Avoid multiple decode/encode of certificates

[erikgb]

          I have a feeling we are decoding/encoding certs in bundles multiple times already, and this adds another round of decode/encode. What about using a "Cert Pool" as an intermediate format between functions - instead of `[]string` or `[]byte`? We could fix this in a follow-up PR. Just wanted to note it here.

Originally posted by @erikgb in #303 (comment)

[arsenalzp]

Hello,
Can I work on this?
If we have the final decision, I cat work on this issue to improve code usage and performance.
I would appreciate any suggestions how can we improve.
As far as I remember CertPool usage was suggested.

[erikgb]

Please do, @arsenalzp! 🎉 Anything that needs to be decided?

[arsenalzp]

Please do, @arsenalzp! 🎉 Anything that needs to be decided?

As usually, during the work process we finding out the best decision :)

[arsenalzp]

/assign

[erikgb]

@arsenalzp Are you working on this issue? If not, I am considering to take a stab. 🤠

[arsenalzp]

@arsenalzp Are you working on this issue? If not, I am considering to take a stab. 🤠

Hello,
Sure, I don't forget about this issue, however I pushed all my strengths to pass one of AWS Architect exam.
Please, don't re-assign this issue to anyone else, I will manage with it soon :)

[arsenalzp]

Sorry for delayed response, I started working on this issue.
I found some changes were made for source.go and sync.go.

[arsenalzp]

Hello,
How do you think, should be maintain CertPool structure between all functions: in source.go, pem.go, cert_pool.go? Or just between pem.go and cert_pool.go and keep as it is in source.go?

[erikgb]

Hello,

How do you think, should be maintain CertPool structure between all functions: in source.go, pem.go, cert_pool.go? Or just between pem.go and cert_pool.go and keep as it is in source.go?

I am a bit out of context, but if it makes sense to keep the cert pool from source to target, I will imagine that would make the code more compact and easier to read/understand?