Repository navigation
fix: load the sample dataset via arrayBuffer under storage pressure - #32
Merged
Merged
Conversation
response.blob() fails with "TypeError: Failed to fetch" once cached models have filled the origin's storage quota, because Chromium may back a large fetched blob with disk. Reading the body into an ArrayBuffer works in the same state. Reproduced in Brave with Gemma 4 E2B, Llama 3.2 3B and the embeddings model cached (5.12 GB used).
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
…ge-pressure # Conflicts: # web_interface/index.html # web_interface/sw.js
This branch was successfully deployed
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Bug: with large models cached, Load sample fails with
TypeError: Failed to fetch. The fetch itself succeeds (HTTP 200, and the existing retry loop never triggers), butresponse.blob()throws, presumably because Chromium backs a large fetched blob with disk and the storage quota is already used up.Repro (Brave): cache Gemma 4 E2B + Llama 3.2 3B + embeddings (5.12 GB used, 2.1 GB reported quota), then click Load sample. It failed six times in a row. On the same response,
response.arrayBuffer()succeeds andnew File([buffer])parses all 506 documents.Fix: read the body with
arrayBuffer()and build theFilefrom it. This was the only fetched-blob read in the app.Verification
npm run checkpassesThis will come up more now that the default model is 2.84 GB.