@gscdump/lakehouse 0.36.2 → 0.36.4

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
@@ -106,13 +106,13 @@ async function connectIcebergCatalog(config, opts = {}) {
106
106
  await cachePut(opts.cache, catalogConfigKey(config), toCache, CATALOG_CONFIG_TTL_MS, now);
107
107
  }
108
108
  }
109
- const resolver = cachingResolver(s3SignedResolver({
109
+ const resolver = withVerifiedWriterByteLengths(cachingResolver(s3SignedResolver({
110
110
  accessKeyId: config.s3.accessKeyId,
111
111
  secretAccessKey: config.s3.secretAccessKey,
112
112
  region: config.s3.region ?? "auto",
113
113
  endpoint: config.s3.endpoint,
114
114
  pathStyle: true
115
- }));
115
+ })));
116
116
  return {
117
117
  catalog,
118
118
  resolver,
@@ -120,6 +120,34 @@ async function connectIcebergCatalog(config, opts = {}) {
120
120
  cacheScope: catalogCacheScope(config)
121
121
  };
122
122
  }
123
+ function withVerifiedWriterByteLengths(resolver) {
124
+ if (!resolver.writer) return resolver;
125
+ const baseWriter = resolver.writer;
126
+ return {
127
+ ...resolver,
128
+ writer(path, options) {
129
+ const writer = baseWriter(path, options);
130
+ const finish = writer.finish.bind(writer);
131
+ writer.finish = async function() {
132
+ await finish();
133
+ const actual = bufferedByteLength(writer);
134
+ if (actual == null || writer.offset === actual) return;
135
+ console.warn(`[lakehouse] corrected Iceberg writer byte length for ${path}: offset=${writer.offset} actual=${actual}`);
136
+ writer.offset = actual;
137
+ };
138
+ return writer;
139
+ }
140
+ };
141
+ }
142
+ function bufferedByteLength(writer) {
143
+ if (typeof writer.flush === "function") return null;
144
+ try {
145
+ const bytes = writer.getBytes();
146
+ return Number.isFinite(bytes.byteLength) ? bytes.byteLength : null;
147
+ } catch {
148
+ return null;
149
+ }
150
+ }
123
151
  function catalogCacheScope(config) {
124
152
  return `${config.catalogUri}\0${config.warehouse}`;
125
153
  }
package/package.json CHANGED
@@ -1,7 +1,7 @@
1
1
  {
2
2
  "name": "@gscdump/lakehouse",
3
3
  "type": "module",
4
- "version": "0.36.2",
4
+ "version": "0.36.4",
5
5
  "description": "Dataset-agnostic Iceberg lakehouse layer + dataset registry for R2 Data Catalog producers (ADR-0021).",
6
6
  "author": {
7
7
  "name": "Harlan Wilton",