Ada Programming Language/Alire

Ada Library Repository (a package manager for Ada)

Maxim Reznik

@reznikmm

@chriseberly_twitter ada language server uses libadalang_tools internally, so formatting should be the same as with gnatpp.

1 reply

Jeremy Grosser

@JeremyGrosser

I changed rp2040_hal to depend on gnat_arm_elf ^12.2. it works fine on my machine, but the github action is getting stuck trying to solve that dependency... what am I missing? https://github.com/JeremyGrosser/rp2040_hal/actions/runs/3394269053/jobs/5642615851

oh, I needed to update the github action to alire 1.2.1... I didn't realize we have separate index branches for minor releases

weird that it got stuck in the solver rather than just erroring out

Fabien Chouteau

@Fabien-Chouteau

Solver should timeout eventually

Stephen Leake

@stephe-ada-guru

I'm trying to run the ada_language_server tests, so I can duplicate a bug I'm getting with Emacs ada-mode and ada_language_server. it wants python module e3, but python pip says there's no such module. Anyone know where that might be? (I'll also post an ada_language_server issue, but I've had better response time here).

Irvise

@irvise:matrix.irvise.xyz

[m]

@stephe-ada-guru: I think it is called e3-core pip install e3-core

1 reply

Taken from https://github.com/alire-project/GNAT-FSF-builds

Stephen Leake

@stephe-ada-guru

Gnu Emacs Ada mode 8.0 beta is now available in GNU ELPA devel for
beta testing.

All Ada mode executables can now be built with Alire
(https://alire.ada.dev/); this greatly simplifies that process.

gpr-query and gpr-mode are split out into separate GNU ELPA packages.
You must install them separately (Emacs install-package doesn't
support "recommended packages" like Debian does).

Ada mode can now be used with Eglot; this is controlled by new variables:

ada-face-backend - one of wisi, eglot, none

ada-xref-backend - one of gnat, gpr_query, eglot, none

ada-indent-backend - one of wisi, eglot, none

The the indent and face backends default to wisi if the wisi parser is
found in PATH, to eglot if the Ada LSP server is found, and none
otherwise. The xref backend also looks for the gpr_query executable in
PATH.

The current AdaCore language server (23) support face but not indent.
The current version of eglot (19) does not support face. So for now,
eglot + ada_language_server only provides xref.

The AdaCore language server ada_language_server is installed with
GNATStudio (which ada-mode will find by default), or can be built with
Alire. If you build it with Alire, either put it in PATH, or set
gnat-lsp-server-exec.

I have not tested ada-mode with lsp-mode. You can set ada-*-backend to
'other to expermiment with that, or tree-sitter, or some other
backend.

To access the beta version via Gnu ELPA, add the devel archive to
package-archives:

(add-to-list 'package-archives (cons "gnu-devel" "https://elpa.gnu.org/devel/"))

Then M-x list-packages; the beta release shows as ada-mode version
8.0.3.0.20221106.55317, wisi version similarly.

Please report success and issues to the Emacs ada-mode mailing list
https://lists.nongnu.org/mailman/listinfo/ada-mode-users.

The required Ada code requires a manual compile step, after the normal
list-packages installation:

cd ~/.emacs.d/elpa/ada-mode-7.3beta*
./build.sh
./install.sh

If you have Alire installed, these scripts use it. Otherwise, this
requires AdaCore gnatcoll packages which you may not have installed;
see ada-mode.info Installation for help in installing them.

Maxim Reznik

@reznikmm

@stephe-ada-guru Any idea what request eglot uses for indent?

Stephen Leake

@stephe-ada-guru

LSP defines the textDocument/rangeFormatting request, which does pretty-printing, which includes indent. So that's what ada-eglot uses. Currently partly broken in ada_language_server 23; I just submitted a patch for it. The libadalang format subprogram takes options to control what changes to allow, including "indent only"; it would be nice if there was a way to pass those options to ada_language_server thru some extension.

1 reply

Nathan Rignall

@NathanRignall

Is there any plans for Alire to provide prebuilt executables and prebuilt toolchains for arm64? Now that there are many devices that use this architecture and that it is clear Microsoft is also aiming to make Arm Windows more of a thing it would be cool to see!

