Repository navigation
compression.threshold_tokens: null takes effect live; cap survives a failed config load (follow-up to #115986) - #119406
Merged
kshitijk4poor merged 3 commits intoSep 22, 2026
Conversation
…d of waiting for a restart The agent-cache signature reads the raw config file (no DEFAULT_CONFIG merge) and mapped an absent key to None. Since NousResearch#115986 shipped compression.threshold_tokens=256K, the documented opt-out (`hermes config set compression.threshold_tokens null`) produced the same None as "never set", so the signature did not change and live gateway sessions kept compacting at 256K until the process restarted (community report on NousResearch#115986). Absent keys now carry DEFAULT_CONFIG's value — what the agent was actually built with — so absent == explicit default (no spurious rebuild) and absent != explicit null.
…ke every other compression key
`_parse_compression_config` documents "Defaults here MUST match DEFAULT_CONFIG" because the
config-load failure path hands it `{}`. Every key had an inline fallback except the cap
added in NousResearch#115986, so a broken config.yaml kept `threshold=0.50` but silently dropped the
256K cap — the pre-NousResearch#115986 500K trigger on 1M-window models. An explicit null still
means ratio-only.
…reshold_percent fallback `min(threshold_tokens_cap, context_length)` behind the same `cap is not None and cap > 0` guard was written in both `_derive_trigger` and `_apply_threshold_tokens_cap`, contradicting `_derive_trigger`'s "one place the trigger math lives". Both now call `_effective_threshold_cap(context_length)`. The `getattr(self, "_config_threshold_percent", ...)` fallback was dead: `__init__` sets the attribute unconditionally and no bare-`__new__` test instance reaches `_derive_trigger`. Behaviour-preserving; the cap tests still fail when the helper is mutated to return None.
alt-glitch
added
type/bug
kshitijk4poor
added a commit
that referenced
this pull request
Sep 22, 2026
…t limits, one cfg_get per key #119406 made absent keys carry DEFAULT_CONFIG's value in the cache signature, but the legacy bool branch still emitted None for max_snapshots/max_total_size_mb/max_file_size_mb while `_checkpoint_agent_kwargs` builds that agent from DEFAULT_CONFIG — so migrating to `checkpoints: {enabled: true}` rebuilt the cached agent for nothing. The dict/absent branches were also re-implementing cfg_get's own semantics; one `cfg_get(cfg, section, key, default=…)` covers absent section, non-dict section, absent key and explicit null. DEFAULT_CONFIG is imported at module top (hermes_cli.config already is; gateway/restart.py sets the precedent).
kshitijk4poor
added a commit
that referenced
this pull request
Sep 22, 2026
…trim comments and tests `_cap_binds` was the last inline copy of the cap-clamped-to-window predicate that #119406 centralised; the banner still prints the raw configured cap. Plugin engines without the helper report "not binding", as before when no cap was set. The threshold_tokens comment and the new test module stop narrating the incident; the parametrize takes the whole agent config so the `{}` failure path is spelled out rather than derived from a ternary.
7 tasks
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Setting
compression.threshold_tokens: nullnow takes effect on a running gateway, and a failed config load no longer silently drops the 256K cap.Follow-up to #115986 (the 256K default cap). The community report on that PR noted that the documented opt-out —
hermes config set compression.threshold_tokens null— only took effect after a gateway restart. Confirmed in code, plus two smaller seams found while re-reviewing the merged range.Changes
gateway/run_agent_cache.py::_extract_cache_busting_config— the cache signature reads the raw config file and mapped an absent key toNone, sothreshold_tokens: nullproduced the same signature as "never set" and the cached agent kept the cap. Absent keys now carryDEFAULT_CONFIG's value (what the agent was built with): absent == explicit default (no spurious rebuild), absent != explicit null (the opt-out rebuilds). Applies to every key in_CACHE_BUSTING_CONFIG_KEYS; only keys with a non-Nonedefault change behaviour.agent/agent_init.py::_parse_compression_config— every compression key had an inline fallback matchingDEFAULT_CONFIGexceptthreshold_tokens, so the{}config-load-failure path keptthreshold=0.50but dropped the cap. Now reads the default like its siblings; explicitnullstill means ratio-only.agent/context_compressor.py— themin(cap, context_length)clamp was duplicated in_derive_triggerand_apply_threshold_tokens_cap; both call one_effective_threshold_cap(). Dropped a deadgetattr(self, "_config_threshold_percent", …)fallback (__init__sets it unconditionally; no bare-__new__test instance reaches the method). Behaviour-preserving.Validation
origin/maintest_explicit_null_differs_from_absent_when_default_is_set,test_missing_keys_yield_the_shipped_default,test_absent_threshold_tokens_falls_back_like_every_other_key[absent]fail on base, pass on the branch_effective_threshold_cap→return Nonefails 5 cap tests across 3 files; restored from snapshotruff,check-windows-footguns.py --all,check_compat_pointers.py,git diff --checkNot in this PR: whether the default cap should yield to a user-set
compression.threshold/model_thresholds(#117915,needs-decision; #117929 is the leading fix). Docs already describenullas the ratio-only opt-out and make no restart claim, so no doc change.