paul (Paul Klumpp)

paul запушил(а) main в paul/hermes-chatto-plugin

0c41688b19 Document contact info and authorship README now points users to Paul Klumpp (PaulK) and the #chatto-hackers room on ChattoHQ, the same Chatto instance the bot defaults to. plugin.yaml author field names the actual maintainer so the plugin manifest stops attributing the work to a generic 'Chatto community guys' string that nobody owns.

2 дней назад

paul запушил(а) main в paul/hermes-chatto-plugin

0a69a54fb3 Document inbound media and reaction forwarding in the README The capability matrix and Features sections omitted two behaviours the adapter implements: inbound attachments being downloaded into the media cache and handed to the agent as typed media, and human reactions being forwarded through the gateway reaction hook. Both exist in adapter.py (_cache_attachments / _forward_reaction), so the docs now match reality.

5 дней назад

paul запушил(а) main в paul/hermes-chatto-plugin

36b53a9eef Add acknowledgements Credit the plugin's origins where readers look for them: this began as a fork of jrimmer/hermes-chatto-plugin, whose groundwork we still build on and a regular stop for feature ideas worth adopting. chattolib and our adapter's early shape trace back to Felix (TheCodemancerLtd), so the links name their repositories directly.
fdb9544161 Credit the plugin's origins and fix the dead chattolib link The fork credit now reads as what it is: groundwork we still build on and a regular stop for feature ideas worth adopting. Both chattolib links pointed at github.com/chattocorp/chattolib, a repository that never existed — chattolib's actual home is TheCodemancerLtd/chattolib, so both now point there. TheCodemancerLtd is named as Felix, who also wrote chattolib; their branch link is pinned because their adapter only exists on feat/chatto-integration.
cf76c34436 Codify the commit-message rule: why over what The rewrite of the early history (courtesy commits, amendonitas) made the house style explicit, but nothing in this file asked for it — the next contributor had no rule to imitate on purpose. State it where the other readability judgements live: body carries the bug, constraint or rejected alternative; imperative subject; one logical change per commit.
122b2cbb61 Correct stale docs: no resume cursor, no home-channel fallback Two README claims had outlived their truth: - 'Resume cursor support' - protocol v1 reserves resume_cursor and the adapter carries no such state; reconnects rely on duplicate-event suppression, and events missed during downtime are simply gone. - 'First joined room' as the CHATTO_HOME_CHANNEL default (env table, Home Channel section, after-install.md step 2 and the plugin.yaml wizard prompt) - nothing implements that fallback. Unset means context-less cron/notification deliveries have no Chatto target and are not sent.
338566df5e Deliver the room roster on every channel turn, cap it at 20 members The per-room dedup kept the roster out of follow-up dispatches once a room was announced, but threads hold isolated sessions: a thread opened after the announcement never saw an audience at all. Cache and delivery had different scopes - the cache is room-scoped (one directory fetch, presence patched live), but the consumer is turn-scoped. Fix it by dropping dedup entirely: every channel turn now carries the rendered roster line from the cached projection. Repetition is free of API calls; membership events, /leave, stale-room pruning and reconnects still evict a room so its next turn refetches once. ROSTER_MEMBER_LIMIT drops from 100 to 20 to keep the prompt block small now that it rides on every message.
Просмотр сравнение для этих 234 коммитов »

1 неделя назад

paul запушил(а) main в paul/hermes-chatto-plugin

fdb9544161 Credit the plugin's origins and fix the dead chattolib link The fork credit now reads as what it is: groundwork we still build on and a regular stop for feature ideas worth adopting. Both chattolib links pointed at github.com/chattocorp/chattolib, a repository that never existed — chattolib's actual home is TheCodemancerLtd/chattolib, so both now point there. TheCodemancerLtd is named as Felix, who also wrote chattolib; their branch link is pinned because their adapter only exists on feat/chatto-integration.
de6d75c6bd Credit the plugin's origins and fix the dead chattolib link The fork credit now reads as what it is: groundwork we still build on and a regular stop for feature ideas worth adopting. chattolib has no GitHub repository of its own — both links pointed at one that never existed; PyPI is where the package actually lives. TheCodemancerLtd is named as Felix, who also wrote chattolib; their branch link is pinned because their adapter only exists on feat/chatto-integration.
Просмотр сравнение для этих 2 коммитов »

