@PhelypeOleinik I took a lot of care when writing each paragraph and drawing each figure. :) Some of them were printed and manually measured with a ruler.
@PhelypeOleinik also, I had the pleasure of using one of the best fonts available to TeX up to date. :) The result on paper was simply gorgeous.
@DavidCarlisle ;-). I saw his talk - one of the many about "how to extract with a computer information from a document" ;-)
@PauloCereda: you would perhaps be interested in the music talk: Augmenting Music Sheets with Harmonic Fingerprints. I didn't understand everything because of my missing music skill, but it sounded quite sensible. They add some color code to music themes:
@UlrikeFischer no I didn't have much time last week, I did look but I failed to understand the original issue, I think. (the one that caused the original change)
@DavidCarlisle the original change was to get a warning about a non-working \hypersetup{pdftitle=xxxx} on page two. Without the change the command does nothing but silently. Imho that is less problematic than an error ...
@JosephWright @UlrikeFischer Perhaps we should have a list to check the status of the packages (no action taken, PR sent, and updated, or something like that)...
@JosephWright imho the largest and most difficult bunch is acro etc. There PR should probably be done. But quite a lot of the other should work with a mail/ticket, e.g. luatexja, luatexko.
@UlrikeFischer I'd not got a local checkout: it's more-or-less 'clean' expl3 code, uses no recent functions, so it's not been updated for about 3 years!
@Skillmon All of \pdftex_...:D, \xetex_...:D, etc. (except for \tex_...:D). Integer constants \c_zero, \c_one..., \c_ten_thousand, \lua_(now|escape|shipout)_x:n, \msg_log:n, \msg_term:n, \msg_interrupt:nn, \str_case_x:nn(TF), \str_if_eq_x(_p):nn(TF), and \tl_show_analysis:(n|N).
@Skillmon Yes, \c_zero_int stays. But all integer constans (which do not end with _int) will be deprecated: zero|one|two|three|four|five|six|seven|eight|nine|ten|eleven|twelve|thirteen|fourteen|fifteen|sixteen|thirty_two|one_hundred|two_hundred_fifty_five|two_hundred_fifty_six|one_thousand|ten_thousand
@Skillmon We moved all of the driver stuff to an 'internal-only' set up in l3backend, so for example that's now \__color_backend_gray:n, as it's an internal part of \color_...
@JosephWright and in which package are those \color_... macros now? They are not part of expl3 are they? If so their documentation is missing in interface3.
@Skillmon They are used by the experimental package l3color. The backend code is used across l3kernel/l3experimental, hence the rearrangement
@Skillmon Probably the colour code is more-or-less ready to move to expl3, but there are a few other things to sort out first (@DavidCarlisle: thoughts?)
@JosephWright just found l3color in the texmf. Unfortunately texdoc l3color opens up interface3, though l3color has its own documentation. Sometimes texdoc is strange.
@JosephWright sure. Funny thing is, that I used \driver_color_.... in the documentation of duckuments, just for the sake of it. Now I'm loading l3color (why again don't I just use xcolor?).
@JosephWright not really, just want to make sure that the documentation of duckuments builds without issues. I use one \color_select:nn { gray } { 0.5 } in the documentation.
@Skillmon Clearly we need some ideas from cellprops, but also probably to avoid using \halign at all, and do everything 'by hand', hence 'it's a Bruno job'
@Skillmon no real drafts, we'll need something like 2e tabular (if only for compatibility) but it might be good to have something as well that isn't so tied to \halign primitive "features" to avoid usual issues with things needing to be at start of a cell, or coloured backgrounds that don't eat rules, etc
@DavidCarlisle never having issues with coloured backgrounds eating (horizontal) rules. And no one in their right minds uses vertical rules (except the stupid university forces them)
@DavidCarlisle but what's with my verbatim columns?
@JosephWright if you want to stack everything up by hand without using the \.align primitives you'd need to process the widths yourself, clearly possible to do, but well... and I don't think you can get spreadtab like possibilities without eating as arguments.
@Skillmon you can't use \cline and \hhline is ... somewhat strange, and even \hline is somewhat at the mercy of your pdf renderer due to the way colortbl adds the coloured panels.
@Skillmon what do you do to avoid the over-printing? (but the fact that you need to do anything is a sign the base code isn't as helpful as it could be)
@JosephWright in general some of his answers here, read some of the token manipulation things he did. Didn't start to read l3regex nor l3fp, but they are on my list. Especially the latter one should be really interesting, though most likely I'll have a hard time with it.
