Re: [PATCH 1/9] qapi: New special feature flag "unstable"

[Markus Armbruster <armbru@redhat.com>]

Kashyap Chamarthy <kchamart@redhat.com> writes:

> On Mon, Oct 25, 2021 at 07:25:24AM +0200, Markus Armbruster wrote:
>> By convention, names starting with "x-" are experimental.  The parts
>> of external interfaces so named may be withdrawn or changed
>> incompatibly in future releases.
>> 
>> Drawback: promoting something from experimental to stable involves a
>> name change.  Client code needs to be updated.
>> 
>> Moreover, the convention is not universally observed:
>> 
>> * QOM type "input-barrier" has properties "x-origin", "y-origin".
>>   Looks accidental, but it's ABI since 4.2.
>> 
>> * QOM types "memory-backend-file", "memory-backend-memfd",
>>   "memory-backend-ram", and "memory-backend-epc" have a property
>>   "x-use-canonical-path-for-ramblock-id" that is documented to be
>>   stable despite its name.
>
> Looks like there's another stable property with an "x-" prefix:
> "x-remote-object", part of QOM type @RemoteObjectProperties.

The union branch 'x-remote-object' isn't flagged 'unstable' (because
union branches can't have feature flags), but the enumeration value
'x-remote-object' is.  Sufficient, because you can't use the branch
without using the enumeration value.  Admittedly subtle.

I wrote a bit of code (appended) to make sure I don't miss names.

> Given the above "x-" properties are now stable, I take it that they
> cannot be renamed now, as they might break any tools using them?  My
> guess is the tedious way is not worth it: deprecate them, and add the
> non-x variants as "synonyms".  

"x-use-canonical-path-for-ramblock-id" goes back to commit fa0cb34d22
"hostmem: use object id for memory region name with >= 4.0" (v4.0).  It
may have been intended to be internal back then.  It wasn't anymore when
commit 8db0b20415 "machine: add missing doc for memory-backend option"
(v6.0) documented it:

    And document that x-use-canonical-path-for-ramblock-id,
    is considered to be stable to make sure it won't go away by accident.
    
    x- was intended for unstable/iternal properties, and not supposed to
    be stable option. However it's too late to rename (drop x-)
    it as it would mean that users will have to mantain both
    x-use-canonical-path-for-ramblock-id (for QEMU 5.0-5.2) versions
    and prefix-less for later versions.

Igor's reasoning still applies.

"x-origin" has always been stable.  Same argument.

>> We could document these exceptions, but documentation helps only
>> humans.  We want to recognize "unstable" in code, like "deprecated".
>>
>> Replace the convention by a new special feature flag "unstable".  It
>> will be recognized by the QAPI generator, like the existing feature
>> flag "deprecated", and unlike regular feature flags.
>
> FWIW, sounds good to me.

Thanks!

>> This commit updates documentation and prepares tests.  The next commit
>> updates the QAPI schema.  The remaining patches update the QAPI
>> generator and wire up -compat policy checking.
>> 
>> Signed-off-by: Markus Armbruster <armbru@redhat.com>
>> ---
>>  docs/devel/qapi-code-gen.rst            | 9 ++++++---
>>  tests/qapi-schema/qapi-schema-test.json | 7 +++++--
>>  tests/qapi-schema/qapi-schema-test.out  | 5 +++++
>>  3 files changed, 16 insertions(+), 5 deletions(-)
>
> [...]


commit 415b71a9f6e5bc37e84895d2e767cf4cfacd279b (HEAD)
Author: Markus Armbruster <armbru@redhat.com>
Date:   Sat Oct 9 09:01:21 2021 +0200

    qapi: Find x- without feature unstable DBG

diff --git a/scripts/qapi/schema.py b/scripts/qapi/schema.py
index b7b3fc0ce4..f2af1d7eea 100644
--- a/scripts/qapi/schema.py
+++ b/scripts/qapi/schema.py
@@ -14,6 +14,8 @@
 
 # TODO catching name collisions in generated code would be nice
 
+import sys
+
 from collections import OrderedDict
 import os
 import re
@@ -118,6 +120,11 @@ def describe(self):
         return "%s '%s'" % (self.meta, self.name)
 
 
+def check_have_feature_unstable(name, info, features):
+    if name.startswith('x-') and 'unstable' not in (f.name for f in features):
+        print(QAPISemError(info, f"XXX %{name} %{features}"), file=sys.stderr)
+
+
 class QAPISchemaVisitor:
     def visit_begin(self, schema):
         pass
@@ -718,6 +725,7 @@ def __init__(self, name, info, ifcond=None, features=None):
         self.features = features or []
 
     def connect_doc(self, doc):
+        check_have_feature_unstable(self.name, self.info, self.features)
         super().connect_doc(doc)
         if doc:
             for f in self.features:
@@ -745,6 +753,7 @@ def __init__(self, name, info, typ, optional, ifcond=None, features=None):
         self.features = features or []
 
     def check(self, schema):
+        check_have_feature_unstable(self.name, self.info, self.features)
         assert self.defined_in
         self.type = schema.resolve_type(self._type_name, self.info,
                                         self.describe)
@@ -789,6 +798,7 @@ def __init__(self, name, info, doc, ifcond, features,
 
     def check(self, schema):
         super().check(schema)
+        check_have_feature_unstable(self.name, self.info, self.features)
         if self._arg_type_name:
             self.arg_type = schema.resolve_type(
                 self._arg_type_name, self.info, "command's 'data'")
@@ -844,6 +854,7 @@ def __init__(self, name, info, doc, ifcond, features, arg_type, boxed):
 
     def check(self, schema):
         super().check(schema)
+        check_have_feature_unstable(self.name, self.info, self.features)
         if self._arg_type_name:
             self.arg_type = schema.resolve_type(
                 self._arg_type_name, self.info, "event's 'data'")