1 неделя назад

paul запушил(а) main в paul/hermes-chatto-plugin

de6d75c6bd Credit the plugin's origins and fix the dead chattolib link The fork credit now reads as what it is: groundwork we still build on and a regular stop for feature ideas worth adopting. chattolib has no GitHub repository of its own — both links pointed at one that never existed; PyPI is where the package actually lives. TheCodemancerLtd is named as Felix, who also wrote chattolib; their branch link is pinned because their adapter only exists on feat/chatto-integration.
cf76c34436 Codify the commit-message rule: why over what The rewrite of the early history (courtesy commits, amendonitas) made the house style explicit, but nothing in this file asked for it — the next contributor had no rule to imitate on purpose. State it where the other readability judgements live: body carries the bug, constraint or rejected alternative; imperative subject; one logical change per commit.
122b2cbb61 Correct stale docs: no resume cursor, no home-channel fallback Two README claims had outlived their truth: - 'Resume cursor support' - protocol v1 reserves resume_cursor and the adapter carries no such state; reconnects rely on duplicate-event suppression, and events missed during downtime are simply gone. - 'First joined room' as the CHATTO_HOME_CHANNEL default (env table, Home Channel section, after-install.md step 2 and the plugin.yaml wizard prompt) - nothing implements that fallback. Unset means context-less cron/notification deliveries have no Chatto target and are not sent.
338566df5e Deliver the room roster on every channel turn, cap it at 20 members The per-room dedup kept the roster out of follow-up dispatches once a room was announced, but threads hold isolated sessions: a thread opened after the announcement never saw an audience at all. Cache and delivery had different scopes - the cache is room-scoped (one directory fetch, presence patched live), but the consumer is turn-scoped. Fix it by dropping dedup entirely: every channel turn now carries the rendered roster line from the cached projection. Repetition is free of API calls; membership events, /leave, stale-room pruning and reconnects still evict a room so its next turn refetches once. ROSTER_MEMBER_LIMIT drops from 100 to 20 to keep the prompt block small now that it rides on every message.
8af1cef5c3 Render the room roster from a cached projection, presence patched live The previous approach re-fetched the member directory whenever a presence_changed event marked an announced room stale — with real-world presence churn that meant one list_room_members call per inbound message after any flip, and every room was invalidated regardless of who moved. Rework it as a miniature projection of server state, modelled on how the Chatto web frontend handles presence (per-user cache patched in place, member lists never re-fetched): each announced room keeps member IDs plus its users in the existing _user_cache and the last rendered line. presence_changed patches the cached user; the next turn in the room re-renders from cache and re-delivers only when the line actually changed. Membership events (user_joined_room/user_left_room), /leave, stale-room pruning and reconnects discard a room's projection so its next turn refetches once - protocol v1 has no presence snapshot on subscribe, so discarding is the bootstrap. Also corrects the event router: projection_event belongs to realtime protocol v2 on unreleased Chatto main and can never arrive over chattolib v1; noted where to revisit it. AGENTS.md gains two guidance entries born from this thread: check existing state fields before adding new ones (_user_cache instead of a second user store), and treat the Chatto web frontend as the architecture reference for client-side behaviour.
Просмотр сравнение для этих 245 коммитов »

1 неделя назад

paul запушил(а) main в paul/hermes-chatto-plugin

2cdee25277 Credit the plugin's origins and fix the dead chattolib link The fork credit now reads as what it is: groundwork we still build on and a regular stop for feature ideas worth adopting. chattolib has no GitHub repository of its own — both links pointed at one that never existed; PyPI is where the package actually lives. TheCodemancerLtd is named as Felix, who also wrote chattolib; their branch link is pinned because their adapter only exists on feat/chatto-integration.
d37aec3f2c Rework the acknowledgements and fix the dead chattolib link chattolib has no GitHub repository of its own — both links pointed at one that never existed; PyPI is where the package actually lives. The fork credit now reads as what it is (groundwork we still build on and check for feature ideas), and TheCodemancerLtd is named as Felix, who also wrote chattolib.
a681fa483d Credit the fork origin and the adapter inspiration The plugin no longer resembles jrimmer's original much, but attribution belongs in the README before someone digs through history for it; the Codemancer branch link is pinned because that adapter only exists there.
Просмотр сравнение для этих 3 коммитов »

1 неделя назад

paul запушил(а) main в paul/hermes-chatto-plugin

d37aec3f2c Rework the acknowledgements and fix the dead chattolib link chattolib has no GitHub repository of its own — both links pointed at one that never existed; PyPI is where the package actually lives. The fork credit now reads as what it is (groundwork we still build on and check for feature ideas), and TheCodemancerLtd is named as Felix, who also wrote chattolib.

