Italic correction #57
Conversation
…pt, and over- and underscripts.
| if (options.ic) { | ||
| styles['.MJX-TEX mjx-mi:not([noIC="true"])' + char.substr(1) + ':last-child::before'] = | ||
| styles['.MJX-TEX mjx-mo:not([noIC="true"])' + char.substr(1) + ':last-child::before'] = { | ||
| width: this.em(w + options.ic) |
There was a problem hiding this comment.
This is confusing. Can we have some auxiliary variables handling substrings? E.g., everything after mo/mi is the same.
Also, I think multiple assignments for initialisation is fine, but I would avoid them for more meaningful assignments.
| let m = 0; | ||
| for (const i of widths.keys()) { | ||
| dw[i] = (w - widths[i]) / 2 + delta[i]; | ||
| if (dw[i] < m) m = -dw[i]; |
| /* | ||
| * @param{string} variant The variant in which to look for the character | ||
| * @param{number} n The number of the character to look up | ||
| * @param{CharData} The full CharData object, with CharOptions guaranteed to be defined |
| * @param{CharData} The full CharData object, with CharOptions guaranteed to be defined | ||
| */ | ||
| protected getChar(variant: string, n: number) { | ||
| const char = this.font.getChar(variant, n) || [0, 0, 0, null]; |
There was a problem hiding this comment.
The default in the data arrays is null. It gets set to {} in the next line anyway, so does it really matter?
| let [u, v] = (this.isCharBase() ? [0, 0] : [this.getU(basebox, supbox), | ||
| Math.max(basebox.d + tex.sub_drop * subbox.rscale, tex.sub2)]); | ||
| const subscriptshift = this.length2em(this.node.attributes.get('subscriptshift'), tex.sub2); | ||
| const drop = (this.isCharBase() ? 0 : basebox.d + tex.sub_drop * subbox.rscale); |
There was a problem hiding this comment.
These (also below) seem important computations. Can you add a comment what exactly goes on for the benefit of others (e.g, this reviewer).
There was a problem hiding this comment.
These are the computations given in the TeXbook (Appendix G). I've cited that in the comment for the function itself. It's not always clear to me what they are up to, either, as the comments in the TeXBook are pretty sparse as well. I'll see what I can do.
| const basebox = this.base.getBBox(); | ||
| const underbox = this.script.getBBox(); | ||
| const [k, v] = this.getUnderKV(basebox, underbox); | ||
| const delta = DELTA * this.baseCore.bbox.ic / 2; |
There was a problem hiding this comment.
If you can't use the distinction between upper and lower case, why have them be distinct? The all-caps version is a constant, and the lower case one is a variable. I'm not sure why that is such a problem. In any case, I've changed the delta to something less meaningful.
| */ | ||
| protected getOffset(bbox: BBox, sbox: BBox) { | ||
| return 0; | ||
| return [0, 0]; |
There was a problem hiding this comment.
I am wondering if we should have a type for pairs and triples of numbers. They appear a lot and might help finding bugs. number[] could be too general. Just a thought, maybe for the future.
There was a problem hiding this comment.
OK, we can consider that for the future.
|
I've made the requested changes. |
Better tall square roots
This PR adds support for italic correction, based on TeX's rules. We use the difference between a characters right-bearing and its width as a surrogate for italic-correction (since we don't have the actual TeX metrics for this).
The italic correction is added to the last character in
miandmonodes (and should probably be added inmnas well, but numbers are usually in roman and so don't have italic correction). So anmiwrapper is added and themowrapper modified to do that. The italic correction is added via the CSS.Italic correction is not to be used in some superscript and under-over situations. So there is CSS set up to remove the italic correction when needed in those cases.
Also, the italic correction is used to properly place super- and subscripts and under-over constructs, as per TeX placement rules. We adjust the italic correction value in theses cases to more closely match the actual TeX placements (since we don't have the actual TeX values used).