Improve process spawn speed and remove trailing lines.

[matu3ba]

No description provided.

[matu3ba]

Should I file more issues for all the formatting and linting issues ruff shows or what is the policy?

With ruff.toml (ruff version ruff 0.8.2)

line-length = 100

[format]
quote-style = "single"
indent-style = "space"

and running ruff format I get

cpuset/commands/common.py |   5 +--
 cpuset/commands/mem.py    |  44 +++++++++++-----------
 cpuset/commands/proc.py   | 319 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++-----------------------------------------------------------------------
 cpuset/commands/set.py    | 283 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++--------------------------------------------------------
 cpuset/commands/shield.py | 196 +++++++++++++++++++++++++++++++++++++++++++++++++++++-------------------------------------------
 cpuset/config.py          |  22 +++++++----
 cpuset/cset.py            | 288 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++---------------------------------------------------------------
 cpuset/main.py            |  77 +++++++++++++++++++++-----------------
 cpuset/util.py            |  54 ++++++++++++++++-----------
 setup.py                  |  44 +++++++++++-----------
 t/test_cset.py            |  66 ++++++++++++++++-----------------
 t/test_util.py            |  11 +++---
 12 files changed, 789 insertions(+), 620 deletions(-)

ruff check shows me

Found 160 errors.
[*] 24 fixable with the `--fix` option (12 hidden fixes can be enabled with the `--unsafe-fixes` option).

You might want to run this yourself to have less stuff to review.

[Werkov]

The whitespace cleanup (c8d6482) looks good to me, thanks.
The subprocess substitution isn't likely speeding up it as expected.

There is no coding style policy. Would unified style help you before further modifications or what is it helpful for (it's non-functional change, right)?

[Werkov]

Are you sure this is really faster?
I think the comment suggests reading info from procfs directly instead of spawning a helper program.

[matu3ba]

Would unified style help you before further modifications or what is it helpful for (it's non-functional change, right)?

Unfortunately ruff as linter shows on linting both formatting and non-formatting problems. From my experience linting catches mostly unhandled cases and type confusion.
Personally I disagree very often with what Python decided in their formatting pep(s), but configuring it feels too painful/annoying.

The subprocess substitution isn't likely speeding up it as expected.

My bad, will read from procfs directly.