Regression: E474 on certain lhs characters in abbreviations

[adigitoleo]

Regression: E474 on certain lhs characters in abbreviations

[adigitoleo]

Notably, :iabbrev _foo bar always works, this is the only non-alphanumeric character that is not affected it seems.

[zeertzjq]

How is this a regression? The behavior is the same in v0.4.0.

[zeertzjq]

I can't do :iabbrev @ii @include even with @ being present in my &iskeyword.

This works if you :set iskeyword+=@-@. This is because @ has a special meaning, as mentioned in :h 'isfname' which is referenced by :h 'iskeyword':

	If the character is '@', all characters where isalpha() returns TRUE
	are included.  Normally these are the characters a to z and A to Z,
	plus accented characters.  To include '@' itself use "@-@".  Examples:
		"@,^a-z"	All alphabetic characters, excluding lower
				case ASCII letters.
		"a-z,A-Z,@-@"	All letters plus the '@' character.

And the behavior matches :h abbreviations:

There are three types of abbreviations:

full-id	  The "full-id" type consists entirely of keyword characters (letters
	  and characters from 'iskeyword' option).  This is the most common
	  abbreviation.

	  Examples: "foo", "g3", "-1"

end-id	  The "end-id" type ends in a keyword character, but all the other
	  characters are not keyword characters.

	  Examples: "#i", "..f", "$/7"

non-id	  The "non-id" type ends in a non-keyword character, the other
	  characters may be of any type, excluding space and tab.

	  Examples: "def#", "4/7$"

So this is working as intended.

[adigitoleo]

How is this a regression? The behavior is the same in v0.4.0.

Hmm, perhaps. I've had an :iabbrev :foo bar type of thing in my config for a while and this only came up recently (last couple of weeks). I don't mind, I just changed them to :iabbrev _foo bar, but that made me think it was a regression. It's possible this was caused by some more complex interaction.