Retry on failure

[garrettmflynn]

fix #613

This was suggested when merging #612

[CodyCBakerPhD]

OK, well the action is working: https://github.com/NeurodataWithoutBorders/nwb-guide/actions/runs/7960795322/job/21737344076?pr=617#step:12:575 -> https://github.com/NeurodataWithoutBorders/nwb-guide/actions/runs/7960795322/job/21737344076?pr=617#step:12:1138

But not helping alleviate the underlying problem...

[garrettmflynn]

That specific run has an issue because of No module named 'flask'. Seems like the create and activate environment step didn't run.

I've never noticed this become a problem before. Is there a reason this was excluded?

[CodyCBakerPhD]

Hmm... Could it be that the action isn't running in the same shell? (one in which the conda environment is not active?)

Adding a printout of the installed packages to be sure; the cache was restored in the logs I linked so you'd hope so

[garrettmflynn]

Ah looks like that's the case. Just updated to the same shell.

Fingers crossed!

[garrettmflynn]

Hmm all attempts here don't seem to help. According to the docs, this should default to bash on Mac and Linux—and otherwise it seems likely that it'd use the default set at the top of the workflow.

[CodyCBakerPhD]

Right, but specifying it as bash would be a fresh shell I believe, not the currently active conda environment; maybe add a conda activate nwb-guide as first step of the run command in the retry?

[CodyCBakerPhD]

Tabling for now to see if we actually need this - #621 has been working well so far

[CodyCBakerPhD]

Yeah I don't think we actually need this any more