Clean up error output: color reset sequence and stat failure message - #482

Two unrelated-to-each-other-but-both-cosmetic fixes to how errors reach
the user, split out of psss#469 where they did not belong.
did/utils.py: the reset sequence emitted by color() was "033円[1;m", two
SGR parameters where the second is empty. Per ECMA-48 an empty parameter
defaults to 0, so a conforming terminal does end up reset, and this is
not a bug on one. But it relies on that defaulting rule, which parsers
that are not full terminals (log scrapers, CI log viewers, less without
-R, some ANSI-stripping libraries) handle inconsistently, and it briefly
enables bold on the way to resetting. "033円[0m" is the canonical reset
and does not read as a typo. The three test_utils.py assertions follow
from the change.
did/stats.py: the failure path logged "Skipping %s due to %s" with a
concurrent.futures.Future as the first argument, so it rendered a repr:
 Skipping <Future at 0x1063a4830 state=finished raised ReportError>
 due to Unable to fetch token
The address and state tell the user nothing, and the one useful piece of
information (which stat failed) is not there at all, since Future has no
back-reference to the submitted callable. Dropping the useless half
leaves the error text. Naming the stat properly would need a
future -> stat mapping, which is a larger change than this.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>