Support loading comments
Arianna2028 opened this issue · 3 comments
Context
We use this library to parse and modify Terraform files. We do this by reading a particular resource using this library, then using the data from that to write a new resource back to the file in the same place, removing the old one (using with_meta
to know where it was). However, since no comments are output, we end up removing any that were present inside the original block.
Possible Enhancement
If possible, it would be really helpful to optionally include parsed comments in the output from loads
/load
.
For example, if I parse a resource like this:
resource "aws_ec2_tag" "example" {
# This is a comment
resource_id = aws_vpn_connection.example.transit_gateway_attachment_id
key = "Name"
value = "Hello World"
}
I currently get the following dictionary representation:
{
"resource": [
{
"aws_ec2_tag": {
"example": {
"resource_id": "${aws_vpn_connection.example.transit_gateway_attachment_id}",
"key": "Name",
"value": "Hello World",
"__start_line__": 1,
"__end_line__": 6
}
}
}
]
}
But maybe comment data could be added alongside it, something like:
{
"resource": [
{
"aws_ec2_tag": {
"example": {
"resource_id": "${aws_vpn_connection.example.transit_gateway_attachment_id}",
"key": "Name",
"value": "Hello World",
"__start_line__": 1,
"__end_line__": 6,
"__comments__": [
{
"__start_line__": 2,
"__end_line__": 2,
"text": "This is a comment"
}
]
}
}
}
]
}
Seems would be helpful for some cases but requires making some keys reserved.
for instance, what if comments will be key itself?
Yeah, I guess it would mean the __comments__
key would be reserved. Isn't this technically true for __start_line__
and __end_line__
already as well, though?
Alternatively, maybe __comments__
could be renamed to something with characters that are not considered valid in Terraform, so there is no possible conflict? I believe symbols are not considered valid syntax, so that might work.