Felix Bünemann

The underlying S3 multipart upload works in chunks and allows uploading chunks in arbitrary order, but I don't think that's possible with the ACD resume API, because you have only...

The above info is spread out over multiple posts, so here's a short summary of how resume works: - Add `localId` to the query when creating a new node (not...

Yes, you have to abort the upload at least once. What you could do is abort once after sending the first complete. That way you wouldn't have to re-transfer any...

Yes, you have to upload at least one chunk. I think this isn't a problem, because waiting for smaller files after upload is not needed (the ACD API can already...

> However it occurs to me that I can solve this problem in a simpler way potentially. Letting rclone 'see' nodes in PENDING state should allow rclone to verify that...

```sql CREATE TABLE phrases (phrase tsquery); INSERT INTO phrases (to_tsquery('simple', 'quick red fox')); CREATE INDEX phrases_idx on phrases using RUM (phrase); -- ERROR: Indexing of phrase tsqueries isn't supported yet...

I think a recheck would be perfectly fine. Thanks for looking into it!

I can contribute a German translation, I can make a PR.

Done, see #54. I already translated it two days ago, but forgot to do the PR, because this issue said after 1.0.

For what it's worth thin, unicorn and puma share similar issues, while webrick is not affected, probably because it lacks a request queue. @FooBarWidget Do you consider this a bug,...