Allow initializing a chain of protos using only a hash in Ruby #3627
Conversation
|
Thanks for your pull request. The automated tests will run as soon as one of the admins verifies this change is ok for us to run on our infrastructure. |
1 similar comment
|
Thanks for your pull request. The automated tests will run as soon as one of the admins verifies this change is ok for us to run on our infrastructure. |
|
Can one of the admins verify this patch? |
zanker
force-pushed
the
zanker/add-submsg-hash-init
branch
7 times, most recently
from
4f0d23c
to
399ab27
Compare
|
ok to test |
|
Looks like this change failed Travis? |
zanker
force-pushed
the
zanker/add-submsg-hash-init
branch
from
399ab27
to
8771483
Compare
|
@haberman Fixed the test that failed. It was one of the compatibility tests, which now fails with Is the fact that the exception changed ok? |
ruby/ext/google/protobuf_c/message.c
Outdated
| @@ -248,16 +260,26 @@ int Message_initialize_kwarg(VALUE key, VALUE val, VALUE _self) { | |||
| Map_merge_into_self(map, val); | |||
| } else if (upb_fielddef_label(f) == UPB_LABEL_REPEATED) { | |||
| VALUE ary; | |||
| VALUE entry; | |||
There was a problem hiding this comment.
This is only used inside the for loop, so please declare it there.
I saw #3120 also asked for this. We're in the process of trying to migrate people from a pure Ruby proto gem to the Google one, and the requirement to initialize all protos is a blocker. Sounds like there was interest in this, so figured I'd take a stab at implementing it in.
Taking the examples from #3120, this lets you do:
MyMessage.new(submessage_field: { foo: 42 })and it will automatically turn the{ foo: 42 }into aSubMessage.new(foo: 42).Since a common scenario is also to be turning JSON into protos (for proto backed JSON APIs), I also changed the initialization to support symbol or string keys to prevent having to do a deep symbolization pass before initializing the protos.