Note: I have not built the package, I have low spec hardware.

I just wanted to bump the update quickly, so I just changed the name and got the hash from running:

nix-prefetch-url "https://git.savannah.gnu.org/cgit/emacs.git/snapshot/emacs-29.1.tar.gz"

Let me know if there is some mistake or some major changes, as to take more steps, rather than simple name change, so I can close this PR.

Note: I have not built the package, I have low spec hardware.

I just wanted to bump the update quickly, so I just changed the name and got the hash from running:

nix-prefetch-url "https://git.savannah.gnu.org/cgit/emacs.git/snapshot/emacs-29.1.tar.gz"

Let me know if there is some mistake or some major changes, as to take more steps, rather than simple name change, so I can close this PR.

For hash attribute, value needs to be in SRI format, which you can fetch using this.

Also, don't forget to add the label backport release-23.05.

#245068 (comment)

Thank you for the clarification!

Also all-packages.nix should be change to emacs = emacs29;

But only on unstable, right?

Also all-packages.nix should be change to emacs = emacs29;

But only on unstable, right?

Should I change that?

Then there is no point for pkg.emacs29-* (which currently exist on unstable?)
Also in that case, new addition is also expected:

emacs-pgtk = emacs29-pgtk

Also all-packages.nix should be change to emacs = emacs29;

But only on unstable, right?

Should I change that?

Then there is no point for pkg.emacs29 (which currently exist on unstable?) Also in that case, new addition as, is also expected:

emacs-pgtk = emacs29-pgtk

The problem is that the version bump should apply to both unstable and release-23.05.
But if you add these other changes as well, then that's going to break release-23.05, since the release should not alter the (major) versions of the packages included.

I would do these changes in a separate PR, pertaining only to unstable.

But I'm a NixOS novice, so let's hear also to more experienced people.

Thanks for the PR, BTW!

Please don't change the main attribute in this PR. Let's just get the bump merged first and then do a follow-up changing the main attribute.

That requires release notes and more thorough testing.

I'm newbie at this matter lol, I just opened pr to notify and just change the hash.

Glad to contribute to nixos, and it is such a wonderful place.

Somebody, or you guys can continue with other steps/PR.

Emacs overlay auto-updated a couple of hours ago and we've been running hydra jobsets for 29.1 on the nix-community hydra in https://hydra.nix-community.org/eval/400298?filter=-unstable&compare=399847&full= with no visible regressions.

         specified: sha256-KYy0T/wEMEE0NdiIjzVQB1vTkmtwJGdJ0w28tLyf0sw=
            got:    sha256-3HDCwtOKvkXwSULf3W7YgTz4GV8zvYnh2RrL28qzGKg=

@idlip While I appreciate your enthusiasm you need to apply a bit of testing before opening a PR.
This is too low of an effort to meet our bar. Next time at the very least build the derivation that you want to contribute.

While not all checkboxes in the PR template are mandatory you should at least check some of them.

I am going to close this one and open my own bump PR as you already state that you are unable to build anything.

#246203

         specified: sha256-KYy0T/wEMEE0NdiIjzVQB1vTkmtwJGdJ0w28tLyf0sw=
            got:    sha256-3HDCwtOKvkXwSULf3W7YgTz4GV8zvYnh2RrL28qzGKg=

@idlip While I appreciate your enthusiasm you need to apply a bit of testing before opening a PR. This is too low of an effort to meet our bar. Next time at the very least build the derivation that you want to contribute.

While not all checkboxes in the PR template are mandatory you should at least check some of them.

I am going to close this one and open my own bump PR as you already state that you are unable to build anything.

TIL: The output of nix store prefetch-file "url" (or nix-prefetch-url "url", for that matter), don't produce the same hash (assuming same encoding of the hash) as the evaluation of the derivation. It seems that nix does not just hash the downloaded .tar.gz, it first expands the archive, normalizes it, and then obtains the hash of the tree.

To sum up, there is no alternative to a) bump the version and b) run nixos-rebuild test or switch and obtain the expected hash from there.

Again, I'll leave up to more experienced people to comment on this.

To sum up, there is no alternative to a) bump the version and b) run nixos-rebuild test or switch and obtain the expected hash from there.

Well, you can just build the package without having to eval the entire nixos.

cd $WHEREEVER_YOU_CLONED_NIXPKGS
git checkout -b my_local_version upstream/nixos-unstable # less rebuilding when working against unstable compared to master
# now bump the version, and then
nix build -f . emacs-29  # build whatever thing you are working on - it will show you the proper hash here

Much faster than running nixos-rebuild.

I will write a tutorial about this.
Nonetheless, everything was already explained on the wiki:

https://nixos.wiki/wiki/Nixpkgs/Create_and_debug_packages