fix(sessions): let a compression continuation reclaim its base title by xxxigm · Pull Request #48989 · NousResearch/hermes-agent

xxxigm · 2026-06-19T11:14:10Z

Summary

Reported on Discord (desktop app, repro also affects CLI/TUI): a user opened a
session titled fingerprint-scanner across the desktop app and hermes --tui.
On reopening, it showed as fingerprint-scanner #2 and renaming it back failed
with:

Rename failed — Title 'fingerprint-scanner' is already in use by session 20260619_131108_c3c5fd

…a session they couldn't find in their list.

Root cause (confirmed in the uploaded `agent.log`, reproducible on `main`)

The log shows context compression mid-session:

13:53:29 context compression started: session=20260619_131108_c3c5fd
13:54:13 context compression done:    session=20260619_135413_c7bf98

Compression intentionally rotates the session id: it ends the original,
creates a continuation child (parent_session_id), and auto-numbers the title
via get_next_title_in_lineage (name → name #2). The session list
(list_sessions_rich, project_compression_tips=True) then projects the
ended root behind its live tip, so the predecessor is invisible to the user.

But SessionDB.set_session_title()'s uniqueness check compared the new title
against all sessions (WHERE title = ? AND id != ?), including that hidden,
ended ancestor. So renaming the visible tip back to the base name dead-ended,
citing a session the user has no way to see or free. This isn't an intended
constraint — the uniqueness check is simply lineage-unaware.

Fix

When the conflicting title is held by a compression ancestor of the session
being renamed, transfer the title instead of raising: clear it from the
ended predecessor and apply it to the continuation. Ancestry is detected with
the same edge definition as get_compression_tip (parent end_reason = 'compression' and child.started_at >= parent.ended_at), so delegate/branch
children are not treated as continuations.

Invariants preserved:

Exactly one session ever carries the exact title (no duplicates) → resolve_session_by_title / tip projection keep working.
The parent_session_id lineage is untouched.
Genuine conflicts (unrelated sessions, non-compression children) still raise.

Test plan

tests/test_hermes_state.py::TestSessionTitleLineage — single- and multi-level chains transfer the title off the ended predecessor; unrelated sessions and non-compression children still conflict
Full tests/test_hermes_state.py (278 tests) green
tests/gateway/test_title_command.py + tests/cli/test_cli_new_session.py green (existing conflict behavior unchanged)

When context compression rotates a session, the original is ended and the continuation is auto-numbered (e.g. "name" -> "name #2"). The session list projects the ended root behind its live tip, so the user never sees the predecessor. But set_session_title's uniqueness check compared against ALL sessions, so renaming the visible tip back to "name" dead-ended with "Title 'name' is already in use by session <id the user can't find>". When the conflicting title is held by a compression ancestor of the session being renamed, transfer the title instead of raising: clear it from the ended predecessor and apply it to the continuation. Uniqueness is preserved (still exactly one session carries the title) and the parent-link lineage is untouched, so resume-by-title and tip projection keep working. Genuine conflicts with unrelated sessions, and with non-compression children (delegate/branch), still raise as before.

Regression tests for renaming a compression continuation back to its base title: single- and multi-level chains transfer the title off the ended predecessor, while unrelated sessions and non-compression children (created while the parent was live) still raise the uniqueness conflict.

…neage-48989 fix(sessions): let a compression continuation reclaim its base title (salvages #48989)

kshitijk4poor · 2026-06-19T12:19:36Z

Merged via #49000 (#49000). Your two commits were cherry-picked with authorship preserved; I added one refactor commit on top that expresses the compression-ancestor check as a single recursive CTE reusing the canonical _COMPRESSION_CHILD_SQL edge fragment (instead of a per-hop Python walk that re-encoded the continuation edge). Thanks for the clear root-cause analysis and the lineage tests — both made this an easy salvage.

…id (#38763) Context compression today rewrites the message list AND rotates the session id — it ends the session, forks a parent_session_id child, and renumbers the title (name -> name #2). That moving identity key is the root cause of a whole bug cluster: /goal lost (#33618), pending response lost at the split (#14238), orphan sessions (#33907), TUI sid desync (#36777), FTS search gaps + duplicate sidebar entries (#45117), null continuation cwd (#42228), and title-rename dead-ends (#48989). It also forced a large defensive apparatus (compression lock, contextvar/env/ logging triple-sync, orphan finalization, gateway SessionEntry re-propagation, tip projection) whose only job is surviving a mid-conversation id change. Add a compression.in_place config flag (default False during rollout). When True, compaction rewrites the transcript and rebuilds the system prompt but keeps the SAME session_id: no end_session, no child row, no title renumber, no contextvar/logging re-sync, no memory/context-engine session-switch. The conversation keeps one durable id for life, like Claude Code / Codex. Compaction is lossy by design — the pre-compaction transcript is summarized away, not archived. The rotation path is unchanged when the flag is off (moved verbatim into an else branch). Staged rollout: this PR ships the option behind a default-off flag for live validation; a follow-up flips the default and deletes the now-redundant rotation machinery, superseding the 14 open band-aid PRs in this area. - hermes_cli/config.py: add compression.in_place (default False), documented - agent/agent_init.py: resolve the flag -> agent.compression_in_place - agent/conversation_compression.py: branch compress_context() on the flag - tests/run_agent/test_in_place_compaction.py: in-place invariants + rotation regression guard + config default The pre-flush of current-turn messages (#47202) runs in BOTH modes, so no boundary data loss. Prompt-cache invariant preserved: the system-prompt rebuild is the same single sanctioned invalidation that already happens during compaction — no NEW invalidation. Message alternation preserved.

