Message ID | 1443760312-656-4-git-send-email-eblake@redhat.com |
---|---|
State | New |
Headers | show |
Woot! Eric Blake <eblake@redhat.com> writes: > Commit ac88219a had several TODO markers about whether we needed > to automatically create the corresponding array type alongside > any other type. It turns out that most of the time, we don't! > > As part of lazy creation of array types, this patch now assigns > an 'info' to array types at their point of first instantiation, > rather than leaving it None. I guess our general for info should be: For explicitly defined entities, info points to the (explicit) definition. For implicitly defined entities, it points to a place that triggers implicit definition. For some kinds of entities, multiple places may exist, and info points to one of them. First one we encounter is as good as any, and easy to compute. > There are a few exceptions: 1) We have a few situations where we > use an array type in internal code but do not expose that type > through QMP; fix it by declaring a dummy type that forces the > generator to see that we want to use the array type. > > 2) The builtin arrays (such as intList for QAPI ['int']) must > always be generated, because of the way our QAPI_TYPES_BUILTIN > compile guard works: we have situations (at the very least > tests/test-qmp-output-visitor.c) that include both top-level > "qapi-types.h" (via "error.h") and a secondary > "test-qapi-types.h". If we were to only emit the builtin types > when used locally, then the first .h file would not include all > types, but the second .h does not declare anything at all because > the first .h set QAPI_TYPES_BUILTIN, and we would end up with > compilation error due to things like unknown type 'int8List'. Yes. > Actually, we may need to revisit how we do type guards, and > change from a single QAPI_TYPES_BUILTIN over to a different > usage pattern that does one #ifdef per qapi type - right now, > the only types that are declared multiple times between two qapi > .json files for inclusion by a single .c file happen to be the > builtin arrays. But now that we have QAPI 'include' statements, > it is logical to assume that we will soon reach a point where > we want to reuse non-builtin types (yes, I'm thinking about what > it will take to add introspection to QGA, where we will want to > reuse the SchemaInfo type and friends). One #ifdef per type > will help ensure that generating the same qapi type into more > than one qapi-types.h won't cause collisions when both are > included in the same .c file; but we also have to solve how to > avoid creating duplicate qapi-types.c entry points. So that > is a problem left for another day. Yes, let's leave it for later. > Interestingly, the introspection output is unchanged - this is > because we already cull all types that are not indirectly > reachable from a command or event, so introspection was already > using only a subset of array types. The subset of types > introspected is now a much larger percentage of the overall set > of array types emitted in qapi-types.h (since the larger set > shrunk), but still not 100% (proof that the array types emitted > for our new Dummy struct, and the new Dummy struct itself, don't > affect QMP). Of course, the rest of the generated output is > drastically shrunk (a number of unused array types are no longer > generated) [for best results, diff the generated files with > 'git diff --patience --no-index pre post']. Here, you're talking about the effect on generated code. I'd put the specific case of introspection last, in a paragraph of its own. How much is "drastically"? Here's what I get: $ git-diff --patience --no-index --stat o n {o => n}/qapi-types.c | 2239 +----------------------------- {o => n}/qapi-types.h | 1800 +----------------------- {o => n}/qapi-visit.c | 3605 +------------------------------------------------ {o => n}/qapi-visit.h | 151 +-- 4 files changed, 23 insertions(+), 7772 deletions(-) $ wc o/* | tail -n 1 34738 90018 985738 total $ wc n/* | tail -n 1 26989 71981 788721 total That's a >20% shrink. $ grep -c 'typedef .