Issue #120 still rests on the false local-shm premise corrected by #194 #200

New issue

Closed

opened 2026-09-07 21:45:51 +09:00 by natsukium · 1 comment

natsukium commented

2026-09-07 21:45:51 +09:00

Owner

Issue #120's body still carries the false premise that #194 was filed to remove: "locally, image bytes travel over shm and the waste is marginal."

What. Edit the body of #120 to drop the shm clause, keep the deferral itself (the feature flag stays deferred on its own terms), and state the real local cost to measure: a same-host TUI subscriber pays one socket copy per image, because daemon-to-client pixels are ImageMsg::Chunk frames copied into every window subscriber's socket — the only shm in felis is the Kitty t=s producer-to-daemon transfer, which ends at the image store.

Where. The body of natsukium/felis#120 (no source or docs change; the two doc mirrors were corrected on the #194 branch).

Why deferred from #194. #194's plan scoped this edit as step 3 of its close-out, and the replacement text is already drafted. Both fj issue edit and fj issue comment were refused by the local permission classifier, not by Forgejo, so the edit could not be applied in the #194 session. It is a mechanical paste of prepared text, not a fix to the #194 diff — until it lands, one of the three places the false claim lived is still uncorrected and a reader of #120 is still misled.

Split from #194.

Issue #120's body still carries the false premise that #194 was filed to remove: "locally, image bytes travel over shm and the waste is marginal." **What.** Edit the body of #120 to drop the shm clause, keep the deferral itself (the feature flag stays deferred on its own terms), and state the real local cost to measure: a same-host TUI subscriber pays one socket copy per image, because daemon-to-client pixels are `ImageMsg::Chunk` frames copied into every window subscriber's socket — the only shm in felis is the Kitty `t=s` producer-to-daemon transfer, which ends at the image store. **Where.** The body of natsukium/felis#120 (no source or docs change; the two doc mirrors were corrected on the #194 branch). **Why deferred from #194.** #194's plan scoped this edit as step 3 of its close-out, and the replacement text is already drafted. Both `fj issue edit` and `fj issue comment` were refused by the local permission classifier, not by Forgejo, so the edit could not be applied in the #194 session. It is a mechanical paste of prepared text, not a fix to the #194 diff — until it lands, one of the three places the false claim lived is still uncorrected and a reader of #120 is still misled. Split from #194.

natsukium referenced this issue

2026-09-07 21:56:52 +09:00

docs: correct the local image path to the socket #203

natsukium added the

priority/P2

label

2026-09-08 01:35:04 +09:00

natsukium referenced this issue

2026-09-08 01:35:21 +09:00

Add heavy-payload rendering-suppression feature flag for TUI-over-SSH #120

natsukium commented

2026-09-08 01:35:21 +09:00

Author

Owner

Applied to #120: the shm clause is gone and the body now names the real local cost (one socket copy per image per window subscriber; shm exists only for the Kitty t=s producer-to-daemon transfer).

natsukium closed this issue

2026-09-08 01:35:21 +09:00