[14:42:36] <Krinkle>	 RoanKattouw: I haven't confirmed, but this might also fix the calc() issue you found via Codex. https://gerrit.wikimedia.org/r/c/mediawiki/libs/less.php/+/896212
[17:07:58] <RoanKattouw>	 Krinkle: hmm I'll try
[17:08:16] <RoanKattouw>	 > LESS should only handle math when the input uses PHP variables or literals, and thus effectively has a constant outcome.
[17:09:06] <RoanKattouw>	 I would add to that 1) "and the units are the same" and/or 2) "the calculation isn't wrapped inside calc()"
[17:10:05] <RoanKattouw>	 Right now if I write padding: calc( @foo + @bar ); , where @foo is 2em and @bar is 1px, less.php will output padding: calc( 3em );
[17:11:01] <RoanKattouw>	 That's wrong because adding up px and em doesn't work that way, and also because if I used calc() I clearly intended for the calculation to happen at runtime in the browser 
[17:11:28] <RoanKattouw>	 Newer versions of Less handle this correctly but I'm not sure if upstream did at the version we're targeting (2.5.3)
[17:12:01] <RoanKattouw>	 So I'll look into this today
[18:25:36] <Krinkle>	 It didn't back then http://ecomfe.github.io/est/fiddle/#version=2.5.3&autoprefix=false&est=false&autorun=false
[18:26:43] <Krinkle>	 And 3.0 excludes calc more generally as I'm doing
[18:28:31] <Krinkle>	 Outside calc, 3.8 still just uses left hand side unit
[18:29:57] <Krinkle>	 Preserving non-calc 3px - 5% as CSS would produce invalid syntax.
[22:15:05] <RoanKattouw>	 Yeah that makes sense
[22:15:34] <RoanKattouw>	 If your patch excludes calc generally that would be great! I'll review it today and also add a test case for the non-var() case