Markus Armbruster <arm...@redhat.com> writes:
> Kevin Wolf <kw...@redhat.com> writes:
>
>> Am 16.02.2021 um 16:14 hat Markus Armbruster geschrieben:
>>> Kevin Wolf <kw...@redhat.com> writes:
>>>
>>> > Signed-off-by: Kevin Wolf <kw...@redhat.com>
>>> [...]
>>> > diff --git a/tests/qapi-schema/alias-name-bad-type.err
>>> > b/tests/qapi-schema/alias-name-bad-type.err
>>> > new file mode 100644
>>> > index 0000000000..489f45ff9b
>>> > --- /dev/null
>>> > +++ b/tests/qapi-schema/alias-name-bad-type.err
>>> > @@ -0,0 +1,2 @@
>>> > +alias-name-bad-type.json: In struct 'AliasStruct0':
>>> > +alias-name-bad-type.json:1: alias member 'name' requires a string name
>>>
>>> Would "'aliases' member 'name'..." be more consistent?
>>
>> 'aliases' is a list, not a single alias definition, so technically it
>> would have to be "'aliases' member member 'name'...", which I feel is a
>> bit too confusing.
>
> Indeed.
>
> I think glossing over the list is excusable.
>
>> I think I have consistently used "alias" for "'aliases' member"
>> everywhere, though. At least, that was the intention.
>
> A different way of glossing over details. Should do as well. I'll
> double-check consistency.
I did, and it looks okay:
$ grep "'alias" *err
alias-bad-type.err:alias-bad-type.json:1: 'aliases' members must be objects
Okay; we are talking about members of array 'aliases' here.
alias-missing-source.err:alias-missing-source.json:1: 'aliases' member
misses key 'source'
Likewise.
alias-unknown-key.err:alias-unknown-key.json:1: 'aliases' member has
unknown key 'known'
Likewise.
aliases-bad-type.err:aliases-bad-type.json:1: 'aliases' must be an array
Okay; we are talking about 'aliases'.
double-type.err:Valid keys are 'aliases', 'base', 'data', 'features', 'if',
'struct'.
unknown-expr-key.err:Valid keys are 'aliases', 'base', 'data', 'features',
'if', 'struct'.
Okay.
