Skip field when `bcl` tag is equal to `"-"`

[condrove10]

Hello,

As the title says, it would be great it the parser would skip the block if the struct has bcl:"-":

type logging struct {
	Level     string     `bcl:"level"`
	SlogLevel slog.Level `bcl:"-"`
	Format    string     `bcl:"format"`
}

---

• Also could it possible to add an ulterior check for the primitive type of an enum, this way special fields like slog.Level could be parsed?

[wkhere]

Hi @condrove10!

I think the struct tag bcl:"-" is not needed, because BCL's reflection code allows having a struct field not covered by the content of a block. (On the other hand, every block field needs to have corresponding struct field - I plan to relax it via an option, so it'd be more like with json/yaml).

See the commit f3e9403 - I didn't need to treat the DontCare struct field in any special way, it is simply not filled.

By the way, if you excuse me being picky about the vocabulary, I believe you meant to "skip the field", not a block. Skipping the block would make sense for an nested block, but then the field being skipped should be a nested struct as well. I believe this is not what you meant now :) More on this will be in another thread - btw, I am preparing a sketch of a new syntax to cover all common bind cases, we'll discuss it separately (edit: that's the issue #6).

[wkhere]

As for the 2nd request about types mismatch, I guess the AssiblebleTo check is failing. Can you construct a failing test case?

[wkhere]

Hey again.

I guess what you want to achieve is to get BCL's level field parsed straight to slog.Level.

I think what the reflection code could do is to check if a struct field implements TextUnmarshaler and simply run UnmarshalText on a field with a string value from BCL.

[condrove10]

By the way, if you excuse me being picky about the vocabulary, I believe you meant to "skip the field", not a block. Skipping the block would make sense for a nested block, but then the field being skipped should be a nested struct as well. I believe this is not what you meant now :) More on this will be in another thread - btw, I am preparing a sketch of a new syntax to cover all common bind cases, we'll discuss it separately (edit: that's the issue #6).

Hey, you're totally right. I meant skipping the field. Also, allow me to thank you for all the fixes and features you've added this week. I will definitely update the package version and come back with some feedback in the coming week/s.

[wkhere]

Yes please try the new code.

Btw, I also recommend reading proposals in #6

Enjoy!