No check for an unsaved script

[jeffeaton]

A common silly thing I do is to create an orderly task, create a script and do a lot of development in it, but then forget to save it and run the report.

In orderly1, this would throw a "script my-script.R does not exist" error or something like that. In orderly2, this doesn't give me any indication that I've probably done a silly thing.

root1 <- tempfile()
orderly_init(root1)

setwd(root1)
orderly_new("hello", template = FALSE)

## Open the script and develop a lot
## Forgot to save it though!

orderly_run("hello")

## > orderly_run("hello")
## ℹ Starting packet 'hello' `20240704-225610-cc4efc1c` at 2024-07-05 00:56:10.803153
## ✔ Finished running hello.R
## ℹ Finished 20240704-225610-cc4efc1c at 2024-07-05 00:56:10.820007 (0.01685405 secs)
## [1] "20240704-225610-cc4efc1c"
## 
## Oh no! That has just run an empty script
## and committed it into my archive! Oops,
## how silly!

Minor thing, but I missed that old error message saving me today, if there's anything that could be checked to warn me not done what I want.

[richfitz]

I am not sure that there was something we detected previously, unless you had literally nothing in the directory. The idea that orderly reports can be trivial is one of the aims here, so there's not a lot we can warn about.

[jeffeaton]

Yeah, fair. Previously the orderly task would not run if (1) orderly.yml was not saved with a script: <script>.R specified, or (2) the file <script.R> was not found (which in orderly1 was not created by orderly_new()). This saved me a surprising amount of grief; but appreciate that it's just a user silliness and not particularly important thing to flag.

[plietar]

If you are using RStudio we might be able to use the API to detect an unsaved report and show a warning. I don't know how useful or desirable it is though?

[jeffeaton]

Thanks, yeah, I don't know how to use RStudio unfortunately :). I think we can close this out. Pain free deleting from the archive will make this more of a shrug.