*List;' {o,n}/qapi-types.h o/qapi-types.h:219 n/qapi-types.h:69 Less than a third of the array types are still around. Includes 14 built-in array types. Compiled size with -O2 for me, before: text data bss dec hex filename 39235 4576 0 43811 ab23 qapi-types.o 159424 0 0 159424 26ec0 qapi-visit.o After: text data bss dec hex filename 24931 4576 0 29507 7343 qapi-types.o 115581 0 0 115581 1c37d qapi-visit.o A 30% shrink. Lovely. Further reductions are probably possible. One step at a time. Generating code makes profligate wastefulness so much easier. > Signed-off-by: Eric Blake <eblake@redhat.com> > > --- > v6: new patch > --- > qapi-schema.json | 10 +++++++++ > scripts/qapi.py | 39 ++++++++++++++++----------------- > tests/qapi-schema/qapi-schema-test.json | 4 ++++ > tests/qapi-schema/qapi-schema-test.out | 3 +++ > 4 files changed, 36 insertions(+), 20 deletions(-) > > diff --git a/qapi-schema.json b/qapi-schema.json > index 8b0520c..98bf0b2 100644 > --- a/qapi-schema.json > +++ b/qapi-schema.json > @@ -3396,6 +3396,16 @@ > 'features': 'int' } } > > ## > +# @Dummy > +# > +# Not used by QMP; hack to let us use X86CPUFeatureWordInfoList internally > +# > +# Since 2.5 > +## > +{ 'struct': 'Dummy', 'data': { 'unused': ['X86CPUFeatureWordInfo'] } } DummyToForceArrays? > + > + > +## > # @RxState: > # > # Packets receiving state > diff --git a/scripts/qapi.py b/scripts/qapi.py > index 8123ab3..15640b6 100644 > --- a/scripts/qapi.py > +++ b/scripts/qapi.py > @@ -1143,7 +1143,7 @@ class QAPISchema(object): > def _def_builtin_type(self, name, json_type, c_type, c_null): > self._def_entity(QAPISchemaBuiltinType(name, json_type, > c_type, c_null)) > - self._make_array_type(name) # TODO really needed? > + self._make_array_type(name, None) Should we keep a TODO here? > > def _def_predefineds(self): > for t in [('str', 'string', 'char' + pointer_suffix, 'NULL'), > @@ -1170,10 +1170,10 @@ class QAPISchema(object): > self._def_entity(QAPISchemaEnumType(name, None, values, None)) > return name > > - def _make_array_type(self, element_type): > + def _make_array_type(self, element_type, info): > name = element_type + 'List' > if not self.lookup_type(name): > - self._def_entity(QAPISchemaArrayType(name, None, element_type)) > + self._def_entity(QAPISchemaArrayType(name, info, element_type)) > return name > > def _make_implicit_object_type(self, name, role, members): > @@ -1190,20 +1190,19 @@ class QAPISchema(object): > data = expr['data'] > prefix = expr.get('prefix') > self._def_entity(QAPISchemaEnumType(name, info, data, prefix)) > - self._make_array_type(name) # TODO really needed? > > - def _make_member(self, name, typ): > + def _make_member(self, name, typ, info): > optional = False > if name.startswith('*'): > name = name[1:] > optional = True > if isinstance(typ, list): > assert len(typ) == 1 > - typ = self._make_array_type(typ[0]) > + typ = self._make_array_type(typ[0], info) > return QAPISchemaObjectTypeMember(name, typ, optional) > > - def _make_members(self, data): > - return [self._make_member(key, value) > + def _make_members(self, data, info): > + return [self._make_member(key, value, info) > for (key, value) in data.iteritems()] > > def _def_struct_type(self, expr, info): > @@ -1211,19 +1210,19 @@ class QAPISchema(object): > base = expr.get('base') > data = expr['data'] > self._def_entity(QAPISchemaObjectType(name, info, base, > - self._make_members(data), > + self._make_members(data, info), > None)) > - self._make_array_type(name) # TODO really needed? > > def _make_variant(self, case, typ): > return QAPISchemaObjectTypeVariant(case, typ) > > - def _make_simple_variant(self, case, typ): > + def _make_simple_variant(self, case, typ, info): > if isinstance(typ, list): > assert len(typ) == 1 > - typ = self._make_array_type(typ[0]) > + typ = self._make_array_type(typ[0], info) > typ = self._make_implicit_object_type(typ, 'wrapper', > - [self._make_member('data', typ)]) > + [self._make_member('data', typ, > + info)]) Consider a hanging indent here: typ = self._make_implicit_object_type( typ, 'wrapper', [self._make_member('data', typ, info)]) > return QAPISchemaObjectTypeVariant(case, typ) > > def _make_tag_enum(self, type_name, variants): > @@ -1240,16 +1239,15 @@ class QAPISchema(object): > variants = [self._make_variant(key, value) > for (key, value) in data.iteritems()] > else: > - variants = [self._make_simple_variant(key, value) > + variants = [self._make_simple_variant(key, value, info) > for (key, value) in data.iteritems()] > tag_enum = self._make_tag_enum(name, variants) > self._def_entity( > QAPISchemaObjectType(name, info, base, > - self._make_members(OrderedDict()), > + self._make_members(OrderedDict(), info), > QAPISchemaObjectTypeVariants(tag_name, > tag_enum, > variants))) > - self._make_array_type(name) # TODO really needed? > > def _def_alternate_type(self, expr, info): > name = expr['alternate'] > @@ -1262,7 +1260,6 @@ class QAPISchema(object): > QAPISchemaObjectTypeVariants(None, > tag_enum, > variants))) > - self._make_array_type(name) # TODO really needed? > > def _def_command(self, expr, info): > name = expr['command'] > @@ -1272,10 +1269,11 @@ class QAPISchema(object): > success_response = expr.get('success-response', True) > if isinstance(data, OrderedDict): > data = self._make_implicit_object_type(name, 'arg', > - self._make_members(data)) > + self._make_members(data, > + info)) Hanging indent? > if isinstance(rets, list): > assert len(rets) == 1 > - rets = self._make_array_type(rets[0]) > + rets = self._make_array_type(rets[0], info) > self._def_entity(QAPISchemaCommand(name, info, data, rets, gen, > success_response)) > > @@ -1284,7 +1282,8 @@ class QAPISchema(object): > data = expr.get('data') > if isinstance(data, OrderedDict): > data = self._make_implicit_object_type(name, 'arg', > - self._make_members(data)) > + self._make_members(data, > + info)) Hanging indent? > self._def_entity(QAPISchemaEvent(name, info, data)) > > def _def_exprs(self): > diff --git a/tests/qapi-schema/qapi-schema-test.json b/tests/qapi-schema/qapi-schema-test.json > index abe59fd..020ff2e 100644 > --- a/tests/qapi-schema/qapi-schema-test.json > +++ b/tests/qapi-schema/qapi-schema-test.json > @@ -31,6 +31,10 @@ > 'data': { 'string0': 'str', > 'dict1': 'UserDefTwoDict' } } > > +# dummy struct to force generation of array types not otherwise mentioned > +{ 'struct': 'ForceArrays', > + 'data': { 'unused1':['UserDefOne'], 'unused2':['UserDefTwo'] } } > + > # for testing unions > # Among other things, test that a name collision between branches does > # not cause any problems (since only one branch can be in use at a time), > diff --git a/tests/qapi-schema/qapi-schema-test.out b/tests/qapi-schema/qapi-schema-test.out > index 8f81784..2a8c82e 100644 > --- a/tests/qapi-schema/qapi-schema-test.out > +++ b/tests/qapi-schema/qapi-schema-test.out > @@ -86,6 +86,9 @@ object EventStructOne > member struct1: UserDefOne optional=False > member string: str optional=False > member enum2: EnumOne optional=True > +object ForceArrays > + member unused1: UserDefOneList optional=False > + member unused2: UserDefTwoList optional=False > object NestedEnumsOne > member enum1: EnumOne optional=False > member enum2: EnumOne optional=True
On 10/02/2015 02:06 AM, Markus Armbruster wrote: > Woot! > > Eric Blake <eblake@redhat.com> writes: > >> Commit ac88219a had several TODO markers about whether we needed >> to automatically create the corresponding array type alongside >> any other type. It turns out that most of the time, we don't! >> >> As part of lazy creation of array types, this patch now assigns >> an 'info' to array types at their point of first instantiation, >> rather than leaving it None. > > I guess our general for info should be: s/general/general description/ ? > > For explicitly defined entities, info points to the (explicit) > definition. For implicitly defined entities, it points to a place that > triggers implicit definition. For some kinds of entities, multiple > places may exist, and info points to one of them. > Right now, we don't document any of the fields, but I can certainly add that comment. > > Here, you're talking about the effect on generated code. I'd put the > specific case of introspection last, in a paragraph of its own. > Improving the commit message is easier than redoing bad code :) >> +++ b/qapi-schema.json >> @@ -3396,6 +3396,16 @@ >> 'features': 'int' } } >> >> ## >> +# @Dummy >> +# >> +# Not used by QMP; hack to let us use X86CPUFeatureWordInfoList internally >> +# >> +# Since 2.5 >> +## >> +{ 'struct': 'Dummy', 'data': { 'unused': ['X86CPUFeatureWordInfo'] } } > > DummyToForceArrays? Sure. > >> + >> + >> +## >> # @RxState: >> # >> # Packets receiving state >> diff --git a/scripts/qapi.py b/scripts/qapi.py >> index 8123ab3..15640b6 100644 >> --- a/scripts/qapi.py >> +++ b/scripts/qapi.py >> @@ -1143,7 +1143,7 @@ class QAPISchema(object): >> def _def_builtin_type(self, name, json_type, c_type, c_null): >> self._def_entity(QAPISchemaBuiltinType(name, json_type, >> c_type, c_null)) >> - self._make_array_type(name) # TODO really needed? >> + self._make_array_type(name, None) > > Should we keep a TODO here? Sure, with better text explaining the guard issue on types shared in multiple qapi-types.h files. >> typ = self._make_implicit_object_type(typ, 'wrapper', >> - [self._make_member('data', typ)]) >> + [self._make_member('data', typ, >> + info)]) > > Consider a hanging indent here: > > typ = self._make_implicit_object_type( > typ, 'wrapper', [self._make_member('data', typ, info)]) Okay. pep8 and pylint like it (I'm not used to doing it, but it does look better, and emacs didn't fight me too hard).
