Vyas Ramasubramani
Vyas Ramasubramani
I think the open question is how should this behave with respect to the suffix? For consistency with the current behavior I would expect the return to the original directory...
@csadorf sorry if I wasn't clear enough, I don't think I properly stated my position. I am agreeing with pretty much everything that you're saying, I just wanted to make...
In that case the logic for this issue wouldn't just go in get_prefix, right? You need it to wrap the entire command, including the `cmd_suffix`.
> This means that in the majority of cases, API quality and implementation clarity should be correlated with each other. I agree with that statement when starting from a blank...
I can explain why this is probably happening, but I'm not immediately sure how to fix it (Sphinx at this level of complexity is always a bit of trial and...
@kidrahahjo no, the parallelization will speed up the call to `_fetch_status`, but the code that I'm referring to here is run after `_fetch_status`. You can see the difference by running...
Thanks for the feedback! Responses inline: > My understanding is that names like `.flow/config` and `.flow/${FLOWPROJECT_NAME}_bundles` are relative to the `signac.rc` file, not `project.py`. That means that this directory/file structure...
I agree with most of that. I think my question about moving some of the config information into the class definition was made in the hope of simplification without sufficient...
Interesting, could you share what systems they are using it on? If people are actually making use of this feature I'm fine preserving it, but we're almost guaranteed to get...
> Really I think the only thing we could do better is change the logic a bit so that if more than one scheduler executable is detected, we print a...