Closed antonowano closed 2 months ago
In what situation do you intend to use it?
I think this is a great idea, because when storing the file in several places (local, s3), I only have to configure one chain and do not have to rewrite the code.
Sure, but can you explain what situation you're using it in/for? Is this a migration strategy? Or is it a redundancy strategy? Previously we had an implementation that helped during migrations by self replicating. Writing to multiple locations is only a fraction of the logic behind these types of implementations. And saying "I think it's a great idea" is not providing any context as to the problem you're trying to solve. Also, "storing the file in several places" is a solution to a problem, I'm trying to figure out what problem you're solving.
The only problem I'm currently solving is the cleanliness of my code. I generate a file and store one copy in local, the other in s3.
Why not hide it behind a function in that case?
There is no problem with this. Perhaps I formatted the issue incorrectly. This issue is related to the proposal and in my case does not solve any serious problems. But perhaps there will be people whose problems this feature will help solve.
The only problem I'm currently solving is the cleanliness of my code. I generate a file and store one copy in local, the other in s3.
@antonowano Maybe my replicate adapter can help you.
@antonowano Maybe my replicate adapter can help you.
Interesting solution, I'll definitely take a look!
@ajgarlag
What differences have I noticed between replication adapter and ChainWriter.
replication adapter is good:
ChainWriter is good for:
Summary
We can bring out the best in everyone and do something new. I want to hear what @frankdejonge has to say about this.
Feature Request
I propose to implement a file system chain class based on the FilesystemWriter interface. To write and edit files in multiple file systems at once.
Scenario / Use-case