1 неделя назад

paul запушил(а) main в paul/hermes-chatto-plugin

a681fa483d Credit the fork origin and the adapter inspiration The plugin no longer resembles jrimmer's original much, but attribution belongs in the README before someone digs through history for it; the Codemancer branch link is pinned because that adapter only exists there.

1 неделя назад

paul запушил(а) main в paul/hermes-chatto-plugin

c1be128c3b Codify the commit-message rule: why over what The rewrite of the early history (courtesy commits, amendonitas) made the house style explicit, but nothing in this file asked for it — the next contributor had no rule to imitate on purpose. State it where the other readability judgements live: body carries the bug, constraint or rejected alternative; imperative subject; one logical change per commit.
c6d893dd7d Correct stale docs: no resume cursor, no home-channel fallback Two README claims had outlived their truth: - 'Resume cursor support' - protocol v1 reserves resume_cursor and the adapter carries no such state; reconnects rely on duplicate-event suppression, and events missed during downtime are simply gone. - 'First joined room' as the CHATTO_HOME_CHANNEL default (env table, Home Channel section, after-install.md step 2 and the plugin.yaml wizard prompt) - nothing implements that fallback. Unset means context-less cron/notification deliveries have no Chatto target and are not sent.
9f11357774 Deliver the room roster on every channel turn, cap it at 20 members The per-room dedup kept the roster out of follow-up dispatches once a room was announced, but threads hold isolated sessions: a thread opened after the announcement never saw an audience at all. Cache and delivery had different scopes - the cache is room-scoped (one directory fetch, presence patched live), but the consumer is turn-scoped. Fix it by dropping dedup entirely: every channel turn now carries the rendered roster line from the cached projection. Repetition is free of API calls; membership events, /leave, stale-room pruning and reconnects still evict a room so its next turn refetches once. ROSTER_MEMBER_LIMIT drops from 100 to 20 to keep the prompt block small now that it rides on every message.
bdd6acb0a9 Render the room roster from a cached projection, presence patched live The previous approach re-fetched the member directory whenever a presence_changed event marked an announced room stale — with real-world presence churn that meant one list_room_members call per inbound message after any flip, and every room was invalidated regardless of who moved. Rework it as a miniature projection of server state, modelled on how the Chatto web frontend handles presence (per-user cache patched in place, member lists never re-fetched): each announced room keeps member IDs plus its users in the existing _user_cache and the last rendered line. presence_changed patches the cached user; the next turn in the room re-renders from cache and re-delivers only when the line actually changed. Membership events (user_joined_room/user_left_room), /leave, stale-room pruning and reconnects discard a room's projection so its next turn refetches once - protocol v1 has no presence snapshot on subscribe, so discarding is the bootstrap. Also corrects the event router: projection_event belongs to realtime protocol v2 on unreleased Chatto main and can never arrive over chattolib v1; noted where to revisit it. AGENTS.md gains two guidance entries born from this thread: check existing state fields before adding new ones (_user_cache instead of a second user store), and treat the Chatto web frontend as the architecture reference for client-side behaviour.
30bc4f75ec Drop getattr masking of the base-class _reaction_handler The attribute is declared on BasePlatformAdapter (base.py:3017) with a None default, so it is always present. Direct access is typed and matches the AGENTS.md typed-access rule; the getattr fallback only hid that.
Просмотр сравнение для этих 238 коммитов »

