On Mon, May 27, 2019 at 11:57:20AM +0200, Phil Sutter wrote: > Hi Pablo, > > On Fri, May 24, 2019 at 10:45:14PM +0200, Pablo Neira Ayuso wrote: > > On Wed, May 22, 2019 at 06:14:51PM +0200, Phil Sutter wrote: > > > Round three of JSON validation enhancement. > > > > > > Changes since v2: > > > - Make enhancement to nftables module Python3 compliant. > > > - Complain in nft-test.py if --schema was given without --json. > > > > > > Changes since v1: > > > - Fix patch 2 commit message, thanks to Jones Desougi who reported the > > > inconsistency. > > > > > > Changes since RFC: > > > - Import builtin traceback module unconditionally. > > > > > > Phil Sutter (2): > > > py: Implement JSON validation in nftables module > > > tests/py: Support JSON validation > > > > > > py/Makefile.am | 2 +- > > > py/nftables.py | 29 +++++++++++++++++++++++++++++ > > > py/schema.json | 17 +++++++++++++++++ > > > py/setup.py | 1 + > > > tests/py/nft-test.py | 25 ++++++++++++++++++++++++- > > > > Where is ruleset-schema.json? > > > > + "id": "http://netfilter.org/nftables/ruleset-schema.json";, > > + "description": "libnftables JSON API schema", > > Oh, I forgot about that. There are actually two problems with it: On one > hand, current draft version suggests to use "$id" instead of "id" for > the property name. On the other, the URL should point to an online > location of the document itself, which is obviously not correct. We can upload it to exactly the location you specify above, that won't be a problem. > Given that it is optional according to the draft, I would just drop it > for now. What do you think? Drop for now is fine with fine. Thanks.
