Add "Validating attribute values" section to user_doc #54
Conversation
Adapted the discussion thread [1] to document the expected `valid` key behavior. [1]: kaitai-io/kaitai_struct#435
|
Thanks a lot for starting this one! I will add some more, but it's a good clean start. |
|
|
||
| When a value does not meet the specified criteria, Kaitai Struct raises a validation error, halting further parsing. This preemptive measure ensures the data being parsed is within the expected domain, providing a first layer of error handling. | ||
|
|
||
| NOTE: The actual implementation of validation checks is language-dependent and may vary in behavior and supported features across different target languages. |
There was a problem hiding this comment.
I don't understand why this note is here - it goes against the principles of Kaitai Struct. Once everything is correctly implemented in all target languages (which may with any KS feature temporarily not be the case, but it's simply the reality that individual target languages are in various stages of development at any point in time), the checks will not vary in behavior across different target languages. The status (passed/failed) of any check in one language must be consistent with the status of the same check in the same situation in any other language, any violation is a bug that will be eventually fixed.
@mblsha Did you have something specific in mind when you wrote this sentence? I've just checked the Valid* tests at https://ci.kaitai.io/ and I don't see any current issues with valid, except for Nim, Perl and Construct (all of which are somewhat niche targets), where valid presumably isn't implemented at all, but I don't see any behavior variations.
I remember that valid still doesn't have type checks (kaitai-io/kaitai_struct_formats#678 (comment)), so perhaps you mean the implications of that.
There was a problem hiding this comment.
I believe what @mblsha meant here was that exception handling by itself is not a universal mechanism by any stretch, and indeed individual exception names and notation will vary widely between languages. That's at least how I interpreted this. I certainly won't argue that we should also have language-dependent explanations on how these validation exceptions (or error codes, or whatever it will manifest in) work for specific languages.
|
@generalmimon should I delete this "NOTE:" from the doc then? |
Adapted the discussion thread 1 to document the expected
validkey behavior.