csscalc always return false if usePrefixes is false

[zenatti]

csscalc test always return false if I set usePrefixes false on modernizr 3.3.1

"el.style.cssText = prop + prefixes.join(value + prop);" doesn't check if Modernizr._config.usePrefixes is true or false.

[patrickkettner]

I am not not able to replicate this

can you give a bit more information?

[zenatti]

I have a build with usePrefixes:false

http://jsbin.com/leceyiyoti/edit?html,js,output

[patrickkettner]

ah, im very sorry. I messed up the config when I made the bin :(

the issue is that we are joining the prefixes to get cssText, which is not working when prefixes is an empty array.

@ryanseddon @SlexAxton a hack fix would be to change prefixes from [] to ['','']. Otherwise we will need to reformat a handful of detects to check that config value and react accordingly, or change them from something like

var prop = 'width:';
var value = 'calc(10px);';
var el = createElement('a');

el.style.cssText = prop + prefixes.join(value + prop);

to something like

var prop = 'width:';
var value = 'calc(10px);';
var el = createElement('a');

el.style.cssText = prop + value + prefixes.join(value + prop);

the ['','']is the smallest amount of code, so I am leaning towards that. it just feels a bit dirty.

thoughts?

[SlexAxton]

I don't see a huge need to byteshave here. I'd be happy to to just check for the length of 'prefixes' and do something different in the zero case. But I'm not against ['',''] if you feel strongly.

[patrickkettner]

its a bit of a yak shaving moment, but the main reason I think about it is we would have to do the same check in every detect that uses prefixes.join method (about a dozen or so), or just short circuit the whole issue with a few chars.

if neither you or @ryanseddon are against the ['',''] way, i'll probably go that path with a comment

[ryanseddon]

Nope not against it.

[SlexAxton]

👍

On Tue, Jan 26, 2016 at 7:04 PM Ryan Seddon notifications@github.com
wrote:

Nope not against it.

—
Reply to this email directly or view it on GitHub
#1852 (comment)
.

[patrickkettner]

@zenatti fixed in master, should be released as 3.4.0 in a few days

[zenatti]

Thanks @patrickkettner !!