diff options
Diffstat (limited to 'modules')
| -rw-r--r-- | modules/core/core-bundle.js | 47 | ||||
| -rw-r--r-- | modules/core/module-loader.js | 2 | ||||
| -rw-r--r-- | modules/core/module-predictor.js | 15 | ||||
| -rw-r--r-- | modules/core/virtual-fs.js | 32 | ||||
| -rw-r--r-- | modules/test/test-predictor-accuracy.js | 33 | ||||
| -rw-r--r-- | modules/test/test-vfs-checkpoint.js | 55 |
6 files changed, 169 insertions, 15 deletions
diff --git a/modules/core/core-bundle.js b/modules/core/core-bundle.js index cd510b7..84d194d 100644 --- a/modules/core/core-bundle.js +++ b/modules/core/core-bundle.js @@ -942,6 +942,17 @@ class VirtualFileSystem { } checkpoint(name) { if (!name) name = new Date().toISOString(); + name = String(name).trim(); + if (!name || name.indexOf('/') !== -1) { + throw new Error('checkpoint name must be non-empty and must not contain "/"'); + } + // Two unnamed calls can land in the same millisecond; without this the second silently + // overwrites the first (measured: 6 calls -> 3 distinct names). + if (this.listCheckpoints().indexOf(name) !== -1) { + let n = 2; + while (this.listCheckpoints().indexOf(name + '-' + n) !== -1) n++; + name = name + '-' + n; + } const createdAt = new Date().toISOString(); // Strip system dirs to prevent fractal bloat const rootClone = JSON.parse(JSON.stringify(this.root)); @@ -961,12 +972,15 @@ class VirtualFileSystem { } this.fsWriteJSON('/_snapshots/checkpoint-' + name + '.json', data); this.saveToStorage(); - // Auto-manage: keep max 3 checkpoints, delete oldest + // Auto-manage: keep the 3 NEWEST checkpoints — by createdAt, not by name, so a custom + // name ("aaa-oldest") cannot pin itself forever while newer snapshots are deleted. const all = this.listCheckpoints(); if (all.length > 3) { - const sorted = all.slice().sort(); - const toDelete = sorted.slice(0, all.length - 3); - for (const old of toDelete) this.deleteCheckpoint(old); + const aged = all.map(n => { + const snap = this.fsReadJSON('/_snapshots/checkpoint-' + n + '.json'); + return { name: n, at: (snap && snap.createdAt) || '' }; + }).sort((a, b) => (a.at < b.at ? -1 : (a.at > b.at ? 1 : 0))); + for (const old of aged.slice(0, aged.length - 3)) this.deleteCheckpoint(old.name); } return name; } @@ -976,8 +990,18 @@ class VirtualFileSystem { console.error('Rollback failed: checkpoint not found or invalid'); return false; } + // The snapshot deliberately omits scaffolding (bloat guard at :944-950). Restoring the + // clone verbatim deletes every checkpoint, every agent session and the trash, so keep the + // live copies and re-attach them after the root swap. + const keep = {}; + for (const dir of ['_snapshots', '_sessions', '.trash']) { + const node = this.root && this.root.content ? this.root.content[dir] : null; + if (node) keep[dir] = node; + } this.root = raw.root; this.currentPath = raw.currentPath || []; + if (!this.root.content) this.root.content = {}; + for (const dir of Object.keys(keep)) this.root.content[dir] = keep[dir]; this.saveToStorage(); return true; } @@ -18261,6 +18285,7 @@ class PredictorModule { return this._dataPoints.filter(dp => { if (!dp.observedDuration) return false; // outcomes only, never predictions if (dp.usable === false) return false; // aborted runs are not evidence + if (!(dp.observedDuration >= 1.5)) return false; // belt and braces: never evidence if (dp.features) return this._similarity(features, dp.features) > 0.3; return false; }); @@ -18412,7 +18437,7 @@ class PredictorModule { // ─── Commands ─── showStatus() { - const actualExecutions = this._dataPoints.filter(dp => dp.observedDuration).length; + const actualExecutions = this._dataPoints.filter(dp => dp.observedDuration && dp.usable !