1 неделя назад

paul удалил ветку review-baseadapter-overrides в paul/hermes-chatto-plugin

1 неделя назад

paul удалил ветку feat/roster-presence-refresh в paul/hermes-chatto-plugin

1 неделя назад

paul удалил ветку feat/roster-presence-cache в paul/hermes-chatto-plugin

1 неделя назад

paul удалил ветку feat/edit-dispatch в paul/hermes-chatto-plugin

1 неделя назад

paul удалил ветку debug/roster-logging в paul/hermes-chatto-plugin

1 неделя назад

paul удалил ветку debug-location-logging в paul/hermes-chatto-plugin

1 неделя назад

paul удалил ветку code-review в paul/hermes-chatto-plugin

1 неделя назад

paul запушил(а) feat/roster-presence-cache в paul/hermes-chatto-plugin

3fc158ab91 Correct stale docs: no resume cursor, no home-channel fallback Two README claims had outlived their truth: - 'Resume cursor support' - protocol v1 reserves resume_cursor and the adapter carries no such state; reconnects rely on duplicate-event suppression, and events missed during downtime are simply gone. - 'First joined room' as the CHATTO_HOME_CHANNEL default (env table, Home Channel section, after-install.md step 2 and the plugin.yaml wizard prompt) - nothing implements that fallback. Unset means context-less cron/notification deliveries have no Chatto target and are not sent.
025aa1d63f Deliver the room roster on every channel turn, cap it at 20 members The per-room dedup kept the roster out of follow-up dispatches once a room was announced, but threads hold isolated sessions: a thread opened after the announcement never saw an audience at all. Cache and delivery had different scopes - the cache is room-scoped (one directory fetch, presence patched live), but the consumer is turn-scoped. Fix it by dropping dedup entirely: every channel turn now carries the rendered roster line from the cached projection. Repetition is free of API calls; membership events, /leave, stale-room pruning and reconnects still evict a room so its next turn refetches once. ROSTER_MEMBER_LIMIT drops from 100 to 20 to keep the prompt block small now that it rides on every message.
99990384aa Render the room roster from a cached projection, presence patched live The previous approach re-fetched the member directory whenever a presence_changed event marked an announced room stale — with real-world presence churn that meant one list_room_members call per inbound message after any flip, and every room was invalidated regardless of who moved. Rework it as a miniature projection of server state, modelled on how the Chatto web frontend handles presence (per-user cache patched in place, member lists never re-fetched): each announced room keeps member IDs plus its users in the existing _user_cache and the last rendered line. presence_changed patches the cached user; the next turn in the room re-renders from cache and re-delivers only when the line actually changed. Membership events (user_joined_room/user_left_room), /leave, stale-room pruning and reconnects discard a room's projection so its next turn refetches once - protocol v1 has no presence snapshot on subscribe, so discarding is the bootstrap. Also corrects the event router: projection_event belongs to realtime protocol v2 on unreleased Chatto main and can never arrive over chattolib v1; noted where to revisit it. AGENTS.md gains two guidance entries born from this thread: check existing state fields before adding new ones (_user_cache instead of a second user store), and treat the Chatto web frontend as the architecture reference for client-side behaviour.
8081f459df Drop getattr masking of the base-class _reaction_handler The attribute is declared on BasePlatformAdapter (base.py:3017) with a None default, so it is always present. Direct access is typed and matches the AGENTS.md typed-access rule; the getattr fallback only hid that.
fa18c2b2c1 Refresh room roster on presence change instead of freezing at thread start The member roster was announced exactly once per thread and then never updated, so presence moves (away/offline) inside a long-lived room never reached the agent. Make the roster room-scoped: the first turn in a room carries it, and a presence_changed event marks the room stale so the next inbound turn re-fetches and re-delivers it. The bot's own 60s presence refresh is filtered out so it never churns announced rooms; a failed lookup leaves the room unannounced for a retry. Delete the stale PLAN.md (edit-dispatch feature is already shipped) and update README/AGENTS docs. Also fix an undefined-name bug in the _make_edited_event test helper.
Просмотр сравнение для этих 11 коммитов »

