Ht 3192 enumbug

[mwarin]

Not a solution to the problem, but a change to make the warnings make more sense.

[mwarin]

offending enumchrons listed in https://tools.lib.umich.edu/jira/browse/HT-3192

[aelkiss]

So this fix makes the warning more readable, but it doesn't prevent the underlying problem with the exception being thrown.

To be clear, I don't think we need to make it so that the offending enumchron is parseable in any meaningful way. However, I think we should try to avoid it throwing the undefined method 'gsub!' for nil:NilClass exception internally. Looking at where that's happening:

holdings-backend/lib/enum_chron_parser.rb

Line 221 in fff71d7

input_str = input_str.gsub!(m[0], "")

it looks like the issue is that input_str is nil at that point. I don't really understand how that would happen just looking at the code (in particular I don't understand how a nil string could match the regexp, and I don't understand how the string could end up being transformed into nil), but I do think it's worth tracing that path with an offending enumchron.

[aelkiss]

The problem is that:

input.gsub!("not_in_string","replacement") 

returns nil if input does not contain not_in_string.

We should change this to either:

input_str = input_str.gsub(m[0], "")

Or, probably my preference at this point given that we're already mutating input_str a bunch:

input_str.gsub!(m[0], "")

There are several other places we're doing the same thing, so we should probably just review all uses of gsub! and make sure we are using it consistently and correctly.

[mwarin]

The problem is that:
...
There are several other places we're doing the same thing, so we should probably just review all uses of gsub! and make sure we are using it consistently and correctly.

Took care of it.

[aelkiss]

LGTM 🌭