What's the best way to get a starred version of a LaTeX3 environment so that it functions the same as a 2e one? This old question has an answer, but is it still valid?
@Robbie well the question there was specifically asking about using an s argument spec but I'd have thought the most natural way is just to define the environment as normal. for example tabular* is just an environment name, it is not a * form of tabular in the way that \section* is a star form of \section.
@DavidCarlisle I was looking at more towards the latter case, defining a starred version of an environment as being an unnumbered version of the unstarred environment.
@Robbie sure but in for example align* compared with align the two environments are quite separately defined, but using common internal code. It is not that align looks for an optional *. so you can do same in l3, your expl3 internal functions handle both cases, then make two xparse declarations for the top level environments
@campa The function \xpatch_main:NN starts off by doing \tl_set:Nx \l__xpatch_name_tl { \cs_to_str:N #2 }; here #2 is the macro to be patched. I need the stringified name in order to do later processing. What happens without \makeatletter is that \@ is stringified, but not setref. If you try \xpatchcmd\@setref{}{}{}{} you'd get errors.
@campa In the further processing only \l__xpatch_name_tl is used.
@campa Maybe I should add a check that the argument is a single token.
Given an Integer array:
Start from the first number
Jump forward n positions where n is the value of the current position
Delete the current position, making what was the next position the current position.
Goto step 2 until there is one number remaining
Print that number
Rules
The array wra...
@PauloCereda It's a cute comment. \begin{soapbox} But the argument being made is total crap. Of course there are bad journals, and some of the open access ones like PLOS One are exactly that, but this is not cause and effect. And being a Society Journal doesn't preclude being Open Access. \end{soapbox}
@PauloCereda Two of the best journals in our field are now open access. One was formed when the entire editorial board of a major Elsevier journal (Lingua) resigned and created their own open access version (Glossa). The old one is now called Zombie Lingua. :)
@JosephWright @PauloCereda How can I tell l3build to copy a sty into the ctan-zip? I don't have a dtx or an ins, only one simple .sty in the build folder and it refuses to go into the zip ;-(.
@DavidCarlisle @JosephWright oh well. It should be "sourcefiles" not "sourcefile" ;-(. (btw: I tried at first with typesetsourcefiles but then I got the sty only in the tds.zip. Is this normal?)
@UlrikeFischer The setting typesetsourcefiles is for where you have some DocStrip sources that need to be unpacked for the docs, but not for installation in the TeX tree. Not sure whether it ends up in TDS doc or sources (without looking)
@PauloCereda Rather the other way around: the manifest is a result of the various matches from the variables
@PauloCereda I meant that somewhere the user had to set up the spec. With the current approach, we can have defaults as wildcards and they largely work, and from those one can generate a file manifest (of greater or lesser complexity). If you start from the manifest, doesn't the user have to type in the name of ever file they want?
@JosephWright must admit I haven't followed the full details of the manifest stuff but doesn't the end result have to be the same as saying build ctan ;unzip -l zzz.zip ?
@DavidCarlisle Will is keen more on a 'full' manifest of what is in the source repo, including all of the test files, etc., and (optionally) with descriptions
@DavidCarlisle I have to say my starting point was not dissimilar to yours: for example, in siunitx I list the source and derived files, which means they match more-or-less the TDS-style zip (though I have to admit that doesn't cover the README)
@JosephWright -- while trying to post an update to ctan for the amscls collection, i was severely chided by petra that the manifest file in the *.tds.zip must be identical to the one in the zip for ctan. (the manifest in the amsmath collection follows that dictum.) so if you're posting to ctan and the manifest files have different content in the tds and ctan (flat) versions, you are likely to get scolded too.
@barbarabeeton Currently, I don't have a separate manifest, just a listing in the source (done by hand). Will is working on automation for l3build, but his aims are perhaps different ...
@barbarabeeton it's OK: once l3build has implemented the ctan upload API it will automatically insert into the "additional notes" field "I'm a bot don't argue with me"
@yo' sorry, I meant to say this version is wearing a lab coat which makes strongbad even badder it also has a little space at the top of the picture so nothing gets cut off.