Eric Blake <eblake@redhat.com> writes: > On 10/02/2015 02:06 AM, Markus Armbruster wrote: >> Woot! >> >> Eric Blake <eblake@redhat.com> writes: >> >>> Commit ac88219a had several TODO markers about whether we needed >>> to automatically create the corresponding array type alongside >>> any other type. It turns out that most of the time, we don't! >>> >>> As part of lazy creation of array types, this patch now assigns >>> an 'info' to array types at their point of first instantiation, >>> rather than leaving it None. >> >> I guess our general for info should be: > > s/general/general description/ ? General idea. Looks like my typing engine doesn't run on all cylinders on Fridays... >> For explicitly defined entities, info points to the (explicit) >> definition. For implicitly defined entities, it points to a place that >> triggers implicit definition. For some kinds of entities, multiple >> places may exist, and info points to one of them. >> > > Right now, we don't document any of the fields, but I can certainly add > that comment. > > >> >> Here, you're talking about the effect on generated code. I'd put the >> specific case of introspection last, in a paragraph of its own. >> > > Improving the commit message is easier than redoing bad code :) > >>> +++ b/qapi-schema.json >>> @@ -3396,6 +3396,16 @@ >>> 'features': 'int' } } >>> >>> ## >>> +# @Dummy >>> +# >>> +# Not used by QMP; hack to let us use X86CPUFeatureWordInfoList internally >>> +# >>> +# Since 2.5 >>> +## >>> +{ 'struct': 'Dummy', 'data': { 'unused': ['X86CPUFeatureWordInfo'] } } >> >> DummyToForceArrays? > > Sure. > >> >>> + >>> + >>> +## >>> # @RxState: >>> # >>> # Packets receiving state >>> diff --git a/scripts/qapi.py b/scripts/qapi.py >>> index 8123ab3..15640b6 100644 >>> --- a/scripts/qapi.py >>> +++ b/scripts/qapi.py >>> @@ -1143,7 +1143,7 @@ class QAPISchema(object): >>> def _def_builtin_type(self, name, json_type, c_type, c_null): >>> self._def_entity(QAPISchemaBuiltinType(name, json_type, >>> c_type, c_null)) >>> - self._make_array_type(name) # TODO really needed? >>> + self._make_array_type(name, None) >> >> Should we keep a TODO here? > > Sure, with better text explaining the guard issue on types shared in > multiple qapi-types.h files. > > >>> typ = self._make_implicit_object_type(typ, 'wrapper', >>> - [self._make_member('data', typ)]) >>> + [self._make_member('data', typ, >>> + info)]) >> >> Consider a hanging indent here: >> >> typ = self._make_implicit_object_type( >> typ, 'wrapper', [self._make_member('data', typ, info)]) > > > Okay. pep8 and pylint like it (I'm not used to doing it, but it does > look better, and emacs didn't fight me too hard). I'm still getting used to hanging indents myself. When in Rome...
