[t2CharStringPen] add "roundTolerance" option #804
Conversation
…argument Rounding of coordinates is now disabled by default. The pen now accepts an optional 'roundTolerance' float, with values between 0 and 1 (default is 0). Values >= 0.5 mean round all coordinates to integers. Values between 0 and 0.5 mean round only when the absolute difference between the original float and the rounded integer is within the tolerance Fixes fonttools#769
|
I think the default should remain rounding behavior and that the new behavior should be explicitly turned on. |
|
Yeah, initially I also thought it'd better to keep the current default, but then got carried away :) Shall I assert the tolerance is comprised between 0 and 0.5? Values between 0.5 and 1.0 don't make much sense.. |
anthrotype
changed the title
…Tolerance hmtx only can have integer values, so it would be weird to have widths as float in CFF...
abs ends up calling fabs for floats anyway, and is a bit faster if 'number' is an int
|
No, please don't assert tolerance. I like to be able to say "UPEM * .005" for example and that should work. Assertions are for other things: for when the input is wrong. Which is not the case here. |
|
I'm confused. I'm using the built in round() to convert float to int. The maximum tolerance (which I define as the max absolute difference between the rounded and unrounded values) is 0.5, and couldn't be greater.. Am I'm missing something? |
| @@ -77,7 +76,7 @@ def _round(number): | |||
| # return rounded integer if the tolerance is 0.5, or if the absolute | |||
| # difference between the original float and the rounded integer is | |||
| # within the tolerance | |||
| if tolerance == .5 or fabs(rounded - number) <= tolerance: | |||
| if tolerance == .5 or abs(rounded - number) <= tolerance: | |||
There was a problem hiding this comment.
Just make this "tolerance >= .5"
Correct. Which means rounding is an acceptable solution for any user-specified tolerance >= .5. A tolerance of 1.5 is NOT by any means an error. |
|
Ok, I got it. Thanks |
I went for the easy route, and simply added a
roundToleranceargument to theT2CharStringPenas @behdad was discussing in #769.I personally think that doing more clever things would be a bit beyond the scope of fontTools.
But I'm happy to be convinced otherwise ;)
PTAL, thanks