3.0.1

3.0.1

Fixes

Copy Deep Link regexes are easier to read

Generated deep links used to join words with \s+:

stream.md:6:/^#\s+2026-08-05\s+"The\s+Containment\s+Brothers"/

Words are now joined by a literal space, so the same link reads:

stream.md:6:/^# 2026-08-05 "The Containment Brothers"/

Only gaps that aren't a single plain space — a tab, or a run of spaces —
still become \s+, so the regex always matches the line it was copied from.
Leading indentation is still ^\s*, which keeps the anchor working when a
line gets re-indented. Existing links with \s+ continue to work.

Deep links containing spaces or quotes now open from any cursor position

With the cursor inside the quoted part of a link like
stream.md:6:/^# 2026-08-05 "The Containment Brothers"/, Open URL would
select just The Containment Brothers instead of the whole link, and fail.
Placing the cursor at the start of the link had the same problem, stopping at
the first space. Deep-link tokens are now matched as a unit, so any cursor
position inside one opens it — including when the link is wrapped in quotes or
backticks, inside a markdown link, or sharing a line with another link.

Deep links whose regex ends in a colon now parse

f.py:12:/^def foo\(x\):/ and f.md:1:/^key: value/ were unparseable — the
colon inside the regex body was mistaken for the location separator, so Python
def/if lines and YAML keys produced broken links. A location suffix now has
to be complete (a bare line number, an N-M range, or a closed /.../ or
"...") before it's treated as one.