Improve performance of creating access rules

[hculea]

In order to check if the optimisation is effectual, I have run the following script, which populated the priorly created optimization-test directory with 1000 secrets:

u, err := secrethub.NewClient()
if err != nil {
	fmt.Print(err.Error())
}

iterations := 1000
for i := 0; i < iterations; i++ {
	_, err := u.Secrets().Write(fmt.Sprintf("<SECRETHUB NAMESPACE>/optimization/optimization-test/secret%d", i), []byte(fmt.Sprintf("value%d", i)))
	if err != nil {
		fmt.Print(err.Error())
	}
}

I have then replaced the CLI's SDK dependency by pasting the following in CLI project's go.mod:

replace (
    github.com/secrethub/secrethub-go => <RELATIVE PATH TO secrethub-go FOLDER>  // in my case, ../secrethub-go
)

and then I rebuilt the CLI.

NB, I was on the feature/refactor-and-test-acl branch on the SDK.

The effect of the pull request can be observed by comparing the running times of an acl set command run on a directory with a large number of secrets (such as the one created by the script above), before and after the change, which, in my experience, differed by quite a large factor.

If anyone considers this necessary, I will measure the actual times for both the before and after cases, making use of the duration library.

[SimonBarendse]

Sweet stuff! 👌

I've tested this on a huge repo (2000 secrets). On develop it takes a whopping 231s to create an access rule for the root, with these changes applied it takes 28s. Although that's still a long while (the encryption itself also takes long), it's a huge improvement! Thanks for adding this! 💙

I've pushed a couple of changes to the test suite. Could you have a look at them? I've added some explanation in the commit messages. If you agree with what I've done, this is ready to go 🚀

[hculea]

Thanks for the touch ups! Everything looks good to me too! I'll merge now.