|
From: | John Snow |
Subject: | Re: [PATCH 04/16] qapi/expr.py: Add assertion for union type 'check_dict' |
Date: | Thu, 24 Sep 2020 21:30:39 -0400 |
User-agent: | Mozilla/5.0 (X11; Linux x86_64; rv:68.0) Gecko/20100101 Thunderbird/68.11.0 |
On 9/24/20 9:08 PM, Eduardo Habkost wrote:
On Thu, Sep 24, 2020 at 08:47:31PM -0400, John Snow wrote:On 9/23/20 3:53 PM, Eduardo Habkost wrote:On Tue, Sep 22, 2020 at 05:13:01PM -0400, John Snow wrote:mypy isn't fond of allowing you to check for bool membership in a collection of str elements. Guard this lookup for precisely when we were given a name. Signed-off-by: John Snow <jsnow@redhat.com> --- scripts/qapi/expr.py | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/scripts/qapi/expr.py b/scripts/qapi/expr.py index f6b55a87c1..67892502e9 100644 --- a/scripts/qapi/expr.py +++ b/scripts/qapi/expr.py @@ -166,7 +166,9 @@ def check_type(value, info, source, raise QAPISemError(info, "%s should be an object or type name" % source) - permit_upper = allow_dict in info.pragma.name_case_whitelist + permit_upper = False + if isinstance(allow_dict, str): + permit_upper = allow_dict in info.pragma.name_case_whitelistWell, this keeps existing behavior, so: Reviewed-by: Eduardo Habkost <ehabkost@redhat.com> But: what exactly is the meaning of allow_dict=False, allow_dict=True, and allow_dict being a string?allow_dict = True -- allows the type to be an object describing the type. allow_dict: str -- allows the type to be an object (like True), but also passes a name in for the purposes of validating the name with the pragma whitelist(!)What.
(lol)
What's going on here is that when you pass in a name, bool(allow_dict) is True -- so we will allow the object being checked here to be a dict.
Also, when you pass in a name, that name is looked up in info.pragma.name_case_whitelist to check if the names of the keys in the dict being checked (again, allow_dict is implicitly true here) are allowed to use uppercase names.
I have some more experimental patches I didn't want to mix in with the type checking patches that try to extract pragma checks from this code and perform them elsewhere; but that's going to come much, much later.
--js
[Prev in Thread] | Current Thread | [Next in Thread] |