@DavidCarlisle feature -> bug

@DavidCarlisle thanks.

@Skillmon I don't like doing that. actually, I used ctrl+c. but I try to do that only as a last resort. (it seems ... rude? unwise? dangerous? no idea ...)

    local new_mcontent = (string.gsub(mcontent,"(%<%[?)" .. jpatt .. "( %<%w+%.pfb \" fontinst%-autoenc[%w%-_]*)( ReEncodeFont)", "%1" .. targenc .. "%2-" .. tag .. "%3"))

is there any significant difference between the above line and the below line if I'm matching lines in a .map files?

local new_mcontent = (string.gsub(mcontent,"(%<%[?)" .. jpatt .. "( %<%w*%.pfb \" fontinst%-autoenc[%w%-_]*)( ReEncodeFont)", "%1" .. targenc .. "%2-" .. tag .. "%3"))

I don't understand why the second works and the first works sometimes since no fonts have empty names.

@cfr just one * to + change, or did I miss something?

I'm trying to test tagging of PDFs, and I'm running into some problems. Clearly, I'm missing something, but I have no idea where to start. I have followed the example from github.com/latex3/tagging-project/blob/main/documentation/… with a simple 'hello world' (using 'article.cls'). I'm getting LaTeX Error: File tagpdf-base.sty not found. as an error. What packages/config am I missing? tlmgr insists that pdfmanagement-testphase is installed.

@enkorvaks you need to install tagpdf

@cfr not this time, it really is a tex feature, in that the logging here is out of latex control, it's primitive tex behaviour.

@DavidCarlisle Thank you, that got it working. Now to test my class (and I may have more questions about making that compatible, later)...

@enkorvaks it's best to test with lualatex-dev first. Usually lualatex is OK and also pdflatex(-dev) but if you just want to test one format to get things initially working start with lualatex-dev

@enkorvaks and of course, if anything breaks, blame @UlrikeFischer not me.

@DavidCarlisle currently just using pdflatex (no -dev), and it is compiling, at least. I'll check for tags later (dinner now).

@DavidCarlisle I was assuming that was where blame goes.

@UlrikeFischer ooh Island of TeX

@enkorvaks you mention dinner and it arrives

@DavidCarlisle oh no

@DavidCarlisle I think so, yes. though I stared at it for so long there could be anything in it. that is, I constructed the second separately because I couldn't figure out what was wrong with the first. I expected what I'd screwed up to be obvious when I compared them, but the + to * was the only thing I could see and I'd have expected either to work?

@cfr well yes, the * one should match everything the + one does.

@DavidCarlisle tagging It will work with xelatex too (I haven't tried it)?

@PabloGonzálezL not really, no. (xetex can not get space characters into the stream so white space is lost even if the actual tags are correct, so it's basically doomed unless the engine is updated which is unlikely)

@DavidCarlisle Ok, does that mean I could do it in the future?

@PabloGonzálezL anything is possible but there isn't really any enthusiasm for maintaining xetex as it's basically just luatex without lua

@DavidCarlisle Hehe, I stopped using it a while ago and moved to lua :D

@DavidCarlisle Since you are the current maintainer (github.com/rf-latex/capt-of) ..can you make it compatible with tagging?

@DavidCarlisle Didn't AMS go for xetex?

@PabloGonzálezL no

@DavidCarlisle jjajaj :D...I knew you would say that :D

@mickep they were veering that way, but as far as I know they still use pdftex in production, I haven't seen any updated author instructions for submitting documents with opentype fonts (unelss @barbarabeeton knows more)

@PabloGonzálezL @UlrikeFischer has write access, then she can take any blame

@DavidCarlisle It seems right to me :D

@DavidCarlisle Ah, OK. I just thought I read it somewhere (inhere?). Probably still pdftex then. And still not stixtwo, it seems.

Of course,everyone will soon use xpdftex anyways.

@PabloGonzálezL it might make sense to keep capt-of compatible with caption to avoid introducing weird package load order artifacts, I'm sure that'll get done sooner rather than later, and do capt-of at same time

@mickep except that that is still a secret format

@DavidCarlisle Yes,but note "soon" :)

@DavidCarlisle and @mickep -- Sorry, I'm totally out of that loop. Yes, XeTeX was being used for at least some work at the time I retired, and the word was that the conditions that required pdflatex had been overcome. But I don't know about later conditions.

@DavidCarlisle have I? I lost a bit the overview

@DavidCarlisle Ok, I understand, at the moment the caption package is not adapted for such a task and capt-off still appears in many of the forum responses.

@UlrikeFischer rf-latex == you me Frank @JosephWright and @egreg, it seems

@UlrikeFischer we could do this ahead of caption package?

