Why not migrate Portage to C language?

[alexander-n8hgeg5e]

That aside, I should probably just stop using portage for installing kernel sources. There's not really a point if my scripts end up doing the building anyway.

I made a kernel install ebuild today,
I had it as shell script befor, now i tried using an ebuild.

Probably not state of the art in ebuild programming,
but it works really well for me.
I did it because i wanted to do it automatic without compiling
with root privileges.
If someone want's to look at, it's here,

https://github.com/alexander-n8hgeg5e/m ... nux-gentoo

[CaptainBlood]

IMHO some features may be lacking sucn as local MAKEOPTS configuration, maybe in /etc/portage/packages.env...
Thks 4 ur attention.

[alexander-n8hgeg5e]

That very interesting and makes sense.
I once looked at the code that nuitka generates,
and it looked very much like full feature blown python.
So one thing is, because of this features the programs
made by python will be good.
I love python.
It needs some few cpu cycles but they are a good investment.
For me it would not make sense to give up the
elegance of this language.
If i emerge something it's no problem
to use 50+ cpus fully loaded.
If portage gets running and the system is
fast enough the packages are flying.
Dependency calculation is the only thing
that feels slow if there are a lot of packages.
I don't really know how i'ts made
but i can imagine that this is the part
where having python comes in handy and makes
the complicated logic easier to implement.

In things of nuitka it's a really cool thing that
it can with almost no effort speed up python code
maybe x1.5.

I have to add:
To make things parallel is allways a good option,
you can find so many things that could
be offloaded to annother cpu in almost any program.
I even coded some "pyworkers" for my editor neovim.
They wait for tasks like communicate with tmux, check the daytime,...
Then i discoverd the python multiprocessing module
and it is really cool.

[Ant P.]

Portage won't get significantly faster until libbash is finished.

[Tony0945]

Portage won't get significantly faster until libbash is finished.

That looks like never:

This page was last edited on 9 June 2018, at 13:03.

[xahodo]

I'd like the dependency on python removed. Perhaps finally at that point python 2.7 gets removed from my system with the next update after that happens.

[Naib]

I'd like the dependency on python removed. Perhaps finally at that point python 2.7 gets removed from my system with the next update after that happens.

not gonna happen in the short or medium term

Portage is written in python and bash. Feel free to rewrite in C but the onus is on you to show compliance to the PMS and then convince people to adopt it

There is a reason COBOL is still used... Inertia

[alexander-n8hgeg5e]

I'd like the dependency on python removed.

Do you want to remove python2 and 3 ?
To run only python3 it's possible i guess.

[crocket]

If you want correctness and reliability and maintainability, haskell is a good choice.

[Hu]

The problem with Haskell is that it's a language built on lazy evaluation and nobody's actually called for it.

[Ant P.]

So maintainable its proponents aren't willing to write or learn it!

[crocket]

I learned haskell and wrote it. And, https://www.xkcd.com/1312/ is not about lazy evaluation. Lazy evaluation can actually be refreshing while it could lead to leak of space and time if you don't know what you are doing.

If you know how to control lazy evaluation, you can process infinite streams(e.g., radio stream) efficiently. You can also disable lazy evaluation in haskell.

To use any language efficiently requires some learning. Haskell is not a simple language, but it makes it easy to write correct programs. If a haskell program compiles, it probably works unless there are logic bugs.

[Hu]

I learned haskell and wrote it.

In this context, by it, do you mean that you wrote some Haskell program as proof that you knew the language, or do you mean you wrote a Haskell port of some substantial portion of the package manager?

[crocket]

I have written and maintained little haskell programs for my own usage.

The fact that I haven't written a package manager in haskell doesn't automatically mean haskell is unsuitable for package managers.

What I can say is that I could feel the benefits of haskell even in small programs. Ease of writing correct programs could be felt even at small scales.
Writing the same program in C or python felt like a drag because even python allows more classes of bugs than haskell does.
If you lose haskell's corectness, you want to at least gain nimbleness. Some lisp languages like janet attain nimbleness and some degrees of correctness. I personally use janet as a python killer.

