My favorite, since I'm not a programmer anymore, is excel
E: Your formula has a circular reference. I ain't doing shit till you fix it
Me: where?
E: In your spreadsheet, I don't fucking know
Excel: taking ages to load a file
Excel: There is a link to another Excel document, but I can't access it to update the value.
Me: Where?
Excel: To this document.
Me: ... Where can I find the cell that contains this link?
Excel: I don't know noises
Me: What if it is a named variable?
Excel: Yes.
And don't even try to do a conversion of text to numbers in a big column. There's a super fast way (name is eluding me) but if you respond to the error popup I imagine it looks at each cell of text, thinks says, "abracadabra you're now a number!" for every row. It takes that long
It’s ok, you run the expression debugger, which says the first step, which is all of the formula, will result in an error. So helpful.
Tbf, you have to be pretty far with Rust to get to a point where Rust's compiler errors stop helping you (at least, as far as I've seen). After that, it's pretty much the same
Yep use a little bit more deeply cascaded generic rust code with a lot of fancy trait-bounds and error messages will explode and be similar as C++ (though to be fair they are still likely way more helpful than C++ template based error messages). Really hope that the compiler/error devs will improve in this area
Rust has better runtime errors, too. If you run a dev build, it should pretty much never segfault unless you use unsafe and will instead tell you what went wrong and where, no valgrind necessary.
Can't have a runtime error if you don't have a compiled binary *taps forehead*
(For the record, I say this as someone who enjoys Rust)
This is actually unironically a major benefit of Rust - compile time errors are supposed to be for dev mistakes and runtime errors supposed to be for user mistakes. Way easier to debug something at compile time instead of runtime.
Then there's Haskell that would remove (well, used to at some point) your source code file if you made any errors: https://gitlab.haskell.org/ghc/ghc/-/issues/163
Reading their page gave me a good laugh. Didn't know about this before, and I'm glad to have learned about its existence
Vigil deleted a function. Won't that cause the functions that call it to fail?
It would seem that those functions appear to be corrupted as well. Run Vigil again and it will take care of that for you. Several invocations may be required to fully excise all bugs from your code.
Yeah. this bit got me
MySQL: you have an error near here.
Me: What's the error?
MySQL: It's near here.
Me: You're not going to tell me what the error is? Okay, near where? Here?
MySQL: warmer... warmer...
Oracle: You have this error in line 1
User: Hey, no, there isn't anything to cause this error in line 1
Oracle: I'm telling you, it's in line 1
User: Hum... How many lines are in my 10 lines query?
Oracle: 1
MySQL: you have an error around here
Me: that's the entire query. If you aren't going to tell me what the error is, can you at least narrow it down?
MySQL: ... Stfu
This joke is never funny; Forgetting a semicolon in c results in compile time errors, not runtime errors
Haskell errors:
Iä! Iä! Cthulhu (b -> (a -> c)) -> (b -> (c -> c)) -> a fhtagn! Ph'nglui mglw'nfah [[a]] Cthulhu R'lyeh wgah'nagl fhtagn!
[45 lines of scopes]
Once you understand the type system really well and know which 90% of the error information to discard it's not so bad, I guess.
I literally had a type-theory themed stress dream a couple nights ago. I'll leave it up to you if that makes this less or more funny.
GHC messages are complete and precise, usually telling you everything you need to know to understand, find, and fix the error, that may not even be on the place it's actually detected.
It's also in an alien language. That's correct.
complete and precise
Exactly. It's a perfectly condensed yet totally complete readout of all the data you might need for debugging. It makes mathematicians everywhere proud.
If you don't actually need a complete set of information about possible exotic type choices just to see you put an infix in the wrong place that's basically not the compiler's problem.
(TBF, I wouldn't want to try and mindread the programmer in my compiler either, but then I am a maths person)
I dunno. That set of information about exotic type choices helps me very often. And I can always ignore it when it's not useful.
The bunch of "yes, compiled that module, everything is all right" messages in between them and warnings not surviving a second compilation bother me much more than the error messages. But learning to read the messages was not easy.
And I can always ignore it when it’s not useful.
I did mention that right off the bat. I made it sound unreasonable for comedic purposes, but breaking the jerk I actually do really like Haskell, and Haskell error messages.
Rust is nice, unless you have a traits compilation error from a 3rd party library using types that are more difficult to write than C++ templates.
yeah as nice as it is what you can achieve with trait-bounds there are definitely trade-offs, being compile time and error messages, and sometimes mental complexity, understanding what the trait-bounds exactly mean... I really hope, that this area gets improvement on at least the error-messages and compile time (incremental cached type-checking via something like salsa)
I much prefer getting told of that it doesn't match a trait than get 600 characters of which the majority is implementation detail of global allocators und from what exactly the string is derived.
Depends on what trait bound error messages you have had yet, I had 1000 lines long already, where it's not obvious at all what is meant (and is often a very simple fix). But I'm sure this will get better over time, there's already a bigger ongoing redesign of the type system solver, so maybe it will be integrated into stable rust soon.
If it were up to me I'd be avoiding templates altogether. Sadly, I don't get that choice with 3rd party code.
I would say that joke only works with template bugs, normally gcc gives fairly good directions.
Template<Instatiation::_1,_2,_3, Instatiation2::_1, _2<closure::wrapped<_1[map::closure_inner]>>, Outer<Inner<Wrapper<float>>>>::static_wrapper<std::map<auto>, spirit::parser::lever<int, std::array<int>>::fuck_you
Syntax error: unmatched thing in thing from std::nonstd::__map<_Cyrillic, _$$$dollars>const basic_string< epic_mystery,mongoose_traits<char>, __default_alloc_<casual_Fridays = maybe>>
(from James Mickens' The Night Watch, highly recommended with his other essays: https://mickens.seas.harvard.edu/wisdom-james-mickens)
I like how this depicts how rust is designed more top down and C++ is designed bottom up.
How compiler builders see peppa:
https://www.deviantart.com/ian-exe/art/Peppa-pig-front-face-743773121
I think these two pigs are the best comparison of rust and c++ I've ever seen. Also considering the aesthetics, it's so accurate.
LISP be like: "There is an error here in this wierd code I just generated and which you never saw before. Wanna hotfix it and try again?"
Ever tried using typenum numerals in Rust? 😅
Try it and see the errors with something like typenum::U500.
"Fuck you ... or not. One day ... or two ... or every day. For certain, when you least expect it"
(C++ errors involving memory pointers)
One of the reasons i started learning rust was bc of how easy it is to get into it, or at least that's how it felt for me. It wasn't until a few months into consistently writing that I started to encounter things that I didn't understand.
The good thing about Rust is that if you have no idea of any problem in your code, it very likely because your code is ok.
On C++ things are different.