Hacker News Reader

Comment by jjnoakes | Hacker News Reader

> Of course, that's what ghc does by default.

ghc by default can generate object files for systems it doesn't know about? That sounds positively magical. How does it work?

> The same way every other compiler does.

Every other compiler that targets native code for these systems has first-hand knowledge of the file formats. If Haskell doesn't require that knowledge, I would love to hear about how it works.

Here's an example. Say Haskell only supported Windows on x86_64 and I'd like to build my Haskell program for Linux on PowerPC. Are you telling me Haskell on Windows could generate Linux on PowerPC executables and shared libraries without first-hand knowledge of the instruction set, calling conventions, and the ELF file format?

View on HN

I don't know why you are being so hostile and evasive at the same time. What systems are you talking about. I will tell you if there is a compiler that supports them. That's the problem right? That it will turn out that your obscure systems are very normal and supported. Silly fallacious replies about magic are not constructive. Just answer the question and I can help you.

I'm not being hostile at all. If such a thing exists as I believe you described it, I would find it to be magical and would love to use it. I would even pay for it.

And I'm not being particularly evasive either. I gave you an example which fits my situation perfectly. I'd prefer not to reveal the names of the actual systems I need to build for, since it would fairly uniquely identify me.

Are you willing to work with the hypothetical I put forward? Or are you dodging that in favor of attacks on me personally?

Every native compiler does this. There is nothing magical about it. Obviously the platform has to be supported. That's why i asked what the platform is. The idea that you are the only person running a system is absurd. Even the most obscure platforms have thousands of active users. Your hypothetical can not be worked with, because it is a concrete question. Is your platform supported or not. Very simple. And again, if you want to have it generate C, then do so.

You think I haven't tried using Haskell on my platform? It isn't supported and it doesn't work. So you misrepresented things.

I mean why would my first post be "Boy I wish I could use Haskell on my platform" if... I could use Haskell on my platform?

Now I could put a ton of effort into it... for each platform... but I don't have that kind of time to maintain Haskell for N platforms.

It would be much easier to compile via portable C or C++ that was human readable. But I would expect, since Haskell has lazy evaluation and a GC, that there is almost no way to map to readable C or C++. Thunks flying around, continuations, ...

>You think I haven't tried using Haskell on my platform?

Correct.

>It isn't supported and it doesn't work

What is "it"?

>I mean why would my first post be "Boy I wish I could use Haskell on my platform" if... I could use Haskell on my platform?

Because you incorrectly believe you can't.

>It would be much easier to compile via portable C or C++ that was human readable

Then do that, like I told you repeatedly. The fact that you have one of the developers of a haskell compiler trying to help you and you are entirely hostile and evasive makes it hard to take you seriously.

> Correct.

Why would you think that when I've said exactly the opposite? Who is being hostile and evasive here?

> What is "it"?

Also already mentioned above (file formats, instruction sets, calling conventions).

> Because you incorrectly believe you can't.

I wish that was true.

> Then do that, like I told you repeatedly.

If you show me how to generate portable C or C++ that's human readable from Haskell, I'd be happy to.

You keep claiming Haskell solves my problems, except every time we look closely at your recommendations, they are full of holes.

And I'm hostile and evasive.

>Why would you think that when I've said exactly the opposite?

Because you refuse to provide even the most basic information. Like what exactly you tried to use even.

>Also already mentioned above (file formats, instruction sets, calling conventions).

Mentioning is not helpful, telling is. So tell. What format? What instruction set? What compiler?

>You keep claiming Haskell solves my problems

I claimed nothing. I asked you a simple question to try to help you. And look at your responses. You still haven't answered a single basic question.

>And I'm hostile and evasive.

Yes. You refuse to answer even the most basic question that is needed to help you. You do not want help, you want to make a vague, baseless complaint and then avoid having that complaint proven baseless by refusing to provide any information.

I've provided enough information to get to the right answer (Haskell isn't magical and can't compile to my platforms if it never heard of them, and it doesn't generate portable and human readable C or C++ so I can't build Haskell code using my tool chain).

That's all that matters to me. I'm not sure what crusade you are on... but if you had read the comments you replied to you would have avoided wasting my time and yours offering up Haskell as a solution when, based on information contained in the very posts you replied to, you should have been able to figure out that such a position was a misrepresentation.

Whether it was intentional or not I will refrain from speculating.

Your characterization of the discussion is not accurate. Read the thread, it is still here. You lamented not being able to use haskell. I took this in good faith and tried to help you to use haskell. You have done nothing but deflect and derail trying to avoid being helped. You insist haskell can't do this, but you still can't even tell me what haskell compiler you even tried. There's dozens of supported platforms, and more than one compiler that produces C. You are being dishonest because you never actually wanted to use haskell, you just want to post vague anti-haskell nonsense.