== false).length; const paired = this._accuracy.total; const unpaired = this._dataPoints.filter(dp => dp.paired === false).length; const accPct = paired > 0 ? ((this._accuracy.correct / paired) * 100).toFixed(0) + '%' : 'n/a'; @@ -18517,8 +18542,16 @@ class PredictorModule { const data = fs.fsReadJSON('/system/metrics/predictor.json'); if (!data) return; - // Prediction-only points are noise for similarity; keep outcomes. - this._dataPoints = (data.dataPoints || []).filter(dp => dp && dp.observedDuration); + // Legacy (pre-v2) points carry no `usable`. Backfill it from the measurement — an + // aborted run must not enter similarity as evidence (measured: a 0.079 s point + // dragged avgDuration to ~0 and floored every prediction at 1 s). + this._dataPoints = (data.dataPoints || []) + .filter(dp => dp && dp.observedDuration) + .map(dp => { + if (typeof dp.usable !== 'boolean') dp.usable = dp.observedDuration >= 1.5; + return dp; + }) + .filter(dp => dp.usable !== false); this._predictions = data.predictions || 0; if (data.v === this._SCHEMA) { diff --git a/modules/core/module-loader.js b/modules/core/module-loader.js index b278f51..18137ec 100644 --- a/modules/core/module-loader.js +++ b/modules/core/module-loader.js @@ -4,7 +4,7 @@ // Core module bundle loaded synchronously in main.html. // This loader only handles version cache-busting, splash, terminal init, and optional ext modules. - const LATEST_VER = 'v4.6.102'; + const LATEST_VER = 'v4.6.103'; const CACHE = '?v=' + LATEST_VER; const stored = localStorage.getItem('webx11_cache_ver'); if (stored !== LATEST_VER) { diff --git a/modules/core/module-predictor.js b/modules/core/module-predictor.js index aebda87..b534d34 100644 --- a/modules/core/module-predictor.js +++ b/modules/core/module-predictor.js @@ -118,6 +118,7 @@ class PredictorModule { return this._dataPoints.filter(dp => { if (!dp.observedDuration) return false; // outcomes only, never predictions if (dp.usable === false) return false; // aborted runs are not evidence + if (!(dp.observedDuration >= 1.5)) return false; // belt and braces: never evidence if (dp.features) return this._similarity(features, dp.features) > 0.3; return false; }); @@ -269,7 +270,7 @@ class PredictorModule { // ─── Commands ─── showStatus() { - const actualExecutions = this._dataPoints.filter(dp => dp.observedDuration).length; + const actualExecutions = this._dataPoints.filter(dp => dp.observedDuration && dp.usable !== false).length; const paired = this._accuracy.total; const unpaired = this._dataPoints.filter(dp => dp.paired === false).length; const accPct = paired > 0 ? ((this._accuracy.correct / paired) * 100).toFixed(0) + '%' : 'n/a'; @@ -374,8 +375,16 @@ class PredictorModule { const data = fs.fsReadJSON('/system/metrics/predictor.json'); if (!data) return; - // Prediction-only points are noise for similarity; keep outcomes. - this._dataPoints = (data.dataPoints || []).filter(dp => dp && dp.observedDuration); + // Legacy (pre-v2) points carry no `usable`. Backfill it from the measurement — an + // aborted run must not enter similarity as evidence (measured: a 0.079 s point + // dragged avgDuration to ~0 and floored every prediction at 1 s). + this._dataPoints = (data.dataPoints || []) + .filter(dp => dp && dp.observedDuration) + .map(dp => { + if (typeof dp.usable !== 'boolean') dp.usable = dp.observedDuration >= 1.5; + return dp; + }) + .filter(dp => dp.usable !