6 replies

DrPi

@drpi:matrix.org

[m]

@reznikmm: In the ada_language_server repository, there are 2 ada.tmLanguage.json files, one in syntaxes folder and another one in advanced folder. What are the differences (except the file size) ?

2 replies

ajdude

@ajdude:matrix.org

[m]

It also may not be needed on the Mac at the moment, but, I was successfully able to build a mac aarch64 alr binary that I use on my m2 air. Very easy with Simon's gnat distribution for the Mac and an existing copy of alire to recompile itself.

ajdude

@ajdude:matrix.org

[m]

:point_up: Edit: It also may not be needed on the Mac at the moment, but, I was successfully able to build a mac aarch64 alr binary that I use on my m2 air. Very easy with Simon's gnat distribution for the Mac and an existing copy of alire to alr build itself.

Bent Bracke

@bracke

Is it possible to build a non-alire package like templates-parser with a Alire installed toolchain?

16 replies

RajaSrinivasan

@RajaSrinivasan

cross build for ARM Cortex M4

ajdude

@ajdude:matrix.org

[m]

My only request, if alire is distributed for other systems, is to keep with the arch naming convention, so an M1/M2 macos release would end in bin-aarch64-macos.zip for example.

RajaSrinivasan

@RajaSrinivasan

Cross build for ARM Cortex M4 - I would like to write some algorithms in Ada to be included in an ARM target. ThreadX executive. Our app is mainly C++ now. The IDE eclipse can allow external objects I think. Question is - what does an alr environment look like? in other words how do i create this alr environment and produce just obj files. Hope this is clear. Thanks for any ideas.

Question about ImageMagick. There are references to something called G2F but that does not lead anywhere. Anyone know of where we can find this. Alternatively anyone know of an Ada binding for imagemagick?

Irvise

@irvise:matrix.irvise.xyz

[m]

@RajaSrinivasan: I do not know if this will be helpful to you... But if you are in Alire project, you can just run alr printenv which will print the environment variables and setup it does to build/run the environment. That should give you the info you need for eclipse afaik.

RajaSrinivasan

@RajaSrinivasan

@irvise:matrix.irvise.xyz thanks. will give it a shot

Irvise

@irvise:matrix.irvise.xyz

[m]

Hi. A new Alire user is having issues with it in Void Linux using MuslC. The Alire binaries cannot run there as they use GlibC. Is there a way we could provide muslc-built Alire?

2 replies

Maxim Reznik

@reznikmm

will alr be supported on ubuntu-18.04 after its deprecation on GitHub Actions actions/runner-images#6002

2 replies

Irvise

@irvise:matrix.irvise.xyz

[m]

@reznikmm: afaik, it should. It is distributed as either a binary or AppImage. There is very little dependence on the base/build OS.

ajdude

@ajdude:matrix.org

[m]

@Fabien-Chouteau: @mosteo

Regarding AJ-Ianozi/getada#1

I am interested in comforting with whatever standards you are interested in setting. A reason why I haven't messed with bash-completion yet a was because I wasn't sure if it should go in .config/ or somewhere else. I can definitely have the binary for alire live in .getada/bin instead of .alire/bin and just leave .alire alone if that's what you want, I was basing it off https://github.com/alire-project/alire/issues/1164#issuecomment-1231799016

I'll work towards leaving .alire out of it

Bent Bracke

@bracke

My fork of alire-index does not know about stable-1.2.1 - it stops at stable-1.1, but according to github it is fully synced. Any pointers?

Rolf Ebert

@RREE

When forking you have to check a box that says sth like 'all branches'. If not your fork does not see any other or newer branch.

Richard Lewis

@smoothsailing_gitlab

Should I be able to create C header files from spec files? This link https://gcc.gnu.org/onlinedocs/gnat_ugn/Running-the-C-Header-Generator.html suggests that I should.

Barry Houser

@barryhouser:matrix.org

[m]

Alire is an acronym for Ada LIbrary REpository. By weird coincidence, it's also the surname of Camila Alire, a Librarian and former president of the American Library Association from 2009 to 2010. She was later appointed as a member of the National Council on the Humanities by President Barack Obama in 2012. She even has her own Wikipedia page:

https://en.wikipedia.org/wiki/Camila\_Alire

A cool easter egg could be launch her wikipage from the browser by typing alr camila.