[Naib]

Feel free to implement it. it is your responsibility to show compliance to PMS and then convince people to switch

[crocket]

When I read RMS, I remember Richard Matthew Stallman, the founder of Free Software Foundation. Show compliance to richard stallman and GPL? Hail RMS.
By the way, I think RMS is out of shape. As a beginning body builder, I can't take him seriously. His body betrays lack of discipline.
His lack of discipline led him to microsoft.

[Naib]

I meant PMS. TLA aside... this isn't some edict, this is the specification on the package manager, just like there is a specification on the ebuild format. How a package manager behaves is bound by its spec so ebuild writers can be sure their ebuilds will be acted upon concisely.

you must show 100% traceability to the PMS because if you cannot then there is a possibility that an ebuild feature will be not acted upon as the ebuild developer is expected and other package managers do.

Once you have shown 100% traceability to the PMS then it is a matter of convincing people why your implementation is better than portage or pkgcore or paludis

[crocket]

Is there any package manager that doesn't require perl-cleaner and haskell-updater?

[Tony0945]

If a haskell program compiles, it probably works unless there are logic bugs.

If a C program compiles, it probably works, unless there are logic bugs. Same for Fortran. Is there a language that that is NOT true for?

[crocket]

With C, you have to worry about violating memory boundaries constantly even if a C program compiles.
A C program crashes if you are not careful about memory boundaries. Just writing a simple for loop in C can lead to buffer overflow.
People who work on computer security know how hard it is to write correct C programs that don't crash or lead to buffer overflow.
C requires a lot of discipline which is expensive. With haskell, you don't have to worry about it.
If you don't have the budget of NASA, maintaining a haskell program in a multi-person project is substantially cheaper and faster than maintaining a C program.

With python, you need to constantly worry about passing wrong types. Haskell prevents you from passing wrong types to functions.

Haskell is a good candidate for a package manager.

[Naib]

Is there any package manager that doesn't require perl-cleaner and haskell-updater?

Feel free to produce one, one that deals with the incompatibility of the target dev-lang. As I said the PMS is the canonical reference as to what the package manager must do when presented with an ebuild

[Naib]

With python, you need to constantly worry about passing wrong types. Haskell prevents you from passing wrong types to functions.

incorrect

Code: Select all

def greet(name: str) -> str:
    return "Hello, " + name

the positional argument can only by of type:str and the return type can only be of type str. Type hinting has been in python since py35

[crocket]

I'd rather use clojure than python for a dynamic language with optinoal typing.

By the way, regarding your political manifesto.

The best argument against democracy is a five-minute conversation with the average voter
Great Britain is a republic, with a hereditary president, while the United States is a monarchy with an elective king

It's actually an argument against government. By relying on government, you sell your soul to the devil. The government is invested in keeping you dumb. It even poisons the water supply with neurotoxin called flouride. If a child ingested flouride regularly, the child's brain would not develop properly. The average IQ of people on earth has been declining for the last few decades. Flouride is best used in mouth wash, not drinking water. The best form of government is no government. The best form of no government is total self governance which takes a lot of discipline and training. No government, no lover, and no parent is going to do the work of obtaining individual sovereignty for you. You must do the work yourself. I intend to establish and maintain individual sovereignty.
Think about becoming jason bourne who obtained individual sovereignty through years of training and efforts.

[Tony0945]

With C, you have to worry about violating memory boundaries constantly even if a C program compiles.
A C program crashes if you are not careful about memory boundaries. Just writing a simple for loop in C can lead to buffer overflow.
People who work on computer security know how hard it is to write correct C programs that don't crash or lead to buffer overflow.
C requires a lot of discipline which is expensive. With haskell, you don't have to worry about it.
If you don't have the budget of NASA, maintaining a haskell program in a multi-person project is substantially cheaper and faster than maintaining a C program.

I've been writing C for about forty years, professionally for 25. I NEVER had those problems. Maybe code cowboys did, professionals didn't. You are spreading FUD.

[Dr.Willy]

I've been writing C for about forty years, professionally for 25. I NEVER had those problems. Maybe code cowboys did, professionals didn't. You are spreading FUD.

oh dear…