Tada - Ada Package Manager has a website now :)

General
1

See: https://tada.pm

2

Nice.

Do you have any plans for further development?

3

Yes, I do! Tada is usable as-is right now, but I still want to:

  1. Implement better toolchain management - right now Tada expects you to have gnat and gprbuild installed, you can configure different toolachains, but thatā€™s a manual step
  2. MOAR PACKAGES!!11 :ogre: - thatā€™s what Iā€™m working on :slight_smile:
  3. Once there is more packages - better dependency graphs
  4. Once there is more packages - packages index querying, etc.
4

Ada has a 2nd package manager?

5

Crazy times, huh? :slight_smile: I built one for myself, ā€œbecause I canā€. Not planning to harm Alire in any way (itā€™s wonderful it exists!)

6

At least three: this thing, Alire, and Aura

7

To be honest I donā€™t understand your position here. You say you make this project for yourself, or on the website ā€œTada is a project for personal useā€. But you post about it a lot (on this forum, reddit, etc.), and you create a website. So it seems you want other people to use it.

I think this is harming the Ada community more than Alire. Like when I see your Tackle, or padlock projects not available to the community, I canā€™t help to think this is a shame.

8

I agree with Fabien here. I fear that something similar might happen like in Haskell. They have one package repo but 2(,5?) package managers (Cabal v1/v2 and Stack) and for beginners it causes a lot of confusion. Especially for preconfigured tools and editors. In my case Doom Emacs is preconfigured for cabal but I use stack.

Please donā€™tā€¦ As someone who is working to get a SuperH cross compiler working with Alire, I can tell you that toolchain management is hard. You would have build and package the tools. If you decide to use the tools from gnat-fsf-builds what actual benefit do you get? Not only do you have to set your default toolchain (presumably gnat_native) but also per project configuration and their .gpr files (Especially Runtime and Driver).

Could you also package your crates for alire?

I donā€™t mean to sound mean. Tada is a cool project. Itā€™s just that long term it might result into a haskell situation which would be less than ideal :confused:

9

I feel torn between cheering for your enthusiasm and voicing my concerns about fragmentation (as others already have). :broken_heart:

10

Keep doing whatever you find useful.

11

Thanks! This looks really cool. Appreciate all the hard work!

12

Might I suggest, before jumping down ā€œmore content!ā€, that you prioritize #4 & #3.
Iā€™m of the opinion the best way to do this would be via a database amiable IR, not ā€˜textā€™ or ā€˜filesā€™; it is more up-front and design heavy work, but the payoff is things like:

  1. No uncompilable (or at least unparsable) packages.
  2. You already have the DB to manage metainfo, like licenses and dependencies.
  3. You can design the system to validate, via query, the combinations of various packages.

Honestly, this is one area Iā€™d like to address ā€” bootstrapping an Ada compiler, one that is formally verified, should be a matter of ā€œfollow this recipe/procedureā€, a triviality.