@DavidCarlisle Where did you get cgloss.sty from? Didn't that give an error? I commented out the font line because I assumed that wasn't relevant, but I just got package not found.

@AlanMunn No, the original had both \ex. and cgloss.

@AlanMunn Like logicians, perhaps. Except they also tend to pick identical names for entirely different packages ....And the documentation, if there is any, is sometimes written by somebody else and located on a different site. Not blaming the somebody at all here. If it wasn't for Peter Smith, it would be impossible to find anything.

@mickep do you have an eg?

@DavidCarlisle Excellent. Thank you.

@cfr In phone mode. Will look when I'm at a computer.

@yo' Try this one. Start with walking through the pdf file.

Morning all

@cfr gitlab.com/smc/fonts/Nupuram/-/blob/master/docs/README.md

@cfr that had alrready been deleted (also the \ex. that causes the error had gone, I put it back in the question)

@JosephWright Hi Joseph, say hello to your wife and tell her that her directions were very helpful to me

@CarLaTeX Will do

@JosephWright Thanks <3

Since everyone else are still on vacation, I'm playing around with beamer templates. In my template I want to do all the configuration and setting using key=val's, \title, \author etc will not be supported. But which key system to use? In an earlier version I used pgfkeys (as I like the keys to have spaces) and in another project I used the expl3 key system. Can the latter also do key names with spaces? Or should I stick with pgfkeys as this is beamer after all.

@daleif Depending on how/where you want to use your key values, using \DeclareOptionBeamer etc. might also be an option.

@samcarter I need both. There are options that needs to be specificed when loading the theme. And there are options to configure the presentation (setting, title, presenter etc)

There might be some overlap. but there will be options that are required to be set at load time (a font option that controls how the fonts are loaded).

@daleif ah ok, then one of the other key value thingies might be better suited

@cfr It's from here: staticweb.hum.uu.nl/medewerkers/alexis.dimitriadis/latex

Stupid question, say #1 or \tmp contains "foo bar", say I now want to use this to build \foobar from #1 or \tmp (which may be complicated but in the end resolve to "foo bar")

Was testing \csname ... \endcsname but that did not always work. I had though that it actually removed all spaces before it generated the macro name.

@daleif \csname\zap@space#1 \@empty\endcsname

@DavidCarlisle that was fast, you seem to know the command.

@UlrikeFischer best command ever, @Skillmon told me.

@DavidCarlisle wtf, never seen that before, thanks

@daleif so in general \expanded{\noexpand\csname\noexpand\zap@space#1} \@empty\endcsname

@daleif it's why \documentclass[a4 pap er]{article} works

@DavidCarlisle and presumably why one cannot get pgfkeys style class options.

@daleif yes it's the cause of decades of pain for anyone trying to extend the option system in any way. One of my finer achievements:-)

\expandafter\def\csname myprefix\zap@space#1 \@empty\endcsname{something with some pgfkeys}, seems to work ok. UI generators are always useful.

@daleif the new kv nased option system avoid it as for some releases now latex has also stored the "raw" option list that has not been passed through zap@space, in addition to the original processed list. In latex2.09 (as you probably do not remember) you could not have spaces around commas in the list. \documentstyle[12pt, leqno] would fail looking for a file <space>leqno.sty

@DavidCarlisle good old days

If one could also get pgfkeys to automatically be global, then one can make some nice interfaces. I tend to think that \begin{env} keys \end{env} is a little cleaner than \config{ keys }, but it is probably not worth the effurt

@daleif \NewDocumentEnvironment{env}{b}{\config{#1}}{} ?

@daleif unless this is actually for some tikz/pgf based layout I'd use l3keys rather than pgfkeys, as they are natively part of latex.

@DavidCarlisle I asked ealier and I'm not sure l3keys can do keys with spaces in their names.

@daleif Of course it can - \keys_define:nn { foo } { bar~baz .code:n = whatever }

@DavidCarlisle That was what I was playing with (almost), but then it seemed that there was no code/data outside of env, so I went back to just \config for now.

@JosephWright I'll see if it can be converted tomorrow. Most works just fine for now.

