On 30/10/2025 11.01, Daniel P. Berrangé wrote:
On Thu, Oct 30, 2025 at 10:45:32AM +0100, Thomas Huth wrote:
On 30/10/2025 10.42, Thomas Huth wrote:
On 30/10/2025 10.33, Daniel P. Berrangé wrote:
On Thu, Oct 30, 2025 at 10:26:38AM +0100, Thomas Huth wrote:
From: Thomas Huth <[email protected]>
The argparse.FileType() type has been deprecated in the latest argparse
version (e.g. the one from Fedora 43), now causing the test_bad_vmstate
functional test to fail since there are unexpected strings in the output.
Change the script to use pathlib.Path instead to fix the test_bad_vmstate
test and to be prepared for the future when the deprecated FileType gets
removed completely.
...
@@ -393,10 +394,10 @@ def main():
help='reverse the direction')
args = parser.parse_args()
- src_data = json.load(args.src)
- dest_data = json.load(args.dest)
- args.src.close()
- args.dest.close()
+ with open(args.src, 'r', encoding='utf-8') as src_fh:
+ src_data = json.load(src_fh)
+ with open(args.dest, 'r', encoding='utf-8') as dst_fh:
+ dest_data = json.load(dst_fh)
This could be
src_data = json.load(args.src.read_text('utf-8'))
dest_data = json.load(args.dest.read_text('utf-8'))
Does not work, looks like the load() function cannot deal with a string:
$ scripts/vmstate-static-checker.py -s
tests/data/vmstate-static-checker/ dump1.json -d
tests/data/vmstate-static-checker/dump2.json
Traceback (most recent call last):
File "../scripts/vmstate-static-checker.py", line 439, in <module>
sys.exit(main())
~~~~^^
File "../scripts/vmstate-static-checker.py", line 397, in main
src_data = json.load(args.src.read_text('utf-8'))
File "/usr/lib64/python3.13/json/__init__.py", line 293, in load
return loads(fp.read(),
^^^^^^^
AttributeError: 'str' object has no attribute 'read'
Ok, there also seems to be a json.loads() function (with "s" at the end)
that seems to do the job ... but IMHO it would be better to continue using
the load() function here to let the json parser decide how to read the file.
ok, I don't mind that much
I'm basically thinking of situations where you have a big dump file: When
using loads(), you have to completely read the file into memory first, maybe
using a lot of memory. With load(), the parser could digest it bit by bit
instead, and e.g. bail out early if there is a bug in the json format of the
file. Well, it likely doesn't matter too much for the file sizes that we use
in QEMU, but still... I think it's nicer with load().
Reviewed-by: Daniel P. Berrangé <[email protected]>
Thanks!
Thomas
