Alan King

We have determined that the only change(s) required to support pam_interactive are in `iinit` at this point. The plugin can be released separately from the server as it is polished,...

A set of requirements has formed... I will number them for easy reference in any further discussion. 1. Introduce 1 or 2 new API plugins to mimic `truncate` and `ftruncate`....

This may require implementing a new resource plugin operation for `ftruncate`. The unixfilesystem resource implementation of the truncate resource plugin operation literally just calls `truncate`. In order to support `ftruncate`...

Do we expect the new truncate APIs to behave like `truncate` and `ftruncate` in the way that they handle output and error codes? Historically, when an iRODS API returns an...

Mostly, I was trying to conform to the behavior of `truncate` and `ftruncate`. I realize this isn't entirely possible due to the fact that there's no global `errno` concept in...

Okay, well, with that, now I need to confirm some other things before I write anything else. Here was my plan for the function signature, cuing off the API plugin...

> I like the name, `rc_replica_truncate`. It's clear it's about a single replica. Although, as I type this, is the use of `rcDataObj` incorrect? _Just wondering._ The existing API is...

I think so... I'll write back here if I think of anything else.

Okay, so, through some experimentation with an initial implementation, it seems we may not be able to avoid changes to the server in order for this to work the way...

We have determined that the way voting for write works is a flawed design, and we will need to change it per #7476. As such, truncate can vote on write...