@barbarabeeton Thanks for a quick answer 👍

@UlrikeFischer I just made you owners as well as github moaned at me for only having one owner for the organisation

@DavidCarlisle perhaps, but captions are bit difficult, as they have to be the first in a structure and you must have a proper structure, not only a div.

@UlrikeFischer that doesn't really fit with \captionof main purpose which is to allow captions pretty much anywhere:-)

@DavidCarlisle (@UlrikeFischer) Everything related to caption or footnote is really a hassle...although the author of footnotehyper (Jean-François Burnol) will make changes for compatibility (as soon as he has some time)

@DavidCarlisle yes.

@PabloGonzálezL what is footnotehyper doing?

@UlrikeFischer footnote pkg + hyperref support

@UlrikeFischer And it has the advantage that you can leave 'footnote' in any environment as if they were part of the text (at the end of the page) for example with minipage

@UlrikeFischer @DavidCarlisle Which makes me wonder if you could have a hook or socket to place \footnote outside of minipage... list+image use case (side by side), I only use minipage for the layout, but list is part of the text and you don't want to show \footnote as it comes by default in minipage.

@DavidCarlisle but the other way? * worked but + didn't. I thought + would match one or more and * none or more.

@cfr yes, do you have an example of a string that only matched the * version?

@DavidCarlisle yes, but I suspect the answer is going to be 'then the problem is caused by code you've not shown'. so the string was the contents of a .map which looked like this:

ylyr-t1 LibrisADFStd-Regular <t1-cfr.enc <ylyr8a.pfb " fontinst-autoenc-t1-cfr ReEncodeFont "
ylyr8s LibrisADFStd-Regular <libris-supp.enc <ylyr8a.pfb " fontinst-autoenc-libris-supp ReEncodeFont "
ylyri-t1 LibrisADFStd-Italic <t1-cfr.enc <ylyri8a.pfb " fontinst-autoenc-t1-cfr ReEncodeFont "
ylyri8s LibrisADFStd-Italic <libris-supp.enc <ylyri8a.pfb " fontinst-autoenc-libris-supp ReEncodeFont "
ylyb-t1 LibrisADFStd-Bold <t1-cfr.enc <ylyb8a.pfb " fontinst-autoenc-t1-cfr ReEncodeFont "
ylyb8s LibrisADFStd-Bold <libris-supp.enc <ylyb8a.pfb " fontinst-autoenc-libris-supp ReEncodeFont "

and I got matches for the ts1-euro.enc lines but not the t1-cfr.enc lines, but I could not figure out why.

@cfr running texlua cf1.lua on the following file, both match the t1-cfr string

local jpatt = "t1%-cfr%.enc"
local tag = "?tag??"
local targenc = "?targenc??"

local mcontent='ylyri-t1 LibrisADFStd-Italic <t1-cfr.enc <ylyri8a.pfb " fontinst-autoenc-t1-cfr ReEncodeFont "'

local new_mcontent = (string.gsub(mcontent,"(%<%[?)" .. jpatt .. "( %<%w+%.pfb \" fontinst%-autoenc[%w%-_]*)( ReEncodeFont)", "%1" .. targenc .. "%2-" .. tag .. "%3"))

print (new_mcontent)


local new_mcontent = (string.gsub(mcontent,"(%<%[?)" .. jpatt .. "( %<%w*%.pfb \" fontinst%-autoenc[%w%-_]*)( ReEncodeFont)", "%1" .. targenc .. "%2-" .. tag .. "%3"))

@DavidCarlisle thanks. so definitely 'code you've not shown'. sorry to waste your time.

@cfr isn't wasting time the reason for having chat sites?

@DavidCarlisle I thought it was to avoid messing around with the question button.

@DavidCarlisle it even works in my actual script ....

(at least, it works when I remake the files from scratch)

@cfr is there a question button?

@DavidCarlisle only if you believe there is

@cfr well, ctrl+C is "rude", but ctrl+d is totally fine. That just means end of file.

@Skillmon oh. thanks. I didn't know it meant that in general. (I knew some programmes tell me to use it, but I thought that was specific to them and that's why they told me.)

@cfr you can close your terminal/shell in the same way.

@cfr @cfr U+0004 is end of transmission

@Skillmon oh.

@DavidCarlisle oh?

I assume that's what ctrl+d does, but how would I know that?

@cfr how would you know what, that Ctrl D is ascii and unicode 4, or that it's called end of transmission?

@cfr D is character 68 and control is "subtract by 64" so control D is character 4

@DavidCarlisle either, I suppose. but I meant the first. (I assumed I could look up the second.)

@DavidCarlisle thanks. (?)