clarkzinzow commented on issue #26685: URL: https://github.com/apache/arrow/issues/26685#issuecomment-1703199976
Hi @jorisvandenbossche, apologies for the delays on my end! The application-level workaround for Ray ended up sufficing, so I never got back to submitting PR, and I must have missed these notifications. I have an out-of-date branch that works e2e for option (1), where the buffer traversal + truncation is shared by both the IPC serialization and pickle paths: a `BufferAggregator` interface is introduced with an IPC implementation and an implementation that accumulates buffers into a payload (similar to the C data interface) that can be serialized with language-specific schemes by Arrow front-ends (e.g. pickle via Python). But this still has a few TODOs: 1. Rebase onto latest master and resolve any pop-up conflicts. 2. Move `BufferAggregator`, `ArrayBufferPayload`, and `ArraySerializerBufferAggregator` definitions/implementations out of the IPC code; these could be top-level, under `arrow/util`, under a new directory, etc. 3. Misc. cleanup. cc @anjakefala https://github.com/apache/arrow/compare/main...clarkzinzow:arrow:arrow-10739-pickle-buffer-truncation-fix -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
