_>

@Pavel If using a Dell laptop (e.g.) you can hold ALt+Fn and use the slanted "Numpad" overlay from 7-9, U-O, J-L, ,-/

@Adám It's cool, I have a numpad.

I just never use it.

Note to readers: This is Windows only.

Crap

But on Linux you can modify your keyboard very easily.

Yeah, I'm still trying to recreate the Windows Russian Mnemonic keyboard.

Linux: Hold down Control and then Shift and press the "u" key. You should get an odd looking character that is waiting for input. Type the numeric code and press space.

(Does not work on all systems)

Good to know.

I'll try it later

Or use vim

I use Atom.

In insert mode, <C-v> then x and then the hex code for character

I wonder if I can hack that functionality into atom.

I haven't tried making my own extensions yet.

@wat @Dennis do we have hello world for wsf?

Ah sorry

did not noticed all that code

in the link

It does not display well ;)

@Dennis timeout: failed to run command ‘/srv/wrappers/wsf’: Permission denied

On that very same link

Can you have a look when you've got a moment?

@AndrewSavinykh Gah, that happened when I switched to arena2... It should work now.

@Dennis thanks, why did it happen?

Because since I couldn't push from arena1 because of the PS issue, I recreated the file manually on arena2, but forgot to chmod.

@Dennis you cannot push from arena1 because of the PS issue?

which file? the wsf wrapper?

Yes, I couldn't pull the changes to the PS wrappers since they require native images now, so I couldn't push anything either.

@Dennis sorry for being slow, I thought you nuked the arena without native images. Is this wrong?

I nuked arena which ran on Fedora 24. arena1 was still using "normal" PowerShell.

out of curiosity, why?

because you did not have time to update it if I have to guess?

or is there a technical reason?

Time, mostly. It's not something I'd want to do while the arena is being used, both because it takes some time and because something can always go sideways.

but you have two of them now, right that you are swapping in and out?

Yes. That's why I asked for the test earlier. arena2 hadn't been tested since the PS upgrade.

sorry for silly questions, I know it's all obvious to you, but sometimes it's dificult for me to picture it from here

Yeah, I get that a lot. ;)

I tend to forget that the other party can only know what I've told them.

so which arena is in now?

arena2

with PS ni`s?

Yes.

so that arena in theory matches current state of git exactly, yes?

Yes.

And in a few minutes, arena1 should as well.

okay, now I think I'm with you

and If I made this test earlier, we would have caught that during that earlier run ;)

what a slacker I am ;)

:P

@AndrewSavinykh Karen should be back in the main office (where I am) today. Did you get your Windows and DSS license sorted?

@Adám I have no idea, I searched high in low on the MyDyalog pages, but I was not able to detect if it's enabled or not. How do I check?

@AndrewSavinykh If you go to Downloads, can you download the Windows version you need?

@Adám yes, both linux and windows. She said that she'd ask Jason to temporary give access to both and then will sort out DSS on the first convinience

the former has happened but I don't know if the latter did

@AndrewSavinykh OK, then it's fine. She probably meant today when she will sit at her desk opposite Jason...

@Adám thank you for your assistance

@AndrewSavinykh No problem at all. Being on DSS, you should never hesitate to contact me adam@ or support@.

@Dennis ^^^ applies to you too, should you want to explore APL.

@Adám you never know how things pan out and it's possible that I won;t have much time for it, but as I said I'm curious. I know I'll be quite busy next few weeks, but I should become less busy after that, when I hope to explore a bit more of APL. I even ordered the printed book, because, well I like books, having it handy increases a chance of me reading a bit here and there and learning more ;)

@AndrewSavinykh Ah great. Yeah, we have been shipping quite a few lately. Hopefully we'll soon get around to updating or amending it with all the newest cool stuff.

@Adám yes, I noticed that it's 8 years old, but I figured not a lot of changes happened during this time, given your yearly release cycle

Wrong channel ;)

@Dennis my backend is broken

@Dennis I sent you the hex dum of response from backend, I cannot see (Again) gzip encoding header in it. This time I made sure not to let fiddler decode it

@Dennis any idea what I'm doing wrong?

@Dennis uhm there is still a chance that fiddler is doing this despite the fact I told it not to

@Dennis okay I cannot see the gzip folder on the working tio either, so I'm assuming this is fiddler output. The problem is something else....

@Dennis okay, I found the issue and it's your fault ;)

github.com/TryItOnline/main-server/blob/master/…

Since my sed expression looks for arena.tryitonline.net there it was unable to change this line properly, can we discuss a fix for that?

@Dennis so I see two options. Either I do this:

sed -i "s/\s*ssh_user_host=.*/ssh_user_host=runner@$ARENA" run

or you make sure that it always says arena.tryitonline.net and not arena1, arena2 or anything else. What do you prefer?

