Setting "option allow_alias = true" causes message compilation to fail about gpb HOT 8 CLOSED

I see. I think options at the time of writing that sentence about being ignored, were maybe only allowed after an enumerator, in any case I didn't foresee it at that position.

Ignoring the option at encoding would be rather easy, I think. On decoding, I suppose a reasonable approach could be to decode to the first enumerator symbol, ENUM1 in this case.

I'll take a look at it, hopefully during the weekend.

from gpb.

Yeah, I agree with your points: ignoring the option at encoding, and decoding to the first enumerator symbol, sounds like very reasonable behavior. Thanks for the library - it's really handy!

from gpb.

I have pushed a solution, to the branch enum-option-allow_alias. My intention is to try to also include the other issue with maps and optional fields, soon. But if that should get delayed, I will merge this to master, so that at least this won't get held up. (My plan is of course that 'soon' will happens soon, but there's always the unexpected, and meanwhile there's a solution on a branch in case you want to test it out.)

from gpb.

Thanks very much! :) I did a quick test of the new enum-option-allow_alias branch, and that certainly looks to solve the issue I had.

from gpb.

Good to hear, thanks for quick feedback!

from gpb.

No problem, thanks for the quick fix. :)

from gpb.

This is now merged to master and in 3.17.8.

from gpb.

Thanks!

from gpb.