@Skillmon not really asking about the implementation so much as timing, how do you put the line over the colour on the following row (delay it until after that row, or put them in at the end of the table?
@DavidCarlisle well, overprinting by stupid PDF readers isn't solved at all, my cline implementation just gets the definition of the entire row and handles it and doesn't issue a \vskip-\arrayrulewidth. In that regard it's similar to hhline. And it doesn't solve doubled rules, only single ones, but those with sufficient flexibility.
@Skillmon ah OK (so the final result will be more or less like \hhline then, OK, thanks) what I'd like to be able to do is have enough global control of the table that all the coloured backgrounds can be set first then the cell contents for the whole table, then the rules over the top (you can do that now with things like tikz, but having it more tabular like would be good)
@DavidCarlisle sounds good, but will that be possible without grabbing the table contents as arguments? I have no idea how that would've been done, but well, I couldn't have implemented longtable...
@DavidCarlisle I could change it to apply the rule after the next line was set, but then I'd need to hook into tabular and I'm afraid I'd break things in conjunction with other packages, so I won't do that. And I'm satisfied with the results as they are.
@Skillmon well yes that's why colortbl is as it is, it's easy to see why it does things wrong, but hard to change that while sitting on the 2e tabular code.
@JosephWright In my school, unfortunately, to create the tests, for economic reasons and for reasons of interest, they want students to use the mobile phone with an App that does not even have an editor to write equations or complex formulas.
Sorry for my absence, but I'm going to the bathroom.
@DavidCarlisle I'm sorry, I thought that you didn't read it. Didn't I understand the message of b over the c of @mickep?
@DavidCarlisle For your opinion Chrome extension that I have finded here golopot.github.io/tex-to-unicode TeX to Unicode is useful or exist another better than to it?
@Sebastiano read the document at the link above, but really it is no different to latex, it is a plain tex encoding of mathematics that is possibly only implemented by Word. converting latex to unicode is not a well defined problem as latex is unicode.
@Sebastiano the chrome extension looks to be doing simple text replacement of \alpha to α you could do that in any text editor, emacs has a mode that does that by default or you could use perl or python or sed or anything that can do string replacements.
@DavidCarlisle I would like to find an optimal solution, among the time available, a better rendering of the formulas, but if it does not exist I will continue to write my test in LaTeX since 2002, when I have knew this wonderful language.
There are many Italian sites that create virtual classes with apps that use MathJaX that perform the same function as this modest App and I should waste time with the management of my school. I personally will continue my work with dignity and I am not interested in any kind of agreement between my manager and this company.
@Sebastiano you need to say what the problem is before you can ask for a solution. latex is already plain Unicode text so there is no encoding needed, you can use either α or \alpha to get an alpha in latex so even if you make that translation the resulting file is still latex, if you have a system of aligned equations, then you will always need some sort of markup, whether that is latex or UTN28 "plain text mathematics" (which as I say is basically microsoft word linear format)
@Sebastiano I do not understand that comment at all, which company?
@DavidCarlisle I'll forward you my warmest regards and hope to sleep tonight. Here it is not normal in Sicily that at this time we are about 25 degrees. At 19 local time we were about 30 degrees. Tomorrow I go to sea with the fish.
@DavidCarlisle No :-( my "expert" colleagues insert images with medcapture, a program that is within the system, of formulas created with equation editors and images of very low quality
@DavidCarlisle @JosephWright I get a failure in the hyperref tests that I don't understand. The build/test folder contains a picture.png, but the test includes a picture from the original testfiles folder: ho-tex-hyperref/test/picture.pdf Why does it find it??
@DavidCarlisle I think we can blame @JosephWright. it is the new texmfdir setting: if the folder doesn't contain a folder texmf, texinputs ends with /otherGit/ho-tex-hyperref// and so everything is in the search path.
@UlrikeFischer I was going to ask you, they are Heiko's original tests I think, I was going to stop them from going to ctan at least, I suppose any useful tests should be converted to l3build (but perhaps you did that already)
@UlrikeFischer actually they are older than that I just looked at test/test1.tex and recognise Sebastian's style
@DavidCarlisle imho I converted everyone that made sense. E.g. test1 has a companion in testfiles. For now added an empty texmf-folder and now the test pass. (But probably it needs a content for git??)
@UlrikeFischer I don't think they are tests in the sense of l3build with testable results, more just example documents used right at the beginning just as a sanity check that linking was happening.
@UlrikeFischer oh sorry I thought the ones in test were remaining ones that had not been converted I see not the case:-). Any converted ones can go, I agree.