Sam Alba
Sam Alba
My 2c, the original name made more sense to me. This one implies a specific secret provider implementation, which wasn't the intent when @jlongtine opened it.
Going through some issues cleaning, I tried to repro this one on a vanilla bash 5.2 with bash-completion and dagger v0.12.4 freshly installed. ```sh $ bash bash-5.2$ . /opt/homebrew/etc/bash_completion bash-5.2$...
Looking at the thread, it looks like we need to expand `~` in the SSH_AUTH_SOCK. However it most cases, it's expected to be an absolute path (at least ssh-agent does...
Just piling up on the consensus for using `-M`, it sounds more consistent. 👍
Closing, feel free to re-open @noah-troncoso if you need help.
> _Quoted from @gerhard:_ > The biggest downsides of this vertical scaling approach is having an instance which we need to pay for even when it is not being used....
Not sure what to think about this one. Even [Github complains about the commit that the tag resolves to](https://github.com/Dudesons/daggerverse/commit/5a0943df5ceca43ab24309e5331f0bac26208a58): ``` This commit does not belong to any branch on this...
Shouldn't the concurrency be managed at the language-level in that case (go routines with a waitgroup)?
Loading the user's \`.gitconfig\` would break the pipeline's portability (it would behave differently based on contextual data). Instead, I'd support it at the API level with the ability to change...
Also, I could not find any mention of that module digest in the docs, I assume it was already solved.