Barry Houser

@barryhouser:matrix.org

[m]

:point_up: Edit: Alire is an acronym for Ada LIbrary REpository. By weird coincidence, it's also the surname of Camila Alire, a Librarian and former president of the American Library Association from 2009 to 2010. She was later appointed as a member of the National Council on the Humanities by President Barack Obama in 2012. She even has her own Wikipedia page:

https://en.wikipedia.org/wiki/Camila\_Alire

A cool Alire easter egg would be to launch her Wikipedia page from the browser by typing: alr camila 😆🤣

https://en.wikipedia.org/wiki/Camila\_Alire

A cool Alire easter egg would be to launch her Wikipedia page from the browser by typing: alr camila Even cooler would be to generate her photo in ASCII text. 😆🤣

https://en.wikipedia.org/wiki/Camila\_Alire

A cool Alire easter egg would be to launch her Wikipedia page from the browser by typing: alr camila It would be even cooler to generate her photo in ASCII text at the prompt. 😆🤣

Bent Bracke

@bracke

@RREE There seems to be no way to set that option after forking. So I guess I will have to "refork"?

Manuel

@mgrojo

I suppose you can pull all branches from the official repository directly to your local repository using git pull --all url and then push to your fork in github.

Bent Bracke

@bracke

@mgrojo I will try that, when I get back to my computer.

fatih incecayir

@MilliUser

Hello. It's been a while since I met ADA. I started a project with Alire, but I noticed something. Even though I have never written any code, there is something like 4 bytes full in RAM. I feel like the program I wrote in C cannot fit in RAM with ADA. The MCU I am using is STMF030C8T6 with 64K flash and 8K RAM. I would like to benefit from your experience. Thanks in advance

4 replies

Irvise

@irvise:matrix.irvise.xyz

[m]

Are you using -Os?

fatih incecayir

@MilliUser

@irvise:matrix.irvise.xyz no I do not use

dkm

@marcat:matrix.org

[m]

The easiest would be to put it in some public git repository (e.g. github, sourceberg, sourcehut, whatever) and point us to it. If you don't have any, then a tgz would be ok too :)

fatih incecayir

@MilliUser

ok. thanks

dkm

@marcat:matrix.org

[m]

:point_up: Edit: The easiest would be to put it in some public git repository (e.g. github, codeberg, sourcehut, whatever) and point us to it. If you don't have any, then a tgz would be ok too :)

fatih incecayir

@MilliUser

sample.tar

8 replies

RomanJanousek

@RomanJanousek

Hi all, can I ask you for help? I reinstalled ALIRE and it does not go. Error message - Cannot find the powershell file. OS WIN 10 HOME, PowShell 7. Thank you in advance for your advice. Roman

dkm

@marcat:matrix.org

[m]

   New solution is incomplete.
   +⚠ stm32f0_hal ~0.1.0-dev (new,missing)

@simonjwright
Which does not surprise me as I don't see the crate in the alire index (the closest is my stm32f0x2_hal crate)

2 replies

fatih incecayir

@MilliUser

stm32f0_hal.tar

sorry, this is my library. Created using alire and svd2ada

Holger Rodriguez

@hgrodriguez

I am using gnat_arm_elf in my RP2040 based projects, and just realised, that the version 12.2.1 gives me a new warning:
main has a LOAD segment with RWX permissions
While I do understand the warning message itself (to some extend), I do not understand where this comes from, as I never changed any settings regarding this. How to fix this, as the code should be RX, and RAM should be RW, so how to segregate this correctly to make this right?

Jeremy Grosser

@JeremyGrosser

The warning was added by the new version of binutils 12.2.1 is built against. The linker flags are meaningless on embedded targets without an ELF loader. This warning is safe to ignore, which is already done on the master branch JeremyGrosser/rp2040_hal@ed21f38

Holger Rodriguez

@hgrodriguez

:thumbsup:

Where communities thrive

People

Repo info

Activity