diff --git a/qapi-schema.json b/qapi-schema.json index 8b0520c..98bf0b2 100644 --- a/qapi-schema.json +++ b/qapi-schema.json @@ -3396,6 +3396,16 @@ 'features': 'int' } } ## +# @Dummy +# +# Not used by QMP; hack to let us use X86CPUFeatureWordInfoList internally +# +# Since 2.5 +## +{ 'struct': 'Dummy', 'data': { 'unused': ['X86CPUFeatureWordInfo'] } } + + +## # @RxState: # # Packets receiving state diff --git a/scripts/qapi.py b/scripts/qapi.py index 8123ab3..15640b6 100644 --- a/scripts/qapi.py +++ b/scripts/qapi.py @@ -1143,7 +1143,7 @@ class QAPISchema(object): def _def_builtin_type(self, name, json_type, c_type, c_null): self._def_entity(QAPISchemaBuiltinType(name, json_type, c_type, c_null)) - self._make_array_type(name) # TODO really needed? + self._make_array_type(name, None) def _def_predefineds(self): for t in [('str', 'string', 'char' + pointer_suffix, 'NULL'), @@ -1170,10 +1170,10 @@ class QAPISchema(object): self._def_entity(QAPISchemaEnumType(name, None, values, None)) return name - def _make_array_type(self, element_type): + def _make_array_type(self, element_type, info): name = element_type + 'List' if not self.lookup_type(name): - self._def_entity(QAPISchemaArrayType(name, None, element_type)) + self._def_entity(QAPISchemaArrayType(name, info, element_type)) return name def _make_implicit_object_type(self, name, role, members): @@ -1190,20 +1190,19 @@ class QAPISchema(object): data = expr['data'] prefix = expr.get('prefix') self._def_entity(QAPISchemaEnumType(name, info, data, prefix)) - self._make_array_type(name) # TODO really needed? - def _make_member(self, name, typ): + def _make_member(self, name, typ, info): optional = False if name.startswith('*'): name = name[1:] optional = True if isinstance(typ, list): assert len(typ) == 1 - typ = self._make_array_type(typ[0]) + typ = self._make_array_type(typ[0], info) return QAPISchemaObjectTypeMember(name, typ, optional) - def _make_members(self, data): - return [self._make_member(key, value) + def _make_members(self, data, info): + return [self._make_member(key, value, info) for (key, value) in data.iteritems()] def _def_struct_type(self, expr, info): @@ -1211,19 +1210,19 @@ class QAPISchema(object): base = expr.get('base') data = expr['data'] self._def_entity(QAPISchemaObjectType(name, info, base, - self._make_members(data), + self._make_members(data, info), None)) - self._make_array_type(name) # TODO really needed? def _make_variant(self, case, typ): return QAPISchemaObjectTypeVariant(case, typ) - def _make_simple_variant(self, case, typ): + def _make_simple_variant(self, case, typ, info): if isinstance(typ, list): assert len(typ) == 1 - typ = self._make_array_type(typ[0]) + typ = self._make_array_type(typ[0], info) typ = self._make_implicit_object_type(typ, 'wrapper', - [self._make_member('data', typ)]) + [self._make_member('data', typ, + info)]) return QAPISchemaObjectTypeVariant(case, typ) def _make_tag_enum(self, type_name, variants): @@ -1240,16 +1239,15 @@ class QAPISchema(object): variants = [self._make_variant(key, value) for (key, value) in data.iteritems()] else: - variants = [self._make_simple_variant(key, value) + variants = [self._make_simple_variant(key, value, info) for (key, value) in data.iteritems()] tag_enum = self._make_tag_enum(name, variants) self._def_entity( QAPISchemaObjectType(name, info, base, - self._make_members(OrderedDict()), + self._make_members(OrderedDict(), info), QAPISchemaObjectTypeVariants(tag_name, tag_enum, variants))) - self._make_array_type(name) # TODO really needed? def _def_alternate_type(self, expr, info): name = expr['alternate'] @@ -1262,7 +1260,6 @@ class QAPISchema(object): QAPISchemaObjectTypeVariants(None, tag_enum, variants))) - self._make_array_type(name) # TODO really needed? def _def_command(self, expr, info): name = expr['command'] @@ -1272,10 +1269,11 @@ class QAPISchema(object): success_response = expr.get('success-response', True) if isinstance(data, OrderedDict): data = self._make_implicit_object_type(name, 'arg', - self._make_members(data)) + self._make_members(data, + info)) if isinstance(rets, list): assert len(rets) == 1 - rets = self._make_array_type(rets[0]) + rets = self._make_array_type(rets[0], info) self._def_entity(QAPISchemaCommand(name, info, data, rets, gen, success_response)) @@ -1284,7 +1282,8 @@ class QAPISchema(object): data = expr.get('data') if isinstance(data, OrderedDict): data = self._make_implicit_object_type(name, 