1 неделя назад

paul запушил(а) main в paul/hermes-chatto-plugin

3fc158ab91 Correct stale docs: no resume cursor, no home-channel fallback Two README claims had outlived their truth: - 'Resume cursor support' - protocol v1 reserves resume_cursor and the adapter carries no such state; reconnects rely on duplicate-event suppression, and events missed during downtime are simply gone. - 'First joined room' as the CHATTO_HOME_CHANNEL default (env table, Home Channel section, after-install.md step 2 and the plugin.yaml wizard prompt) - nothing implements that fallback. Unset means context-less cron/notification deliveries have no Chatto target and are not sent.
025aa1d63f Deliver the room roster on every channel turn, cap it at 20 members The per-room dedup kept the roster out of follow-up dispatches once a room was announced, but threads hold isolated sessions: a thread opened after the announcement never saw an audience at all. Cache and delivery had different scopes - the cache is room-scoped (one directory fetch, presence patched live), but the consumer is turn-scoped. Fix it by dropping dedup entirely: every channel turn now carries the rendered roster line from the cached projection. Repetition is free of API calls; membership events, /leave, stale-room pruning and reconnects still evict a room so its next turn refetches once. ROSTER_MEMBER_LIMIT drops from 100 to 20 to keep the prompt block small now that it rides on every message.
99990384aa Render the room roster from a cached projection, presence patched live The previous approach re-fetched the member directory whenever a presence_changed event marked an announced room stale — with real-world presence churn that meant one list_room_members call per inbound message after any flip, and every room was invalidated regardless of who moved. Rework it as a miniature projection of server state, modelled on how the Chatto web frontend handles presence (per-user cache patched in place, member lists never re-fetched): each announced room keeps member IDs plus its users in the existing _user_cache and the last rendered line. presence_changed patches the cached user; the next turn in the room re-renders from cache and re-delivers only when the line actually changed. Membership events (user_joined_room/user_left_room), /leave, stale-room pruning and reconnects discard a room's projection so its next turn refetches once - protocol v1 has no presence snapshot on subscribe, so discarding is the bootstrap. Also corrects the event router: projection_event belongs to realtime protocol v2 on unreleased Chatto main and can never arrive over chattolib v1; noted where to revisit it. AGENTS.md gains two guidance entries born from this thread: check existing state fields before adding new ones (_user_cache instead of a second user store), and treat the Chatto web frontend as the architecture reference for client-side behaviour.
8081f459df Drop getattr masking of the base-class _reaction_handler The attribute is declared on BasePlatformAdapter (base.py:3017) with a None default, so it is always present. Direct access is typed and matches the AGENTS.md typed-access rule; the getattr fallback only hid that.
fa18c2b2c1 Refresh room roster on presence change instead of freezing at thread start The member roster was announced exactly once per thread and then never updated, so presence moves (away/offline) inside a long-lived room never reached the agent. Make the roster room-scoped: the first turn in a room carries it, and a presence_changed event marks the room stale so the next inbound turn re-fetches and re-delivers it. The bot's own 60s presence refresh is filtered out so it never churns announced rooms; a failed lookup leaves the room unannounced for a retry. Delete the stale PLAN.md (edit-dispatch feature is already shipped) and update README/AGENTS docs. Also fix an undefined-name bug in the _make_edited_event test helper.
Просмотр сравнение для этих 5 коммитов »

1 неделя назад

paul запушил(а) feat/roster-presence-cache в paul/hermes-chatto-plugin

a364330518 Correct stale docs: no resume cursor, no home-channel fallback Two README claims had outlived their truth: - 'Resume cursor support' - protocol v1 reserves resume_cursor and the adapter carries no such state; reconnects rely on duplicate-event suppression, and events missed during downtime are simply gone. - 'First joined room' as the CHATTO_HOME_CHANNEL default (env table, Home Channel section, after-install.md step 2 and the plugin.yaml wizard prompt) - nothing implements that fallback. Unset means context-less cron/notification deliveries have no Chatto target and are not sent.

1 неделя назад

paul запушил(а) feat/roster-presence-cache в paul/hermes-chatto-plugin

622560ac60 Deliver the room roster on every channel turn, cap it at 20 members The per-room dedup kept the roster out of follow-up dispatches once a room was announced, but threads hold isolated sessions: a thread opened after the announcement never saw an audience at all. Cache and delivery had different scopes - the cache is room-scoped (one directory fetch, presence patched live), but the consumer is turn-scoped. Fix it by dropping dedup entirely: every channel turn now carries the rendered roster line from the cached projection. Repetition is free of API calls; membership events, /leave, stale-room pruning and reconnects still evict a room so its next turn refetches once. ROSTER_MEMBER_LIMIT drops from 100 to 20 to keep the prompt block small now that it rides on every message.

1 неделя назад