**Motivation**

Create an ethrex release when a semver tag is pushed to the repo

**Description**

- Modifies already existing L2 releases workflow to build 
- For `ethrex client` (right now only for L2 but can be used for both L1
and L2 after #3381 is merged )
     - Build for `linux-x86_64` , `linux-arm64`, `macos-arm64`
  - For `ethrex prover client`
    - For `exec` mode
       - Build for `linux-x86_64` `linux-arm64` and `macos-arm64`
     - For `sp1 gpu` mode 
         - Build for `linux-x86_64` `linux-arm64`
     - For `risc0 gpu` mode
        - Build for `linux-x86_64`
 - For `ethrex replay`
    - For `exec` mode
       - Build for `linux-x86_64` `linux-arm64` and `macos-arm64`
     - For `sp1 gpu` mode 
         - Build for `linux-x86_64` `linux-arm64`
     - For `risc0 gpu` mode
        - Build for `linux-x86_64`
  - Creates a release with
- Changelog from all the changes between the previous tag and the newly
created one.
- All the built binaries and the rollup L1 and L2 contracts in a tar
archive
- Example from testing repo
https://github.com/LeanSerra/ethrex/releases/tag/v0.0.7-rc.1 before
adding the contracts to the release output
- Another example with the contracts
https://github.com/LeanSerra/ethrex/releases/tag/v0.0.7-rc.2
- Example of all outputs
https://github.com/lambdaclass/ethrex/actions/runs/16077779945
- Non related change: pin the version for the docker image in sp1 
- With #3381 merged we now build ethrex with --all-features so the
default database was changed to libmdbx when initializing the store.
**Other considerations**

> Q: No sp1 cpu?
> A: Right now it's too slow to be viable in production environments

> Q: No sp1 macos?
> A: sp1 does not support the metal api for gpu acceleration. Also
because zkvms are built using docker we'll run into a problem where the
macos github runner does not have docker installed.
> After a quick investigation into this issue we have to use
[colima](actions/runner#1456 (comment))
for docker.
> This leads into another issue where sp1 uses a docker image built only
for amd64 this requires nested virtualization support that is enabled in
apple M3 chips or later but [the runner is currently using M1
chips](https://docs.github.com/en/actions/concepts/runners/about-larger-runners#limitations-for-macos-larger-runners).

> Q: Why does building the sp1 prover for
[arm-64](https://github.com/LeanSerra/ethrex/actions/runs/16010958409/job/45168455867)
take twice as long as building for
[x86-64](https://github.com/LeanSerra/ethrex/actions/runs/16010958409/job/45168455854)
> A: We have to use QEMU inside the `arm-64` runner because the sp1
docker image is only built for `amd64`

> Q: wen risc0
> A: ~After #3172 is merged~
[Now](https://github.com/lambdaclass/ethrex/actions/runs/16036016678/job/45247866997)

…e binary for both L1 and L2 (lambdaclass#3381)

> [!WARNING]
> Merge after lambdaclass#3371

**Motivation**

We want `ethrex` to be a single binary for both running L1 and L2. For
this, we need to remove the `l2` feature flag from `cmd/ethrex`, a work
that includes removing the same feature flag from the following crates:

1. `ethrex-vm`.
2. `ethrex-levm`.
3. `ethrex-blockchain`. 

1 and 2 are removed in lambdaclass#3367,
and 3 is removed in lambdaclass#3371.

**Description**

Removes `l2` feature flag from `cmd/ethrex`

**Motivation**

Create an ethrex release when a semver tag is pushed to the repo

**Description**

- Modifies already existing L2 releases workflow to build 
- For `ethrex client` (right now only for L2 but can be used for both L1
and L2 after lambdaclass#3381 is merged )
     - Build for `linux-x86_64` , `linux-arm64`, `macos-arm64`
  - For `ethrex prover client`
    - For `exec` mode
       - Build for `linux-x86_64` `linux-arm64` and `macos-arm64`
     - For `sp1 gpu` mode 
         - Build for `linux-x86_64` `linux-arm64`
     - For `risc0 gpu` mode
        - Build for `linux-x86_64`
 - For `ethrex replay`
    - For `exec` mode
       - Build for `linux-x86_64` `linux-arm64` and `macos-arm64`
     - For `sp1 gpu` mode 
         - Build for `linux-x86_64` `linux-arm64`
     - For `risc0 gpu` mode
        - Build for `linux-x86_64`
  - Creates a release with
- Changelog from all the changes between the previous tag and the newly
created one.
- All the built binaries and the rollup L1 and L2 contracts in a tar
archive
- Example from testing repo
https://github.com/LeanSerra/ethrex/releases/tag/v0.0.7-rc.1 before
adding the contracts to the release output
- Another example with the contracts
https://github.com/LeanSerra/ethrex/releases/tag/v0.0.7-rc.2
- Example of all outputs
https://github.com/lambdaclass/ethrex/actions/runs/16077779945
- Non related change: pin the version for the docker image in sp1 
- With lambdaclass#3381 merged we now build ethrex with --all-features so the
default database was changed to libmdbx when initializing the store.
**Other considerations**

> Q: No sp1 cpu?
> A: Right now it's too slow to be viable in production environments

> Q: No sp1 macos?
> A: sp1 does not support the metal api for gpu acceleration. Also
because zkvms are built using docker we'll run into a problem where the
macos github runner does not have docker installed.
> After a quick investigation into this issue we have to use
[colima](actions/runner#1456 (comment))
for docker.
> This leads into another issue where sp1 uses a docker image built only
for amd64 this requires nested virtualization support that is enabled in
apple M3 chips or later but [the runner is currently using M1
chips](https://docs.github.com/en/actions/concepts/runners/about-larger-runners#limitations-for-macos-larger-runners).

> Q: Why does building the sp1 prover for
[arm-64](https://github.com/LeanSerra/ethrex/actions/runs/16010958409/job/45168455867)
take twice as long as building for
[x86-64](https://github.com/LeanSerra/ethrex/actions/runs/16010958409/job/45168455854)
> A: We have to use QEMU inside the `arm-64` runner because the sp1
docker image is only built for `amd64`

> Q: wen risc0
> A: ~After lambdaclass#3172 is merged~
[Now](https://github.com/lambdaclass/ethrex/actions/runs/16036016678/job/45247866997)