== false); this._predictions = data.predictions || 0; if (data.v === this._SCHEMA) { diff --git a/modules/core/virtual-fs.js b/modules/core/virtual-fs.js index 09b3548..3ed1b40 100644 --- a/modules/core/virtual-fs.js +++ b/modules/core/virtual-fs.js @@ -940,6 +940,17 @@ class VirtualFileSystem { } checkpoint(name) { if (!name) name = new Date().toISOString(); + name = String(name).trim(); + if (!name || name.indexOf('/') !== -1) { + throw new Error('checkpoint name must be non-empty and must not contain "/"'); + } + // Two unnamed calls can land in the same millisecond; without this the second silently + // overwrites the first (measured: 6 calls -> 3 distinct names). + if (this.listCheckpoints().indexOf(name) !== -1) { + let n = 2; + while (this.listCheckpoints().indexOf(name + '-' + n) !== -1) n++; + name = name + '-' + n; + } const createdAt = new Date().toISOString(); // Strip system dirs to prevent fractal bloat const rootClone = JSON.parse(JSON.stringify(this.root)); @@ -959,12 +970,15 @@ class VirtualFileSystem { } this.fsWriteJSON('/_snapshots/checkpoint-' + name + '.json', data); this.saveToStorage(); - // Auto-manage: keep max 3 checkpoints, delete oldest + // Auto-manage: keep the 3 NEWEST checkpoints — by createdAt, not by name, so a custom + // name ("aaa-oldest") cannot pin itself forever while newer snapshots are deleted. const all = this.listCheckpoints(); if (all.length > 3) { - const sorted = all.slice().sort(); - const toDelete = sorted.slice(0, all.length - 3); - for (const old of toDelete) this.deleteCheckpoint(old); + const aged = all.map(n => { + const snap = this.fsReadJSON('/_snapshots/checkpoint-' + n + '.json'); + return { name: n, at: (snap && snap.createdAt) || '' }; + }).sort((a, b) => (a.at < b.at ? -1 : (a.at > b.at ? 1 : 0))); + for (const old of aged.slice(0, aged.length - 3)) this.deleteCheckpoint(old.name); } return name; } @@ -974,8 +988,18 @@ class VirtualFileSystem { console.error('Rollback failed: checkpoint not found or invalid'); return false; } + // The snapshot deliberately omits scaffolding (bloat guard at :944-950). Restoring the + // clone verbatim deletes every checkpoint, every agent session and the trash, so keep the + // live copies and re-attach them after the root swap. + const keep = {}; + for (const dir of ['_snapshots', '_sessions', '.trash']) { + const node = this.root && this.root.content ? this.root.content[dir] : null; + if (node) keep[dir] = node; + } this.root = raw.root; this.currentPath = raw.currentPath || []; + if (!this.root.content) this.root.content = {}; + for (const dir of Object.keys(keep)) this.root.content[dir] = keep[dir]; this.saveToStorage(); return true; } diff --git a/modules/test/test-predictor-accuracy.js b/modules/test/test-predictor-accuracy.js index 3688b03..74f0fcc 100644 --- a/modules/test/test-predictor-accuracy.js +++ b/modules/test/test-predictor-accuracy.js @@ -247,5 +247,38 @@ function ok(name, cond, detail) { ok('S3 pending consumed', p._pending.length === 0); } +// T. v1 payload carrying an aborted garbage point must not train similarity (2026-10-27) +{ + const { p, term, out } = mk(); + const feats = p._extractFeatures('read the logs'); + term.fs.fsWriteJSON('/system/metrics/predictor.json', { + dataPoints: [{ features: feats, observedDuration: 0.079, observedTurns: 1 }, + { features: feats, observedDuration: 30, observedTurns: 3 }], + predictions: 2 + }); + p._loadFromVFS(); + ok('T1 legacy aborted point dropped on load', p._dataPoints.length === 1, 'n=' + p._dataPoints.length); + const similar = p._findSimilar(p._extractFeatures('read the logs')); + ok('T2 similarity ignores the aborted point', + similar.length === 1 && similar[0].observedDuration === 30, + JSON.stringify(similar.map(s => s.observedDuration))); + out.length = 0; p.showStatus(); + const line = out.find(l => l.includes('Actual executions')) || ''; + ok('T3 status counts only usable executions', /Actual executions:\s*1\s*$/.test(line.trim()), line.trim()); +} + +// U. a lone aborted point cannot floor the next prediction +{ + const { p, term } = mk(); + term.fs.fsWriteJSON('/system/metrics/predictor.json', { + dataPoints: [{ features: { hasRead: true }, observedDuration: 0.079, observedTurns: 1 }], + predictions: 1 + }); + p._loadFromVFS(); + const pred = p.notePrediction('read the logs', 'test'); + ok('U1 prediction is not floored by a lone aborted point', pred && pred.floored === false, + JSON.stringify(pred && { duration: pred.duration, floored: pred.floored })); +} + console.log(passes + '/' + (passes + fails) + ' PASS'); process.exit(fails ? 1 : 0); diff --git a/modules/test/test-vfs-checkpoint.js b/modules/test/test-vfs-checkpoint.js index 03a8c8a..65f32dc 100644 --- a/modules/test/test-vfs-checkpoint.js +++ b/modules/test/test-vfs-checkpoint.js @@ -84,6 +84,61 @@ function ok(name, cond, detail) { 'output=' + JSON.stringify(out.join('\n'))); } +// D. same-millisecond calls must not overwrite each other +{ + const fs = mkFs(); + const names = []; + // 4 calls: one more than the cap of 3, so every name is still stored when the next one is + // minted (at 6+, prune legitimately frees a name and it may be reused). + for (let i = 0; i < 4; i++) names.push(fs.checkpoint('')); + ok('D1 four unnamed checkpoints are four distinct names', new Set(names).size === 4, + 'names=' + JSON.stringify(names)); + ok('D2 prune still caps the store at three', fs.listCheckpoints().length === 3, + 'stored=' + JSON.stringify(fs.listCheckpoints())); +} + +// E. prune keeps the NEWEST three, not the alphabetically last +{ + const fs = mkFs(); + fs.checkpoint('aaa-oldest'); + fs.checkpoint(''); + fs.checkpoint(''); + fs.checkpoint(''); // 4th -> prune + const left = fs.listCheckpoints(); + ok('E1 oldest custom-named checkpoint was pruned', !left.includes('aaa-oldest'), 'left=' + JSON.stringify(left)); + ok('E2 three checkpoints remain', left.length === 3, 'left=' + JSON.stringify(left)); + ok('E3 rollback still works after pruning', fs.rollback(left[0]) === true); +} + +// F. rollback must not destroy scaffolding +{ + const fs = mkFs(); + fs.mkdirRecursive('/_sessions/agent1'); + fs.writeFile('/_sessions/agent1/log.json', '{"turns":3}'); + fs.mkdirRecursive('/.trash'); + fs.writeFile('/.trash/junk.txt', 'x'); + fs.writeFile('/work.txt', 'before'); + const name = fs.checkpoint('keep-me'); + fs.writeFile('/work.txt', 'after'); + ok('F1 rollback succeeds', fs.rollback(name) === true); + ok('F2 file restored', fs.readFile('/work.txt') === 'before', 'got ' + fs.readFile('/work.txt')); + ok('F3 checkpoint survived rollback', fs.listCheckpoints().includes('keep-me'), + 'list=' + JSON.stringify(fs.listCheckpoints())); + ok('F4 sessions survived rollback', !!fs.readDir('/_sessions')); + ok('F5 trash survived rollback', !!fs.readDir('/.trash')); + ok('F6 rollback is repeatable', fs.rollback('keep-me') === true); +} + +// G. a name with a slash is refused, not silently unlistable +{ + const fs = mkFs(); + let threw = false; + try { fs.checkpoint('a/b'); } catch (e) { threw = true; } + ok('G1 slash in name throws', threw); + ok('G2 nothing was written', fs.listCheckpoints().length === 0, + 'list=' + JSON.stringify(fs.listCheckpoints())); +} + if (fails === 0) { console.log('test-vfs-checkpoint: ' + passes + ' passed, 0 failed'); } else { |