The following code written in Nim is valid but it looks like failing to compile.

I think adding github's raw links straight from regular repositories would make a lot of sense instead of just supporting gists. This would make it a lot easier for people with existing repositories of scripts to just paste the link in and see it working instead of having to copy the text over to pastebin.

The following code hasn't finished being compiled into assembly (via ;asm js), and it's been 45 minutes.

for (let i = 0; i<i+++--i; i+=i--+(++i)) {
  console.log(i);
}

Editing a ;compile or ;asm message reply (e.g. to change arguments) doesn't seem to check for a message being replied to:

When you compile in bash
"kill -19 0" the bot stop and if you delete te message it returns "Critical error: Unkown Messege"

Getting this for a simple hello program:

/usr/bin/ld.gold: error: cannot find -lnuma

rts/posix/OSMem.c:370:0: error:
     error: undefined reference to 'mbind'

rts/posix/OSMem.c:748:0: error:
     error: undefined reference to 'numa_num_configured_nodes'

rts/posix/OSMem.c:758:0: error:

Discord command:

;compile haskell
```haskell
main = print "hello"
```

Compiler is offline, as shown in the picture.

project does not compile with current serenity version v0.11.1
could you provide a more specific version to use with the current state of the master branch?

It happend literally yesterday when I wanted to run the .exe release file but instead I get:

INFO discord_compiler_bot > Registering owner(s): 559039176905850880 Error: reqwest::Error { kind: Decode, source: Error("missing field provider", line: 1, column: 1441) }

Now sorry if I don't have more information, because I literally copied and paste the .env example from the root tree and I ran the standalone exe and I get this. It is not my bots fault as I verified other discord code working on my bots. But this might be Discord's new privacy policy or something but it literally work the day before.

Hello, I'm having an issue that whenever I run code with an input(), "I get EOFError: EOF when reading a line", but a simple print() doesn't. I'm compiling python.

This is something that I've been tossing around in my head for a little bit but I'd like to formalize the migration publicly.

We will be migrating to use more of Compiler Explorer and less of Wandbox. I'll explain how it works now and then how it'll work in the future.

As of now all assembly requests are executed by Compiler Explorer and all execution requests are executed by Wandbox. This will be changing to default all execution requests to Compiler Explorer, if Compiler Explorer does not support the language then we will fallback to Wandbox. With this system, we will not lose support for any languages.

With this system, we will move a majority of our requests to a more reliable platform that I think can be better aligned with the spirit of this project. We have great communication and help with Compiler Explorer, so I'm drawn to make this change.

What's also on the table is downtime fallback. If Compiler Explorer were to drop our request, API error, or otherwise unable to fulfill a request we can rely on Wandbox to fulfil the request. This logic could be confusing, but for the large majority of our requests, no one cares what compiler or interpreter is used.

Honestly, and I have heard a decent number of people share this opinion, but the output from code execution would look a lot better if it just used a standard code block. The one thing the embed has going for it is that it provides more color contrast. If the output were made into an embed, there would also be more room for multi-page output. The RichEmbed compresses the code in a small text area, and it isn't great for code viewing.

I hope this FR can be considered.

Problem

Working

Hey! I have this bot in my server and it's great! But what I wanted to know is how to remove the vote and botinfo command since I don't really need them and also how to customise what emotes the bot reacts with

Looks like we're missing a check here for when the compiler cannot be resolved, it was previously assumed that we can already resolve the compiler at this point, but the code path for assembly requests are different than compilation and has different guarentees.

We just need to return a CommandError here for invalid target

Dec 16 00:13:27 gabriel discord-compiler-bot[30098]: thread 'tokio-runtime-worker' panicked at 'called `Option::unwrap()` on a `None` value', src/managers/compilation.rs:99:65
Dec 16 00:13:27 gabriel discord-compiler-bot[30098]: stack backtrace:
Dec 16 00:13:27 gabriel discord-compiler-bot[30098]:    0: rust_begin_unwind
Dec 16 00:13:27 gabriel discord-compiler-bot[30098]:              at /rustc/f1edd0429582dd29cccacaf50fd134b05593bd9c/library/std/src/panicking.rs:517:5
Dec 16 00:13:27 gabriel discord-compiler-bot[30098]:    1: core::panicking::panic_fmt
Dec 16 00:13:27 gabriel discord-compiler-bot[30098]:              at /rustc/f1edd0429582dd29cccacaf50fd134b05593bd9c/library/core/src/panicking.rs:100:14
Dec 16 00:13:27 gabriel discord-compiler-bot[30098]:    2: core::panicking::panic
Dec 16 00:13:27 gabriel discord-compiler-bot[30098]:              at /rustc/f1edd0429582dd29cccacaf50fd134b05593bd9c/library/core/src/panicking.rs:50:5

