make JULIA_DEPOT_PATH work like JULIA_LOAD_PATH?

[StefanKarpinski]

I'm writing long-overdue docs for how JULIA_LOAD_PATH is expanded to populate LOAD_PATH (the current docs are woefully inaccurate at this point) and I noticed that JULIA_DEPOT_PATH doesn't do the same "empty path means insert the default" thing that JULIA_LOAD_PATH does. I really think it should basically for the reasons given here. All the same arguments apply to JULIA_DEPOT_PATH. Technically this is a breaking change, but I think it's a good change and I strongly doubt anyone is relying on the current behavior since most people don't seem to understand what the depot path is or do anything to change it in the first place.

[mgkuhn]

I agree. The behavior that an empty entry after splitting on the separator is expanded to the default value is already well established in other tools that take search paths in environment variables, most notably TeX's TEXINPUT variable, and has withstood the test of time.

Regarding the alternative interpretation of an empty entry as the current working directory, this POSIX sh behavior is actually discouraged from being used by the POSIX standard: “A zero-length prefix is a legacy feature that indicates the current working directory. It appears as two adjacent characters ( "::" ), as an initial preceding the rest of the list, or as a trailing following the rest of the list. A strictly conforming application shall use an actual pathname (such as .) to represent the current working directory in PATH.”

[StefanKarpinski]

All good points. I was not aware of TEXINPUT using the same convention but it makes sense as a convention. I’m inclined to go ahead with this and see if it breaks PkgEval when we tag 1.2-rc1 or if anyone complains. My guess is no on both counts. Will wait for assent from triage.

[StefanKarpinski]

Closed by #31009.