…id (NousResearch#38763) Context compression today rewrites the message list AND rotates the session id — it ends the session, forks a parent_session_id child, and renumbers the title (name -> name NousResearch#2). That moving identity key is the root cause of a whole bug cluster: /goal lost (NousResearch#33618), pending response lost at the split (NousResearch#14238), orphan sessions (NousResearch#33907), TUI sid desync (NousResearch#36777), FTS search gaps + duplicate sidebar entries (NousResearch#45117), null continuation cwd (NousResearch#42228), and title-rename dead-ends (NousResearch#48989). It also forced a large defensive apparatus (compression lock, contextvar/env/ logging triple-sync, orphan finalization, gateway SessionEntry re-propagation, tip projection) whose only job is surviving a mid-conversation id change. Add a compression.in_place config flag (default False during rollout). When True, compaction rewrites the transcript and rebuilds the system prompt but keeps the SAME session_id: no end_session, no child row, no title renumber, no contextvar/logging re-sync, no memory/context-engine session-switch. The conversation keeps one durable id for life, like Claude Code / Codex. Compaction is lossy by design — the pre-compaction transcript is summarized away, not archived. The rotation path is unchanged when the flag is off (moved verbatim into an else branch). Staged rollout: this PR ships the option behind a default-off flag for live validation; a follow-up flips the default and deletes the now-redundant rotation machinery, superseding the 14 open band-aid PRs in this area. - hermes_cli/config.py: add compression.in_place (default False), documented - agent/agent_init.py: resolve the flag -> agent.compression_in_place - agent/conversation_compression.py: branch compress_context() on the flag - tests/run_agent/test_in_place_compaction.py: in-place invariants + rotation regression guard + config default The pre-flush of current-turn messages (NousResearch#47202) runs in BOTH modes, so no boundary data loss. Prompt-cache invariant preserved: the system-prompt rebuild is the same single sanctioned invalidation that already happens during compaction — no NEW invalidation. Message alternation preserved.

…id (NousResearch#38763) Context compression today rewrites the message list AND rotates the session id — it ends the session, forks a parent_session_id child, and renumbers the title (name -> name #2). That moving identity key is the root cause of a whole bug cluster: /goal lost (NousResearch#33618), pending response lost at the split (NousResearch#14238), orphan sessions (NousResearch#33907), TUI sid desync (NousResearch#36777), FTS search gaps + duplicate sidebar entries (NousResearch#45117), null continuation cwd (NousResearch#42228), and title-rename dead-ends (NousResearch#48989). It also forced a large defensive apparatus (compression lock, contextvar/env/ logging triple-sync, orphan finalization, gateway SessionEntry re-propagation, tip projection) whose only job is surviving a mid-conversation id change. Add a compression.in_place config flag (default False during rollout). When True, compaction rewrites the transcript and rebuilds the system prompt but keeps the SAME session_id: no end_session, no child row, no title renumber, no contextvar/logging re-sync, no memory/context-engine session-switch. The conversation keeps one durable id for life, like Claude Code / Codex. Compaction is lossy by design — the pre-compaction transcript is summarized away, not archived. The rotation path is unchanged when the flag is off (moved verbatim into an else branch). Staged rollout: this PR ships the option behind a default-off flag for live validation; a follow-up flips the default and deletes the now-redundant rotation machinery, superseding the 14 open band-aid PRs in this area. - hermes_cli/config.py: add compression.in_place (default False), documented - agent/agent_init.py: resolve the flag -> agent.compression_in_place - agent/conversation_compression.py: branch compress_context() on the flag - tests/run_agent/test_in_place_compaction.py: in-place invariants + rotation regression guard + config default The pre-flush of current-turn messages (NousResearch#47202) runs in BOTH modes, so no boundary data loss. Prompt-cache invariant preserved: the system-prompt rebuild is the same single sanctioned invalidation that already happens during compaction — no NEW invalidation. Message alternation preserved.

…ion-title-lineage-48989 fix(sessions): let a compression continuation reclaim its base title (salvages NousResearch#48989)