John Snow <js...@redhat.com> writes:

> Signed-off-by: John Snow <js...@redhat.com>
>
> ---
>
> The event_enum_members change might become irrelevant after a
> forthcoming (?) patch by Markus, but didn't have it in-hand at time of
> publishing.

It's in my "[PATCH 00/11] Drop support for QAPIGen without a file name",
which includes parts of your v1.  The parts that are new should be
injected into your series so they replace your "[PATCH v2 09/12]
qapi/gen: move write method to QAPIGenC, make fname a str".  Holler if
you need help.

> Signed-off-by: John Snow <js...@redhat.com>
> ---
>  scripts/qapi/events.py |  2 +-
>  scripts/qapi/schema.py | 25 ++++++++++++++-----------
>  scripts/qapi/types.py  |  9 +++++----
>  scripts/qapi/visit.py  |  6 +++---
>  4 files changed, 23 insertions(+), 19 deletions(-)
>
> diff --git a/scripts/qapi/events.py b/scripts/qapi/events.py
> index 9851653b9d1..9ba4f109028 100644
> --- a/scripts/qapi/events.py
> +++ b/scripts/qapi/events.py
> @@ -225,7 +225,7 @@ def visit_event(self,
>                                            self._event_emit_name))
>          # Note: we generate the enum member regardless of @ifcond, to
>          # keep the enumeration usable in target-independent code.
> -        self._event_enum_members.append(QAPISchemaEnumMember(name, None))
> +        self._event_enum_members.append(QAPISchemaEnumMember(name, info))

This "enum" is not supposed to be erroneous.  If it is, it's a bug.

Your patch changes how the code behaves should such a bug bite here.
Before, we crash.  Afterwards, we report the bug using @info, which I'd
expect to produce utterly confusing error messages.

My comments on PATCH 06 apply: how the code should behave here is a
design issue that should be kept out of this patch series.

If you need to pass a value other than None to help with static typing,
then pass a suitable poison info that will crash right where None
crashes now.

>  def gen_events(schema: QAPISchema,
> diff --git a/scripts/qapi/schema.py b/scripts/qapi/schema.py
> index 720449feee4..0449771dfe5 100644
> --- a/scripts/qapi/schema.py
> +++ b/scripts/qapi/schema.py
> @@ -23,6 +23,7 @@
>  from .error import QAPIError, QAPISemError
>  from .expr import check_exprs
>  from .parser import QAPISchemaParser
> +from .source import QAPISourceInfo
>  
>  
>  class QAPISchemaEntity:
> @@ -36,10 +37,10 @@ def __init__(self, name, info, doc, ifcond=None, 
> features=None):
>          self.name = name
>          self._module = None
>          # For explicitly defined entities, info points to the (explicit)
> -        # definition.  For builtins (and their arrays), info is None.
> -        # For implicitly defined entities, info points to a place that
> -        # triggered the implicit definition (there may be more than one
> -        # such place).
> +        # definition. For built-in types (and their arrays), info is a
> +        # special object that evaluates to False. For implicitly defined
> +        # entities, info points to a place that triggered the implicit
> +        # definition (there may be more than one such place).
>          self.info = info
>          self.doc = doc
>          self._ifcond = ifcond or []
> @@ -68,7 +69,7 @@ def check_doc(self):
>  
>      def _set_module(self, schema, info):
>          assert self._checked
> -        self._module = schema.module_by_fname(info and info.fname)
> +        self._module = schema.module_by_fname(info.fname if info else None)

Looks unrelated.

>          self._module.add_entity(self)
>  
>      def set_module(self, schema):
[...]


Reply via email to