Currently custom emojis are required to self-host the bot, as default (unicode based) emojis don't have a 64 bit snowflake-ID. This could be implemented as defaults to fall back to when no emoji is specified in .env

The react roles for the ;languages command work with the default emojis, so hopefully this won't be too hard.

https://vlang.io/

When compiling we're aliasing cpp for c++, but not when searching for compilers.

This should be an easy contribution. We basically just need to run the user's input string for the compilers command through the function below

There's something flawed with how we increment the server count which causes the presence to display an incorrect (sometimes grossly incorrect) value. I haven't investigated much into why this is happening, but something is awry.

On top of this - some shards display server counts that are different than other shards.

Our proper server count is 2,396 servers. This is the display value.

Hello!

I'm a big fan of this bot because of what it can do on discord servers! It can compile almost any language I can think of. Yeah, the almost is bugging you out isn't it? Same for me!

It would be really great if Julia compilers are added too! That's a language that uses math a lot and can sometimes help us do math programmatically. Which calls for Julia's compilers' need.

I hope this issue is considered as soon as possible!

a zwsp is inserted before backticks to make them display correctly, but the length is not checked again, making the message length potentially go over the limit

The following (valid) code fails to compile

This is because the function get_statement_end in src/cppeval/eval.rs keeps track of the number of open brackets to ensure the code is valid, but it fails to ignore brackets that are commented out.

At the current moment, running ;compile haskell will result in a compiler error ghc: failed to create OS thread: Cannot allocate memory

Hi, i found an error while trying to compile a simple javascript hello world program, weird thing is that throws a status code 200. Here's a picture of the issue:

We need a feature that lets users edit requests after their original execution that way users don't have to go back and submit an entirely new request for small mistakes.

Picture telling everything

It'd be great to take advantage of CE's preprocessor tool when -E is used during C/C++ compilation. The preprocessor tool automatically filters out headers, junk line directives which clutter even the most simple of preprocessor results, and can also automatically format the output. Demo: https://godbolt.org/z/bxoj4fzq7.

The API for this is undocumented but a simple request for the preprocessor tool looks like:

{
  source: '...',
  options: [  ],
  backendOptions: {
    producePp: { 'filter-headers': true, 'clang-format': true },
    produceGccDump: {},
    produceCfg: false,
    produceDevice: false
  },
  filters: {
    binary: false,
    execute: false,
    demangle: true,
    intel: true,
    commentOnly: true,
    directives: true,
    labels: true,
    optOutput: false,
    libraryCode: true,
    trim: false
  },
  bypassCache: false,
  tools: [],
  executionParameters: { args: [], stdin: undefined },
  libraries: []
}

Maybe this could be done for both ;compile and ;asm.

Easy fix for someone looking to get contributions.

The following snippet shouldn't execute when max page count is <2

Hi!

For our compilation requests we rely on the service Godbolt or Wandbox to process user requests.

What this means is that our project cannot add new languages or compilers, but any new additions Godbolt or Wandbox make we will inherit. Please search out for these projects and direct the feature requests to them. If the request is approved and completed then the bot will be able to support the new language within 24 hours.

Thanks!

As indicated by the failed tests, the latest commit leaves the bot in an unusable state:

Wandbox request failure 
 FetchError: invalid json response body
at https://wandbox.org/api/compile.json

reason: Unexpected token < in JSON at position 0 
Please try again later

This is the embed currently generated by default:

This can take over a half of available screen height (and on mobile devices, possibly not fit on the screen at all), which can break the flow of discussion. Meanwhile a lot of information provided are either not useful at the moment or (subjectively) not needed in general:

• return status code (it's 0 in a vast majority of uses)
• headers
• reactions like :checkmark2:
• possibly even the "Requested by" part (especially if the code block is directly above the embed)

Meanwhile, the build_small_compilation_embed function added in #98 shows it in a nice compact fashion:

It would be nice to be able to somehow enable this compact embed for all compile output messages, either for a single message message, or in the scope of entire channel/server.