Given that we have both /run and /run-legacy to care about I'd prefer the latter

@AndrewSavinykh I'm struggling with this as well. I'll put the hostname in a configuration file to avoid sed'ing. I'll try to get that done tonight.

@Dennis finally ;)

@Dennis what I meant to say, is thank you, your help is greatly appreciated

> TIO Nexus is the current stable version of Try It Online, a family of online interpreters for over 170 programming languages.

@Dennis ^ from tryitonline.net

Maybe that could be a script that looks at how many languages there are currently?

Also, @Dennis, can you link to the GitHub for TIO anywhere on that front page?

@Pavel once you cleared with Dennis that he is not opposed to the idea in principle you can make a pull request for this feature ;)

I tried, but couldn't find that file >_>

@Pavel which one? I can certainly help you finding it ;) Do you mean this one

I can change it to 200, but I'm not sure how to write a script to update it automatically.

@Pavel yep just changing it to 200 is probably not what you are after. I kind of assumed you are a programmer since tio is all about programming. are you not?

I am, but I don't know JavaScript

I don't know how to query the language list

The language list is here It looks like it is loaded here and this is an example of manipulating it.

I mean you do not have to if you don't want to ;) Just in case you are up to it

I had this brilliant idea of querying Bash for the amount of files in /srv/wrappers/ then decided that that was a stupid idea.

Well, it's true that there are over 170 programming languages.

/srv/wrappers are on a different server - on the arena server. You are displaying the front page from the main server

Hence, stupid idea.

So definitely makes more sense to use languages.js for that

@ATaco exactly my thoughts ;)

I just used a preg_match_all on languages.js, and that got an accurate count.

I have no idea what that is.

I don't JS.

Good thing it's PHP.

It's just a regex match

PHP does not help us a lot since tio is not built on it ;)

I have never done anything webbased.

Well, you can do it with the same regex on JS.

.match(/".*?"\s*?:\s*?{\s*"name"\s*:\s*"(.*?)".*?}/g) will return an array of all language names. (Hopefully)

How do I write that as a script that queries the file?

The bit before the .match is the important one.

shrug

@ATaco you can. for this you need to a) load it as a string b) translate php rergex flavour to javascript one; if you do it in javascript, you do not need to download the file as a string, browser loads it for you. And all javascript you need is this:

screencast.com/t/OoJ9PLfv

So while possible, there are more idiomatic solutions ;)

@Pavel, you can try that snippet yourself in the dev console of your browser

So, can I replace 170 with <script>(langauges.practical.nameIdPairs.length + langauges.recreational.nameIdPairs.length)/10*10</script>? (I want it to round down to the nearest 10)

@Pavel you'll need to test it of course

@Pavel as a minimum you'll need to reference languages.js because they are not currently referenced from the front page. There is also a chance that Dennis wants to keep it this way (so front page is a static site)

How do I reference languages.js?

I linked you an example above

Oh, right.

Well, I tried editing the HTML from f12, it inserted the code literally without evaluation.

you need to insert the value via DOM w3schools.com/js/js_htmldom.asp

I'll try later... maybe...

;)

imgur.com/a/9NCAa

Dynamic TIO Advert gif.

Although not very Dynamic when hosted on Imgur.

I think ran should be run

It depends tense, which isn't apparent from context.

It doesn't sound as nice.

Also, it doesn't appear to cycle through all 202 languages?

Regardless, it's not the idea.

Correct, it cycles through a random selection of them.

15 of them, to be exact.

If it has to be static, why not go for the full 202?

It doesn't have to be static.

I just don't happen to have a webhost to put this on available to me.

Give me somewhere to dump PHP, then I can.

This should totally be the TIO advert tho

@Dennis ^^^^^^^^^^^^^

@Pavel This is where it's inserted in this p element here: screencast.com/t/Uf6Fiy0JDIr If you right click then you will be able to copy selector

Note though, that it is not very reliable method since it depends on particalar order of elements on the page. What you need to do instead is name an element that you would like to query by giving it an id attribute, then you can be able to query it with #yourid

Do you want to write that script? You seem to know what you're talking about.

It'd be nice if the Languages was stored in a JSON and not a .js

@Pavel no I do not want to write it. I have a big back log of other things at the moment, I just thought that it would be nice to share a little knowledge ;)

@ATaco what do you mean by that?

@ATaco this looks like json to me, do you see anything wrong with that?

The language list is a JS file that simply makes a table of languages.

Well, I was looking at this

@ATaco your browser cannot include json files, it can only include js

I'm not using my Browser. :<

@ATaco so my point it is stored as a json

exactly as you said

Well, that's good to know.

it servered as js with this script because browser would not be able to include it otherwise

And what you are looking at is the result of that script ;)

@ATaco This is what you want: tryitonline.net/languages.json

Also, anyone know a free PHP host I can dump this in for testing?