Question
Adding the panel means editing apps/web/src/components/RightPanelTabs.tsx, which upstream owns and
actively develops. That is a new row on docs/t3x/SEAMS.md, and the ledger tripwire says to
re-isolate before adding rows.
Measure it, do not estimate it:
- Churn on
RightPanelTabs.tsx over the 60 days before the merge-base, using the @<epoch> form
from the ledger's Regenerating section (the relative-suffix form silently returns 0).
- How many distinct places must change to register one tab? There are at least two
case switches
(around lines 413 and 482) plus the tab definition list around 175-225.
- Is there any extension point that avoids the edit, or is the switch the only way in?
- What would the row's
risk figure be — fork lines changed × churn?
Then answer the actual decision: is a tab worth the row, or does the overlay precedent in
apps/web/src/coil/AutoResumeOverlay.tsx (fork-owned, near-zero seam) buy enough? Note the
counter-argument already on record: a map is something you keep open while working, which is a tab,
and dodging one ledger row is a poor reason to build the wrong thing.
Question
Adding the panel means editing
apps/web/src/components/RightPanelTabs.tsx, which upstream owns andactively develops. That is a new row on
docs/t3x/SEAMS.md, and the ledger tripwire says tore-isolate before adding rows.
Measure it, do not estimate it:
RightPanelTabs.tsxover the 60 days before the merge-base, using the@<epoch>formfrom the ledger's Regenerating section (the relative-suffix form silently returns 0).
caseswitches(around lines 413 and 482) plus the tab definition list around 175-225.
riskfigure be —fork lines changed × churn?Then answer the actual decision: is a tab worth the row, or does the overlay precedent in
apps/web/src/coil/AutoResumeOverlay.tsx(fork-owned, near-zero seam) buy enough? Note thecounter-argument already on record: a map is something you keep open while working, which is a tab,
and dodging one ledger row is a poor reason to build the wrong thing.