[feature request][archive_file] Allow source_file to be list of files

[hashicorp]

[feature request][archive_file] Allow source_file to be list of files

[silverdaz]

At the moment, it is not possible to archive 2 folders or even one folder and another file (located somewhere else).

If I am missing something, could you please point me into the right direction?
I am currently using the external provider, and created my own shell script that zips the required files and folders. Unfortunately, a terraform plan/apply recreates all the resources depending on it.

I could create a cached version, but isn't it better to include that exact functionality into this provider?
Thanks in advance, that'll be very helpful.
(PS: I like the above syntax for source_files and source_dirs)

[raymondhardynike]

Part of this can be done if you use the excludes functionality that i added. It is not as cool as adding each file one at a time but you can exclude all but the files you want.

I started to look into how to get this type of functionality into the code and it would require some serious changes to the way it gets data into the archive function.

If the core was to take all data objects and file paths and shove them into a hash table before adding each thing to an archive then it would be able to do things like append to zip etc because then it would be as simple as changing that hash table.

example table
/somedir/file1, ondisk
/somedir/file2, ondisk
/somedir/file3, ondisk
/somedir/file5, text, "my text data"
/file7, blob, rawblob

Then the problem becomes what has priority. I.E. does a terraform text data object have priority over a file on disk etc.

[sorjef]

@silverdaz To prevent terraform from recreating resources each time the zip script runs you can use a solution I suggested under another issue #6 (comment). If this does not help, check out the comments in that issue for other options.