Code review for the linguistix package (part 2) - TeX

TopAnswers TeX

Meta

Databases

TeX

Code Golf

APL

C++

.net

db<>fiddle

Java

*nix

PHP

PowerShell

Python

Rust

टेक्-मराठी

Typst

Web Client Dev

Web Server Dev

Code review for the linguistix package (part 2)

Code Review

add tag

निरंजन

In continuation to the [older post](https://topanswers.xyz/tex?q=2114), this post is targeted to the review of key-value interface of package `linguistix`. I am now pretty convinced that `expkv` family is brilliant at handling key-values for different purposes and hence I have shifted from `xkeyval` to `expkv` completely. The old `expex` package uses `xkeyval`. Though the git repository and the complete .sty is shared in the previous post, the following is the code I have used for developing keys for `linguistix`. There are MANY more from `expex` yet to appear in linguistix, but I think the mechanism will remain the same. Any comments regarding developing the following code are welcomed.

```
\ekvdefinekeys{lingset}{%
  skip aboveexskip            = {\lng@aboveexskip},
  skip belowexskip            = {\lng@belowexskip},
  skip numoffset              = {\lng@numoffset},
  skip textoffset             = {\lng@textoffset},
  skip labeloffset            = {\lng@labeloffset},
  skip belowglpreambleskip    = {\lng@belowglpreambleskip},
  skip aboveglpreambleskip    = {\lng@aboveglpreambleskip},
  skip aboveglftskip          = {\lng@aboveglftskip},
  store everygla              = {\lng@gla@font},
  new code exnoformat         = {%
    \def\lng@exnoformat{%
      {%
        \lng@replace{X}{#1}%
      }%
    }%
  },
  new code labelformat        = {%
    \def\lng@labelformat{%
      {%
        \lng@replace{A}{#1}%
      }%
    }%
  },
  choice style                = {%
    wrap   = {\def\lng@begingl@style{wrap}},%
    nlevel = {\def\lng@begingl@style{nlevel}}%
  }
}
\ekvoProcessLocalOptions{lingset}
\NewDocumentCommand{ \lingset }{ m }{%
  \ekvset{lingset}{#1}%
}
\lingset{%
  aboveexskip            = {2.7ex plus 0.8ex minus 0.8ex},%
  belowexskip            = {2.7ex plus 0.8ex minus 0.8ex},%
  exnoformat             = {(X)},%
  everygla               = {\normalfont},%
  labelformat            = {A.},%
  numoffset              = {0pt},%
  textoffset             = {1em},%
  labeloffset            = {1em},%
  belowglpreambleskip    = {1ex},%
  aboveglpreambleskip    = {1ex},%
  aboveglftskip          = {1ex}%
}
```

That's because the replacement of multiple consecutive space characters into a single space token is done at the tokenisation step. But the two spaces are already tokens not characters when they're put next to each other.

There will be no braces in the expansion of `\foo` (one set of braces is stripped by the normal TeX argument-grabbing). Instead this would just look like `␣␣` since `#1 <- ␣` by TeX's macro expansion rules.

Well, the opening braces of for instance `new code exnoformat` need the `%` (else you get a spurious space, since that space would be inside braces and not at the top-level), but the one at the top level will be stripped (but as I said, it does no harm and there is no real performance difference if it's there or not -- unlike in `l3keys`, that would be slower with the space as it would then loop since it removes *all* surrounding spaces, not only the first).

> for instance \def\foo#1{␣#1} and then using \foo{␣} will result in two consecutive space tokens. Strange! Never thought of it like that. Feels like spooky braces just appear during the expansion (`␣{#1}`) :) > * except from the first line ending in % > * note however that you don’t have to use any of the ending % in your \lingset-call here Ah! Completely missed it, it's inconsistent also with what I have done later in the code, but somehow leaving an opening brace without a `%` just makes me uncomfortable these days. I wish spaces _had_ done something wrong here too, so that the absence of `%`s wouldn't have "buggled" me. :joy:

To lift some of the mysteries of space stripping done by `expkv`: TeX will combine multiple consecutive spaces into a single space token, so in TeX `a␣b` and `a␣␣b` are the same (but if you add a space by a macro definition this might not be the same, for instance `\def\foo#1{␣#1}` and then using `\foo{␣}` will result in two consecutive space tokens. And spaces at the start of a line in your source file are ignored as well (but not at the end, and a newline is the same as a space at the end of a line in normal settings). `expkv` will strip *one* space token from either end of each key name and either end of each value and afterwards, if the outermost tokens are `{}` those are removed as well (but only a single brace pair). So your usage of `%` is fine (and absolutely perfect in your `\ekvdefinekeys`-call, except from the first line ending in `%`, that could be dropped but it makes no difference), note however that you don't have to use any of the ending `%` in your `\lingset`-call here (but again, no harm is done by them, it makes no difference).

I always get confused about where the `%`s at the line-end make a difference in expkv and where they don't. In the documentation there were no `%` in the `\ekvdefinekeys`, so I didn't use them there, but are they required in commands like `\lingset` from this code which work with keys developed by `expkv`? I was unsure hence I added them anyways.

0 Answers