On 08/10/2017 09:06 AM, Pradeep Jagadeesh wrote:

>>> It's not "moving it back", it's keeping it where it is. But I see no big
>>> problem with moving it to a common file either.
>> I'd rather not put every struct shared across subsystem boundaries in
>> its own file.
>> We can keep it right where it is for now.  Bonus: more readable diff.
>> If we start sharing more throttle-related material than just a struct,
>> we can reconsider.
>> We could also move it to the existing file for common stuff:
>> qapi/common.json.  Not a great fit, though.
> So, the final conclusion is to move to common.json?


If more than one .json file would benefit by including the definition,
then put it in a separate file that both .json include from.

But if only one .json file would be including a new file, then just
inline the struct directly into that one original file (in this case,
block-core.json) instead of creating a separate file (so no to needing
iothrottle.json), or putting the code in yet a different file than the
one that is using the struct (so no to putting it in common.json).

Eric Blake, Principal Software Engineer
Red Hat, Inc.           +1-919-301-3266
Virtualization:  qemu.org | libvirt.org

Attachment: signature.asc
Description: OpenPGP digital signature

Reply via email to