>Haskell isn't magical and can't compile to my platforms if it never heard of them, and it doesn't generate portable and human readable C

Haskell isn't a compiler. Different compilers support different platforms. You insist "haskell" has never heard of your platform, but I guarentee you can't find a platform that I can't compile haskell code on successfully. You are simply being dishonest. And yet again, yes, multiple haskell compilers produce human readable C code as output.

I have only looked into GHC.

> I [...] tried to help you to use haskell.

No, you made claims like Haskell natively supported compiling to the targets I am using without even knowing what those targets are. That's not helpful to anyone.

> You have done nothing but deflect and derail

Also not true. I provided plenty of information and offered up an analogous situation. You decided to ignore that, and ignore the other things I said, and you continue to attack me instead.

> multiple haskell compilers produce human readable C code as output

If this was true all you had to do (8 posts up) was mention this. Why all the cloak and dagger? Why all of the ad hominem attacks and misrepresentation? Why not just offer up a simple solution that you know about?

I'd appreciate a list of which Haskell compilers produce portable human readable C code. I'd love to use them, if they exist.

Next time someone says "I'd love if Haskell could compile to human readable C or C++" and you know of an implementation of such, it'd be helpful to mention that, instead of tilting off on whatever personal crusade you imagined up against me.

>No, you made claims like Haskell natively supported compiling to the targets I am using

I did not. I asked what they were. You still have not answered.

>That's not helpful to anyone

Neither is lying.

>I provided plenty of information

You have provided one piece of information, after many many posts: ghc. What OS and platform do you think it doesn't support?

> I did not

You did. Do you need a link to it?

https://news.ycombinator.com/item?id=11707328

> You still have not answered

That's right, I refused to, over and over. The answers are irrelevant.

> You have provided one piece of information

That's right, and zero pieces of information were required for you to be helpful. You claimed you were trying to be, but your actions speak much louder than your words.

> Neither is lying.

You claim repeatedly that Haskell compiles to readable C. It doesn't, which is why you can't list any single implementation which does.

Thanks for trolling.

Good bye.

>You did. Do you need a link to it?

That does not say anything of the sort. It says there are haskell compilers that compile to native binaries and interoperate in a native code environment. You simply refuse to specify your platform so you can try to maintain the illusion that someone would still think you aren't full of it. You aren't even fooling yourself.

>That's right, I refused to, over and over. The answers are irrelevant.

Not if you actually wanted help they aren't. But since you actually just want to shitpost, all your refusal does is make it clear you are dishonest.

> That does not say anything of the sort.

Do I have to read it for you too?

Me: "Are you suggesting Haskell can generate object files, shared objects, and binaries for my systems without compiling to C or C++ first?"

You: "Of course, that's what ghc does by default."

And now you claim you were not saying "anything of the sort" when I point out that you claimed "Haskell natively supported compiling to the targets I am using without even knowing what those targets are"?

> Not if you actually wanted help they aren't.

100% they are. Unless you can explain how me naming the systems I support has any relationship to whether or not some Haskell implementation has the ability to generate human-readable C code?

> But since you actually just want to shitpost

The only shitpost here is you backtracking your claims repeatedly.

Unless you listed those Haskell implementations you claimed existed somewhere and I missed it?

I'd appreciate a link if so.

If not: Good bye again.

>Me: "Are you suggesting Haskell can generate object files, shared objects, and binaries for my systems without compiling to C or C++ first?"

You: "Of course, that's what ghc does by default."

And it does that. And that has nothing to do with the random nonsense about magic and knowing nothing about formats you made up afterwards. Your dishonesty is remarkable.

>Unless you can explain how me naming the systems I support has any relationship to whether or not some Haskell implementation has the ability to generate human-readable C code?

It has a pretty serious relationship to whether or not some haskell implementation has the ability to generation objects and libraries and binaries for it. You know, the thing you are still harping on dishonestly?

> It has a pretty serious relationship to whether or not some haskell implementation has the ability to generation objects and libraries and binaries for it.

That's right. And as i just said, zero relationship to the question you keep failing to answer. Where are these multiple Haskell-to-readable-C implementations you tried to make up?

> And it does that.

So now you reversed your position again and you are claiming that Haskell can generate binaries for my system without knowing what my system is?

Keep it classy. I'll let you talk yourself in circles. You don't need me for that.

Start here:

https://news.ycombinator.com/item?id=11708601

(Let me know if I have to read the thread back to you again).

So which is it going to be next? You did or you didn't claim that Haskell supports my platforms?

Inquiring minds want to know!

Reply on news.ycombinator.com