And I like the l3keys structure a bit better

And that is why one builds getter macros, then one should be able to easily replace the backend structure with no changes to the front end.

@daleif oh sorry yes you'd need something like \gdef\foo{\config{#1}}\aftergroup\foo`

@daleif which is an indication that an environment is a dubious syntax as environments are supposed to be local scopes (although latex already has lrbox environment that does the endgroup tricks to make a local assignment that is not scoped by the environment)

@DavidCarlisle I do the same thing for our univesity letter class, though these are set by separate (global) macros, not keys. I just find the structure more appealing

@mickep Interesting. That only defines glyphs, though? Or am I missing something?

@daleif I use lots of l3 keys with spaces in their names.

@AlanMunn Thanks. Is there a LaTeX for Linguists? It sounds as if you need the equivalent of what logic has in lieu of CTAN.

Is it possible to get expl3 indexed properly with ltxdoc? Or should I better drop the expl3? (There is not much expl3 and it does not look easy to switch to l3doc.)

@cfr There was a page a long time ago, hosted out of Essex if I recall, but it was always super out of date. The problem with those sorts of pages is they tend to reflect just what the people made them know rather than what's actually out there unless there's a concerted effort to keep them up to date, and as a result they end up perpetuating old information.

@cfr The younger crowd are more likely to use GitHub which at least makes finding unreleased packages a bit easier. But when something gets a good amount of use, it would be really much better to put it on CTAN. I think though for some people who aren't that invested in TeX they see that as either too much work, or perhaps too much commitment.

@AlanMunn Yes. The logic one runs that risk, but it has been maintained well for a while and is by far and away the best source of information I'm aware of. It is certainly the first place I look for 'a package to do x'. It often provides the only documentation there is for packages. Or a quick-start guide. Or a guide to adapting some more general package.

@cfr As it is, I hate linguex, but it's very popular because it has minimal markup for the formatting you get. But it's so non-semantic, and it will be very hard to add tagging info, I suspect.

@AlanMunn Agree about CTAN. It would also help avoid name conflicts. (Do I have the same nd.sty? Is this nd.sty even distantly related to the one you're using?) There are quite a lot of packages for logic, but very few on CTAN. Especially for natural deduction.

@AlanMunn I think tagging will be a headache for logic since so little of it shows up in the 'official' TeX world. The packages people use for teaching, for example, are not even going to be on the project's radar.

@cfr And there is a degree of scariness to releasing code very publicly. Also you then need to document it better to, and that's not as simple as it seems, and takes some time to do well.

And I suspect symbols will be problematic, too.

@cfr Yes, tagging in linguistics has similar problems. Not so much from the packages used (most of which actually are on CTAN I think) but because it's quite unclear how to deal with the things we represent.

@AlanMunn Yes, I know. :)

@AlanMunn Turnstiles are really important, but I don't know how you get that properly tagged if you rely on something like turnstile, which is the only real option. So they are not even showing up as characters in the syntax, I guess. And lots of introductory logic systems rely quite heavily on visual methods e.g. boxes, scope lines, trees etc., which I assume (though I've not looked) are more difficult to tag.

@AlanMunn The things rather than the representations?

@cfr Well glossed examples are a thing, but not really a representation (in the linguistic sense) they're just a convention for display.

@cfr It seems to me that metapost is used to convert to svg, which in turn is used to make the font files. So, indeed a matter of definition.

@mickep Hmm ... presumably, you could do the same (perhaps less conveniently) with something like Inkscape. So maybe metapost is somewhere in between font and not-font ....

@AlanMunn what is a glossed example exactly? I've seen them, but I don't really know what they are.

@cfr some of the earliest metafont to Type1 font conversions went via interpretting the original metafont as metapost and then converting the generated postscript to type1 outlines, eg google just turned up matijs.net/blog/2005/07/10/…

@cfr So they're used for displaying example sentences in arbitrary languages. So the first line is the sentence in the regular orthography (it a latin based script) or a broad phonetic transcription or accepted romanization (if non-latin based script). The second line then gives a word-for-word or sometimes morpheme-by-morpheme translation into English (or the main language of the article) and the third line gives an idiomatic translation into that same language.