'arg', - self._make_members(data)) + self._make_members(data, + info)) self._def_entity(QAPISchemaEvent(name, info, data)) def _def_exprs(self): diff --git a/tests/qapi-schema/qapi-schema-test.json b/tests/qapi-schema/qapi-schema-test.json index abe59fd..020ff2e 100644 --- a/tests/qapi-schema/qapi-schema-test.json +++ b/tests/qapi-schema/qapi-schema-test.json @@ -31,6 +31,10 @@ 'data': { 'string0': 'str', 'dict1': 'UserDefTwoDict' } } +# dummy struct to force generation of array types not otherwise mentioned +{ 'struct': 'ForceArrays', + 'data': { 'unused1':['UserDefOne'], 'unused2':['UserDefTwo'] } } + # for testing unions # Among other things, test that a name collision between branches does # not cause any problems (since only one branch can be in use at a time), diff --git a/tests/qapi-schema/qapi-schema-test.out b/tests/qapi-schema/qapi-schema-test.out index 8f81784..2a8c82e 100644 --- a/tests/qapi-schema/qapi-schema-test.out +++ b/tests/qapi-schema/qapi-schema-test.out @@ -86,6 +86,9 @@ object EventStructOne member struct1: UserDefOne optional=False member string: str optional=False member enum2: EnumOne optional=True +object ForceArrays + member unused1: UserDefOneList optional=False + member unused2: UserDefTwoList optional=False object NestedEnumsOne member enum1: EnumOne optional=False member enum2: EnumOne optional=True
Commit ac88219a had several TODO markers about whether we needed to automatically create the corresponding array type alongside any other type. It turns out that most of the time, we don't! As part of lazy creation of array types, this patch now assigns an 'info' to array types at their point of first instantiation, rather than leaving it None. There are a few exceptions: 1) We have a few situations where we use an array type in internal code but do not expose that type through QMP; fix it by declaring a dummy type that forces the generator to see that we want to use the array type. 2) The builtin arrays (such as intList for QAPI ['int']) must always be generated, because of the way our QAPI_TYPES_BUILTIN compile guard works: we have situations (at the very least tests/test-qmp-output-visitor.c) that include both top-level "qapi-types.h" (via "error.h") and a secondary "test-qapi-types.h". If we were to only emit the builtin types when used locally, then the first .h file would not include all types, but the second .h does not declare anything at all because the first .h set QAPI_TYPES_BUILTIN, and we would end up with compilation error due to things like unknown type 'int8List'. Actually, we may need to revisit how we do type guards, and change from a single QAPI_TYPES_BUILTIN over to a different usage pattern that does one #ifdef per qapi type - right now, the only types that are declared multiple times between two qapi .json files for inclusion by a single .c file happen to be the builtin arrays. But now that we have QAPI 'include' statements, it is logical to assume that we will soon reach a point where we want to reuse non-builtin types (yes, I'm thinking about what it will take to add introspection to QGA, where we will want to reuse the SchemaInfo type and friends). One #ifdef per type will help ensure that generating the same qapi type into more than one qapi-types.h won't cause collisions when both are included in the same .c file; but we also have to solve how to avoid creating duplicate qapi-types.c entry points. So that is a problem left for another day. Interestingly, the introspection output is unchanged - this is because we already cull all types that are not indirectly reachable from a command or event, so introspection was already using only a subset of array types. The subset of types introspected is now a much larger percentage of the overall set of array types emitted in qapi-types.h (since the larger set shrunk), but still not 100% (proof that the array types emitted for our new Dummy struct, and the new Dummy struct itself, don't affect QMP). Of course, the rest of the generated output is drastically shrunk (a number of unused array types are no longer generated) [for best results, diff the generated files with 'git diff --patience --no-index pre post']. Signed-off-by: Eric Blake <eblake@redhat.com> --- v6: new patch --- qapi-schema.json | 10 +++++++++ scripts/qapi.py | 39 ++++++++++++++++----------------- tests/qapi-schema/qapi-schema-test.json | 4 ++++ tests/qapi-schema/qapi-schema-test.out | 3 +++ 4 files changed, 36 insertions(+), 20 deletions(-)