- clang-format on code
- NULL -> nullptr
- some (char *) changed in std::string favour
- Removed a memory leak.
- Moved struct inside the table namespace
This is in response to #411. Allowing you to specify arguments on the
command-line has more edge-cases than I'd prefer, so I think this is a
bit more of a sustainable solution, especially given that you're already
supplying the tool with a path to your config file (now you can just
track one additional file).
port_inode and socket_inode have caused a few issues lately and, as of
right now, they both have open issues against them. For the time being,
I'm going to blacklist them. When the tables are production-ready, we
can re-add them back in to the base linux build.
My intent in this diff was mostly style linting, so I disabled
non-stylistic pylint linters that fired in the interests of making this
a sane diff with one purpose: stylistic consistency and conformance. If
I disabled them it means they were thrown somewhere and should probably
be looked into some time :-)
This diff adds:
- PEP8 conformance (tested with pep8)
- A cleanup of stuff shown by `pylint`, with quite a few linters
disabled. See above for rationale to disable these -- in the end this
was mostly unused variables, unused imports, etc). These are the
linters I disabled:
- attribute-defined-outside-init
- bad-classmethod-argument
- bare-except
- broad-except
- exec-used
- invalid-name
- logging-not-lazy
- method-hidden
- missing-docstring
- redefined-outer-name
- too-few-public-methods
- too-many-instance-attributes
- too-many-locals
- unused-argument
- flake8 warnings fixed (warnings were about redefinition of previous
variables in a listcomp)
I didn't do anything with git-clang-format since it's from an external
project and it's possible that there may be a wish to merge it in again
later if it gets updated upstream, but you could, of course, apply this
to that script as well if you so wish. Right now it's not at all PEP8
conformant.
