Propose more sass:math module functions #2779
Conversation
nex3
left a comment
nex3
left a comment
There was a problem hiding this comment.
This is a really good start!
Awjin
changed the base branch from
master
to
proposal.forward-with.draft-1.1
Awjin
changed the base branch from
proposal.forward-with.draft-1.1
to
master
proposal/more-math-functions.md
Outdated
| hypot($numbers...) | ||
| ``` | ||
|
|
||
| * Throw an error if any argument is not a [length][]. |
There was a problem hiding this comment.
If we don't allow unitless numbers here, I think that makes it the only Sass function that requires units. That's probably more confusing of an inconsistency than we'd have if we were more lenient than plain CSS and allowed unitless numbers here.
We could even go so far as to allow any compatible units here, since the output unit is always well-defined. I'm tempted to do so, because (unlike CSS) Sass allows custom units for which vector-length calculations could make sense.
There was a problem hiding this comment.
Yeah, makes sense to me. I'll have hypot() accept any arguments that are all compatible with each other, and return a value that takes its unit from the leftmost argument.
|
|
||
| * Throw an error if any argument is not a [length][]. | ||
| * Return `Infinity` if any argument is `Infinity`. | ||
| * Return the square root of the sum of the squares of each argument. |
There was a problem hiding this comment.
All of these functions should define the units the return value uses, even if it's just "as a unitless number".
proposal/more-math-functions.md
Outdated
| * Return `-Infinity` if `$number == 0`. | ||
| * Return `Infinity` if `$number == Infinity`. | ||
| * Return the [natural log][] of `$number`. | ||
| * Otherwise: |
There was a problem hiding this comment.
Style nit:
* Otherwise, if `$base < 0` or `$base == 0` or `$base == 1`, throw an error.
* Otherwise, return the natural log of `$number` divided by the natural log of `$base`.Writing "Otherwise" for each branch of the conditional helps draw the reader's attention to the fact that there are three possible cases here.
proposal/more-math-functions.md
Outdated
| ``` | ||
|
|
||
| * Throw an error if `$number` has units or `$number < 0`. | ||
| * If `$base == null`: |
There was a problem hiding this comment.
Style nit for consistency with existing specs:
| * If `$base == null`: | |
| * If `$base` is null: |
proposal/more-math-functions.md
Outdated
| * Return `1`. | ||
|
|
||
| * Otherwise, if `$exponent == Infinity`: | ||
| * Return `NaN` if the `absolute value of $base == 1`. |
There was a problem hiding this comment.
"Absolute value of" isn't really syntax, so it reads strange to have it code-highlighted... I'd probably either write this in prose:
| * Return `NaN` if the `absolute value of $base == 1`. | |
| * Return `NaN` if the absolute value of `$base` equals `1`. |
or avoid mentioning absolute value altogether:
| * Return `NaN` if the `absolute value of $base == 1`. | |
| * Return `NaN` if `$base == 1` or `$base == -1`. |
There was a problem hiding this comment.
Got it. I'll go with the latter.
proposal/more-math-functions.md
Outdated
| unitless, it is assumed to be in radians. | ||
| * Return `NaN` if `$number == Infinity`. | ||
| * Return `-0` if `$number == -0`. | ||
| * Return `Infinity` if `$number == 90 +/- (360 * n)`, where `n` is any |
There was a problem hiding this comment.
| * Return `Infinity` if `$number == 90 +/- (360 * n)`, where `n` is any | |
| * Return `Infinity` if `$number % 360deg == 90deg`. |
It's important to specify units here, since the values 90 and 360 don't mean anything on their own if $number is in radians.
proposal/more-math-functions.md
Outdated
|
|
||
| > `atan2($y, $x)` is distinct from `atan($y / $x)` because it preserves the | ||
| > quadrant of the point in question. For example, `atan2(1, -1)` corresponds to | ||
| > the point `(-1, 1)` and returns `(0.75 * pi) rad`. In contrast, `atan(1 / -1)` |
There was a problem hiding this comment.
Consider making these radian examples valid SassScript:
| > the point `(-1, 1)` and returns `(0.75 * pi) rad`. In contrast, `atan(1 / -1)` | |
| > the point `(-1, 1)` and returns `0.75rad * $pi`. In contrast, `atan(1 / -1)` |
proposal/more-math-functions.md
Outdated
| > `atan2($y, $x)` is distinct from `atan($y / $x)` because it preserves the | ||
| > quadrant of the point in question. For example, `atan2(1, -1)` corresponds to | ||
| > the point `(-1, 1)` and returns `(0.75 * pi) rad`. In contrast, `atan(1 / -1)` | ||
| > and `atan(-1 / 1)` resolve first to `atan(-1)`, so both return (`-0.25 * pi) | ||
| > rad`. |
There was a problem hiding this comment.
This is a really good explanation 👍.
proposal/more-math-functions.md
Outdated
| * If all arguments are not compatible with each other, throw an error. | ||
| * Let the unit of the return value be the same as that of the leftmost argument | ||
| that has a unit. If all arguments are unitless, let the return value be | ||
| unitless. |
There was a problem hiding this comment.
What do you think about forbidding mixing unitless numbers with numbers that have units? The lack of transitivity there is weird, and while it's probably too late to change it for + and -, we could definitely change it here.
There was a problem hiding this comment.
The only reason I considered allowing unitless numbers was to model the behavior after +/- 🙂 But if the lack of transitivity is undesired, I'm good with forbidding mixing unitless/units.
|
Thanks for the reviews! |
2 participants
First draft for specifying Sassified versions of all the mathematical functions in the Values and Units 4 Draft, logarithms, and $e and $pi variables.
See #851 for context.