Go: introduce special GoOutputWriter and move generation of error check to it #299
Conversation
| @@ -15,7 +15,30 @@ sealed trait TranslatorResult | |||
| case class ResultString(s: String) extends TranslatorResult | |||
| case class ResultLocalVar(n: Int) extends TranslatorResult | |||
|
|
|||
| class GoTranslator(out: StringLanguageOutputWriter, provider: TypeProvider, importList: ImportList) | |||
| class GoOutputWriter(indentStr: String) extends StringLanguageOutputWriter(indentStr) { | |||
There was a problem hiding this comment.
I'm very unsure if this is a way to go. Custom OutputWriters are supposed to be used to encompass very special logic: for example, something with very custom formatting (rather than traditional "keep indentation and go line-by-line").
Here, you use it to implement what can be implemented in any Compiler or Translator — i.e. pretty much vanilla output by calling puts(), inc(), dec(), etc.
If for some reason keeping it in GoTranslator does not work — let's discuss why and if we should move it into, for example, GoCompiler?
There was a problem hiding this comment.
The GoOutputWriter just add the customization point to intercept attempts to write the error check. Currently KS Translator uses custom writer and heuristics to find and replace the returned error with t.Fatal(err) call.
This PR allows do the same but in a more elegant way, and also allow to completely skip the error check which is needed for the translator when the error is an expected result.
Probably it is possible to keep this logic in GoTranslator and make a subclass of the translator in KS Translator to use it here. If that's possible, I'm not sure for what reason that was not done before.
There was a problem hiding this comment.
I agree with @GreyCat that extending StringLanguageOutputWriter makes little sense in this case from the semantic point of view. I don't see why extending GoTranslator in the KST translator wouldn't work just as well.
There was a problem hiding this comment.
Ok, I rewrited kaitai-io/kaitai_struct_tests#123 so this PR is no longer needed, unless you want to accept a couple of lines of documentation
…ue of type It will be used in KST to check that we return default values when error is expected
Mingun
force-pushed
the
go-kst-fixes
branch
from
db49bf5 to
032007a
Compare
This will allow to override this method and do not generate check in KST when performing checks for returned errors using
asserts[i].exceptionkey.This should be merged with this PR: kaitai-io/kaitai_struct_tests#123