Void Linux Drops Systemd & Switches To LibreSSL

[cryptoahash]

My point is that it is a ripoff from arch linux. Just because he has written the package manager based at looking at the code of pacman. If you look at the source code of both pacman and xbps you will see its very similar.
The templates and PKGBUILD files are almost the same only that it has different name conventions. Everyone can edit the package manager and rename some parts of code so that it looks different.

Crosscompiling is supported by makepkg anyhow, you just change the flags of the makepkg.conf and start cross compiling.


edit: I am not saying he is lying about writing the package manager. I am just saying it is a re-spin of the code.

[Michael_S]

My point is that it is a ripoff from arch linux. Just because he has written the package manager based at looking at the code of pacman. If you look at the source code of both pacman and xbps you will see its very similar.
The templates and PKGBUILD files are almost the same only that it has different name conventions. Everyone can edit the package manager and rename some parts of code so that it looks different.

Crosscompiling is supported by makepkg anyhow, you just change the flags of the makepkg.conf and start cross compiling.

edit: I am not saying he is lying about writing the package manager. I am just saying it is a re-spin of the code.

I thought that too, but the following indicates to me it's not a re-spin:
1. pacman has similar features to xbps-install, xbps-query, xbps-rindex, xbps-reconfigure, xbps-remove, and xbps-pkgdb. makepkg has similar features to xbps-create. So in the case of pacman, you have six executables instead of one.
2. pacman and the xbps-* executables share four command flags. All or almost all of the rest are different. (That by itself could be search and replace, but combined with everything else...)
3. pacman does GPG package signing. xbps does RSA package signing.
4. pacman has multiple checksums per package (I'm not sure what that indicates). xbps has one, and it's had that feature since before pacman had it.
5. pacman does not store the maintainer in the database. xbps does.
6. pacman stores the target architecture in the package configuration file. xbps does not.
7. pacman has a build dependency on gpg and on curl. xbps does not.
8. xpbs has its own special syntax for common build script commands. pacman uses regular shell script commands.
9. xbps-create builds using a linux container. makepkg does not.
10. xpbs-create supports building for different target architectures without using distcc. pacman uses distcc.

That's too many differences for a re-spin.

[Stellarwind]

My point is that it is a ripoff from arch linux. Just because he has written the package manager based at looking at the code of pacman. If you look at the source code of both pacman and xbps you will see its very similar.
The templates and PKGBUILD files are almost the same only that it has different name conventions. Everyone can edit the package manager and rename some parts of code so that it looks different.

Crosscompiling is supported by makepkg anyhow, you just change the flags of the makepkg.conf and start cross compiling.


edit: I am not saying he is lying about writing the package manager. I am just saying it is a re-spin of the code.

If you write similar software, you'll likely get similar patterns... Just like Linux is a re-spin of Unix or Dalvik is a re-spin of Java.
Or like some people are saying Russian/Chinese Gen5 fighter jets are F22 ripoff, but if you build planes with similar goals, you'll most likely get similar shapes.

[cryptoahash]

Any experienced Linux user or coder will agree with me. Trust me.
Those points he is referring to are easily recoded.

And him knowing so many of the exact differences with pacman is already something to think about, isnt it?

[cryptoahash]

I thought that too, but the following indicates to me it's not a re-spin:
1. pacman has similar features to xbps-install, xbps-query, xbps-rindex, xbps-reconfigure, xbps-remove, and xbps-pkgdb. makepkg has similar features to xbps-create. So in the case of pacman, you have six executables instead of one.
2. pacman and the xbps-* executables share four command flags. All or almost all of the rest are different. (That by itself could be search and replace, but combined with everything else...)
3. pacman does GPG package signing. xbps does RSA package signing.
4. pacman has multiple checksums per package (I'm not sure what that indicates). xbps has one, and it's had that feature since before pacman had it.

pacman _can_ also use just one. This is up to the pkgbuild builder.

5. pacman does not store the maintainer in the database. xbps does.
6. pacman stores the target architecture in the package configuration file. xbps does not.
7. pacman has a build dependency on gpg and on curl. xbps does not.
8. xpbs has its own special syntax for common build script commands. pacman uses regular shell script commands.

Looking at their templates, those are also just bash shell commands.

9. xbps-create builds using a linux container. makepkg does not.

What does that exactly mean anyhow? a linux container. Is this a chroot env? If so, so does pacman.

10. xpbs-create supports building for different target architectures without using distcc. pacman uses distcc.

distcc has nothing to do with targeting different archs. makepkg _can_ use distcc if wanted. It is not a must.

That's too many differences for a re-spin.

So many very very small differences.

[Stellarwind]

Any experienced Linux user or coder will agree with me. Trust me.
Those points he is referring to are easily recoded.

And him knowing so many of the exact differences with pacman is already something to think about, isnt it?

Well, show us what you mean, which similarities you found exactly? So far it looks more like trolling attempt.

[MoonMoon]

What does that exactly mean anyhow? a linux container. Is this a chroot env? If so, so does pacman.

Sometimes a web search engine is indeed useful: https://linuxcontainers.org/
Of course any experienced Linux coder or user will agree with that. Trust me.

[sbolokanov]

Blind cannot see.
What's the point of arguing with a blind man?

On the topic: This Void thing looks interesting. I might play with it in the future.

[cryptoahash]

Sometimes a web search engine is indeed useful: https://linuxcontainers.org/
Of course any experienced Linux coder or user will agree with that. Trust me.

This is a chroot env. What is your point?

[Michael_S]

pacman _can_ also use just one. This is up to the pkgbuild builder.

You have it backwards. pacman uses just one, xbps uses six to do the same thing. I looked, and if people routinely break pacman into pacman-search, pacman-db, pacman-install, etc... I can't find it.

Looking at their templates, those are also just bash shell commands.

Mostly just bash shell commands, but xbps uses vmove, vdir, etc... as aliases for commands. Maybe they do something special, maybe they're just wrapper on plain old mv, mkdir, etc... But either way, why go through the trouble of looking different from pkgbuild for no benefit?

What does that exactly mean anyhow? a linux container. Is this a chroot env? If so, so does pacman.

Slightly different, per the link MoonMoon provided. Again, why would the developer do something different than chroot just for the sake of being different?

distcc has nothing to do with targeting different archs. makepkg _can_ use distcc if wanted. It is not a must.

I misunderstood, then.

So many very very small differences.

At what point do you draw a line between many dozens of "very very small differences" and two separate projects? I think if we keep enumerating differences and calling them insignificant, soon we can happily declare that XBPS is a blatant ripoff of GCC. Or maybe Firefox.