It only answer the ;ping

Kotlin is used by many developers and continues to grow every year. It may be good to support this language as well.

Here's an example of the input that breaks the parser

;asm python38
wynik = 0
i = 0


while i < 3 :
    x = int(input("podaj liczbe parzystą: "))
    if ( x % 2 == 0):
        wynik += x
        i += 1
    else:
        print("podaj liczbe parzystą.")
    continue

print ("wynik dodawania to: ", wynik)

What's happening is that the parser is constantly looking for the beginning of a code block, which it never reaches due to the malformed input. The parser in turn is also looking for < inputs to indicate a url code request. When it hits the while loop, everything after the < is loaded in as a url. The bot then fails with "Invalid hostname or protocol" linked below
http://i.michaelwflaherty.com/u/ymoOvq3Rwg.png

Please consider replacing the following respectively because people on phones cannot use " and ' also iirc those quotes used in phones aren't used in any of the languages for a particular use although you could add an ability that escaped ones aren't replaced

for example: “ will be replaced but \“ won't be.

“ -> "
” -> "
‘ -> '
’ -> '

It's possible to trip up the command parser if requests are submitted in the following form

;compile c++```cpp
int main() {
  return *(reinterpret_cast<int*>(0));
}```

I don't have time to get into it right now, but it smells like it's coming from here

I saw this request go through and the bot truncated the compiler output. The warnings are the only output visible from discord and the errors were truncated, we should definitely be able to fit the full output in the embed.

Hello!

This bot works great! But it would be much better if the bot supported external libraries (say numpy for Python, etc.) either as arguments:

;compile python -libs=['numpy', 'matplotlib', 'cv2', 'dyepy']
```py
import numpy as np
import matplotlib.pyplot as plt
import cv2
import dyepy

# Code
```

or any other method that lets us use a particular module(s) (any module that is publicly available to others).

I hope this issue is considered as soon as possible.

The compiler doesn't takes input in input based programs

Please, create mbal supprot:
https://github.com/ringwormGO-organization/mbal

ref: http://www.eelis.net/geordi/#setup

In the URL above, I believe supporting the 2.1 Request syntax section is quite simple and can be very useful for folks just wanting to execute little bits instead of having to type a full well-formed c++ program.

Hello,

Suppose we send a single file (a file, not a code block) main.c
and we reply to our message with

;compile c
argv1 argv2

The bot considers argv1 and argv2 as files and not as command line arguments.
(the same behavior occurs when we reference a code in pastebin with <)

That would not occur if we typed in, in a single message

;compile c
argv1 argv2
```c
#include <stdio.h>

int main(int argc, char *argv[]) {
   ...
}
```

Am I missing something?

CE supports URL #includes but for some reason they do not work with compiler bot requests:

;(

It seems that the parsing logic fails with the following case:

;compile c++  -Wall
```cpp
int main() {

}
```

with two spaces preceeding -Wall the output from the bot is always a compiler error in the form of

/usr/bin/ld: cannot find : No such file or directory
collect2: error: ld returned 1 exit status

In the following command, the rs does not get trimmed properly (assume it's intended properly, I can't figure out how GitHub markdown works):

;compile rust
```rs
fn absoluteValue(num: i64) -> i64 {
return if num > 0 {
num
} else {
-num
}
}

fn main() {
println!("{}", abs(-2));
println!("{}", abs(2));
}
```

Leading to the following bizarre error message:

error: expected one of `!` or `::`, found `fn`
 --> prog.rs:2:1
  |
1 | rs
  |   - expected one of `!` or `::` here
2 | fn absoluteValue(num: i64) -> i64 {
  | ^^ unexpected token

error: aborting due to previous error

Meaning it thinks line 1 is rs and line 2 is the function signature.

The correct error message is this:

error[E0425]: cannot find function `abs` in this scope
  --> prog.rs:11:19
   |
11 |    println!("{}", abs(-2));
   |                   ^^^ not found in this scope

error[E0425]: cannot find function `abs` in this scope
  --> prog.rs:12:19
   |
12 |    println!("{}", abs(2));
   |                   ^^^ not found in this scope

error: aborting due to 2 previous errors

For more information about this error, try `rustc --explain E0425`.

This issue is not present if rust is used or if the error made is fixed (renaming absoluteValue to abs).

  ;compile
    ```r
    print(rnorm(10))
    ```

results in