I just came across a fun question asking why does HTML think 'chucknorris' is a color on StackOverflow. This is called "Flex Hex" and is described in detail in a little rant about Microsoft Internet Explorer's color parsing.
It turns out these are all valid colors:
chucknorris sick crap
I thought it would be fun to support parsing colors in this (arguably broken) format in Factor. The algorithm basically breaks down into three parts:
Convert non-hexadecimal digits to zero.
: hex-only ( str -- str' ) [ dup hex-digit? [ drop CHAR: 0 ] unless ] map ;
Group into three equal groups, padding on the right with zero if necessary.
: pad-length ( str -- n ) length dup 3 mod [ 3 swap - + ] unless-zero ; : three-groups ( str -- array ) dup pad-length [ CHAR: 0 pad-tail ] [ 3 / group ] bi ;
Convert each segment into a two-digit hexadecimal value, shortening each segment first to eight chars from the right, padding on the left if only one character.
: hex-rgb ( array -- array' ) [ 8 short tail* 2 short head 2 CHAR: 0 pad-head ] map ;
Putting that together, we have this word to parse "flex hex" colors (removing hash-marks from the left if present):
: flex-hex ( str -- hex ) "#" ?head drop hex-only three-groups hex-rgb "" join ;
And, of course, some tests to verify that we handle lots of different cases:
{ "00b000" } [ "#zqbttv" flex-hex ] unit-test { "0f0000" } [ "f" flex-hex ] unit-test { "000f00" } [ "0f" flex-hex ] unit-test { "000f00" } [ "0f0" flex-hex ] unit-test { "0f0f00" } [ "0f0f" flex-hex ] unit-test { "0ff000" } [ "0f0f0f0" flex-hex ] unit-test { "ad0e0e" } [ "adamlevine" flex-hex ] unit-test { "000000" } [ "MrT" flex-hex ] unit-test { "00c000" } [ "sick" flex-hex ] unit-test { "c0a000" } [ "crap" flex-hex ] unit-test { "c00000" } [ "chucknorris" flex-hex ] unit-test { "6ecde0" } [ "6db6ec49efd278cd0bc92d1e5e072d68" flex-hex ] unit-test
The code for this is on my GitHub.
No comments:
Post a Comment