maybejustjames / yaml Goto Github PK

Input: ""

If I use Yaml.Decoder.list Yaml.Decoder.int I expect to get an empty list, instead I get Expected list, got: Null

Given the input:

 [ '  hello  world  ' : 3 ]

I expect the parser to produce:

Ast.List_ [ Ast.Record_ (Dict.fromList [("  hello  world  ", Ast.Int_ 3)])]

Input:

-

Expected to decode with list null but instead Expected list, got: \"-\"

This causes the comment to be parsed as part of the next element.

"""[ aaa,# A comment
     bbb, # a dumb comment
     ccc
   ]"""

is parsed as:

List_ [String_ "aaa",String_ "# A comment\n           bbb",String_ "ccc"]

but should be parsed as:

List_ [String_ "aaa",String_ "bbb",String_ "ccc"]

From the YAML documentation:

Normally, YAML insists the “:” mapping value indicator be separated from the value by white space.
A benefit of this restriction is that the “:” character can be used inside plain scalars, as long as it is
not followed by white space. This allows for unquoted URLs and timestamps. It is also a potential
source for confusion as “a:1” is a plain scalar and not a key: value pair.

Failing example:

Yaml.fromString Yaml.string "a:b"

Expected result

"a:b"

Actual result

Expected string, got: { a: \"b\" }

I need to encode multi-line strings within a record, that often contain quotes. As well, I have octal permissions such as 0644 that yaml would parse as an octal numeric literal. I would expect Encode.string to either properly escape and quote this, or generate a yaml multi-line string block.

escapeme: "0700"
oneline: "multi line string\nwith \"quoted\" text"
multiline: |-
  multi line string
  with "quoted" text

For now I'm working around this with the following:

safeEncodeString : String -> YE.Encoder
safeEncodeString =
    encodeNewlines
        >> encodeQuotes
        >> String.Extra.quote
        >> YE.string


encodeNewlines : String -> String
encodeNewlines =
    String.replace "\n" "\\n"


encodeQuotes : String -> String
encodeQuotes s =
    if String.contains "\"" s then
        String.replace "\"" "\\\"" s

    else
        s

Edit:

While reading https://ruudvanasseldonk.com/2023/01/11/the-yaml-document-from-hell for other potential pitfalls, I've noted:

• record keys may be parsed in odd ways, and the easy approach may just be to quote them all
• strings starting with an asterisk (references), exclamation point (tags), and ampersand (anchors) must be quoted

I'm of the opinion that yaml encoding here should be as aggressively safe as possible, and parsing and encoded document should be identical regardless of the yaml specification used when parsing. String escaping and quoting should probably be the default, and multi-line strings might just be too problematic to deal with (at the minimum, needing to handle trailing newlines properly)

Anchors and references are documented in the spec here.

• https://github.com/yaml/yaml-test-suite/

Thanks @ypyl

Reference: #25 (comment)

E.g.

Yaml.fromString "0" Yaml.string

Produces this as an error message:

Expected string

A more useful message would be something like:

Expected string, got: 0 (int)

An overview of completeness in comparison to elm/json would be great!

With the encoder:

import Yaml.Encode exposing (toString, record)
toString 2 (record [("name", record [])])

name:{}

but it should be:

name: {}

space after the colon. See here.

Originally posted by @ypyl in #28 (comment)

Given the YAML:

::

I expect the parser to produce:

Ast.Record_ (Dict.fromList [(":", Ast.Null_)])

Hello,

It seems like the library ignores duplicated keys silently

root:
  item1:
    name: Item1
  item1:
    name: Item 2

and it seems like cause Dict.fromList is used on the list without duplication check. Is it expected behavior?
It feels for me that it should produce an error in such cases. Or am I wrong?
Thanks

Another case that, perhaps, can be improved https://ellie-app.com/k5NYB5sL7zDa1

If there is a list of records, each record will be rendered at a new line:

It seems like the following looks better:

If I use Yaml.Decode.value there is no way to Encode the resulting Value.

It seems like the library only put empty line for such cases:

It seems that it is better to use inline type of dict or record in such cases, like:

What do you think?

it seems the parser converts multiline string to simple string, e.g.

description: |
  Some information

will be saved internally as

description: | Some description

My assumption is that it should be like

description: |\n  Some description

The input:

"aaa: aaa"

causes a Parser error:

Err "TODO deadEndsToString"

but should produce:

Record_ (Dict.fromList [("aaa",String_ "aaa")])