@cfr I don't know inkskape, but I guess so. But I also guess that metapost is much nicer for the purpose. Did you by the way look at the font(s)? I'm very impressed by the work.

 disgrifiad    (manwl)  y     gyrrwr o  ’r   ddamwain
 description (detailed) the  driver  of the accident
 ‘the driver’s (detailed) description of the accident’

@cfr ^ A Welsh example. This is just a phrase, of course, not a whole sentence. Sometimes the second line also includes grammatical annotation (case, number, tense etc.) if that's relevant to the data being discussed. So we might have 'driver.SG' if we wanted to make it clear that 'gyrrwr' was a singular noun.

@mickep I didn't, but now I shall.

@AlanMunn Thanks. I didn't know which bit was the 'gloss'.

@cfr Technically the second line is the gloss. Although the whole thing is called a glossed example.

@AlanMunn So actually close to ordinary usage of 'gloss'?

@cfr shiny paint on woodwork, yes exactly

@DavidCarlisle I think I just thought there must be more to it.

@cfr I may have glossed over some of the details

@DavidCarlisle Is it possible to get ltxdoc to index expl3?

@DavidCarlisle really? I can't imagine you'd ever do that.

@cfr answer must be yes in principle, but it might be easier to use l3doc (as that was the point of it)

@DavidCarlisle which means it might be easier to get rid of expl3?

(it doesn't seem as if it would be easy to switch to l3doc)

@cfr that depends how much you have,

@cfr on the other hand doc should mostly work I guess you'd just have to tell it _ and : are letters in the right place....

@DavidCarlisle for this, very little. or I could not have an index.

@cfr I'm not sure what the ordinary uses are (other than @DavidCarlisle's examples).

@DavidCarlisle 'just'

@cfr do you have a small failing example

@cfr If you look at the ExPex documentation you can see some more examples of the kinds of things people do.

@DavidCarlisle I can make one.

@AlanMunn thanks.

@cfr don't put too much effort in to that as I may get bored and not look, but on the other hand I might (or I might have done it before, not sure, seems quite likely)

@DavidCarlisle I only have to delete stuff. I don't even need to make a copy of the files.

% \iffalse meta-comment
% exp.dtx
% \fi
%
% \iffalse
%<*driver>
\RequirePackage{svn-prov}
\ProvidesFileSVN{$Id: exp.dtx 10163 2024-07-16 06:00:15Z cfrees $}[v0.0 \revinfo][testadf DTX: test for 8-bit engines]
\DefineFileInfoSVN
\GetFileInfoSVN*
\documentclass{ltxdoc}
\usepackage{exp}
\EnableCrossrefs
\CodelineIndex
\RecordChanges
\DoNotIndex{\verb,\ProvidesPackageSVN,\NeedsTeXFormat,\ProcessKeyOptions}
\begin{document}
  \DocInput{\filename}
\end{document}
%</driver>
% \fi
%
% \changes{v1.0}{2010/07/17}{First public release.}

@cfr thanks (I think:-)

(\end occurred when \ifx on line 13 was incomplete)

-- $Id: build.lua 10137 2024-07-05 23:44:52Z cfrees $
-- Build configuration for exp
ctanpkg = "exp"
module = "exp"

@DavidCarlisle I don't get an error if I copy the code back into an empty exp.dtx and exp.ins in a new directory? I don't get an error with or without l3build.

@cfr gone now I think

@DavidCarlisle ??

@DavidCarlisle I've got doc to work with expl3 in the index in the past but not with the kind of automated indexing ltxdoc does.

@cfr if I just change the class to l3doc it runs without error and writes a more plausible idx file, but I will look at making it work with ltxdoc, it really can't be that hard:-)

@DavidCarlisle The documentation doesn't really encourage switching .... It also complains a lot about my use of internal variables (in the real document).

@DavidCarlisle It definitely improves the index, though it does need typesetruns = 5 for some reason. I wonder what will break if I switch more complicated cases ....

@cfr yes because you shouldn't use internal variables from other modules:-)

@cfr or if the internal variable is from the module being set up use <@@=test> so it knows that