@DavidCarlisle The 'special case' data: the approach I've taken is to use \lccode/\uccode for the one-one mappings where possible, so the data we need is all of the 'other' outcomes. There are a few that are different between string and token list changes, so they also show up. See the end of l3str
@JosephWright ah yes I remember. The fact that you end up with different data for str and tl seems to imply that at least that part should live in those modules rather than a unified unicode handling module, but haven't really thought about the implications of naming here at all, so do whatever works...
@samcarter @marmot Little ducklings were missing, actually. This was an untenable situation. I cloned (xsavebox) mummy duck and here they are. Watch out for tiny little duckling, he is so cute!
@AlexG @UlrikeFischer @CarLaTeX @PauloCereda The captain of the pirate ship could be \node[businessman,evil,minimum size=zcm] at (x,y) ; from the tikzpeople package. Hmmh but this may cause nightmares so maybe not...
@UlrikeFischer WOW. I didn't know this repository! Now @AlexG has a serious competitor for the Oscars. (And if he makes @CarLaTeX the mummy duck, I'm sure she will get an Oscar as well.;-)
@AlexG Interesting. What could wrong with the original convert? The arara rule says convert: {density: 96, otheroptions: -dispose previous -delay 70 -loop 1, format: gif}. Is -loop 1 wrong? (I don't want to upload tons of gifs for tests ;-)
@AlexG My prediction for the academy awards 2019: The Oscar for the best actress goes to @CarLaTeX, the Oscar for the best actor to @PauloCereda and the Oscar for best picture to @AlexG.
@AlexG You may bring your own food there. What they serve in Hollywood is almost as bad as ocean water soaked pineapple pizza. (Perhaps @AlanMunn may like it ;-)
@AlanMunn This was indeed some award show, but not for acting, and I was only the accompanying marmot. Yet it was in a very fancy Hollywood location, and the food was as described above.
@AlanMunn Overall I agree but there are some notable exceptions. Yet the food there was even bad in comparison to other events of similar size. That is, normally I forget how the food quality was, but not in this case...
@PauloCereda -- thank you. i love it! (i was once stared down by the biggest praying mantis i had ever seen, sitting on top of a parking meter in gatlinburg, tennessee. this bug reminds me of that creature.)
\def\arraystretch{1.0} \begin{table} \begin{tabular}{ l c c c c c c r } Ayy & Bee size & Ceeee & Dee & Eeee & Eff & Geee & Aitch \\ \midrule something something & writing here writing & 1 & 2 & 34685 & another thing & 6999 & 10295 \\ \bottomrule \end{tabular} \end{table}
this actually overflowing into the right margin - would one usually go to landscape in this case? I'm not sure... it seems if i could nudge it so that it violated both margins equally it could be ok
@baxx nothing to do with tables, you would see the same with \fbox{zzzzzzzzzzzzzzzzzzzzzzz} boxes start at the left margin and if they are too wide, tex complains about overfull box and it overprints to the right
@DavidCarlisle ok, just happens to be a table, more generally then I'd ask; how to equally violate margins? So that I force it to overlap the left and right evenly, rather than putting everything over the right
@baxx impossible to tell you what to do, go landscape, use a smaller font, reformat the table with narrower columns, steal space from the magins all work technically but may or may not be appropriate, depending on external factors
Or (in your opinion) would this be a case for landscape mode? It just seems that it's only a little bit, and landscape tables are ( in my opinion ) kind of annoying
@baxx looking at your image it is presumably much too wide wider than textwidth+the width of the margin and still stealing space past the right margin?
@DavidCarlisle hrm, yeah i think space stealing isn't enough - I've just found this though : tex.stackexchange.com/questions/27097/… , and I don't think that it's too bad.
@baxx no. one uses a smaller font and adjusts the tables settings to whatever is specified by your documentclass for that size. and the other gratuitously distorts the output to make it fit a specified space.
@baxx would you justify text in a paragraph by scaling each line of text by whatever arbitrary amount that you need to make the text fit the line width, so having wildy inconsistent font size and line spacing? Hopefully not, so why do that to text in a table?
@DavidCarlisle I don't think I have the same respect / appreciation for tables, what yo said made sense though, if there were two on the same page with different font sizes that might look a little funky
@UlrikeFischer, @DavidCarlisle Hans replied about the bonum font problem. "the fonts have bad glyph dimensions that assume italics always to be applied (as in traditional tex fonts)". With \definefontfeature[mathextra][collapseitalics=yes] it looks much better.
According to Hans some fonts look weird when italic correction is applied. That is why italic correction is disabled by default in ConTeXt. (In fact, most features are disabled by default in ConTeXt)
To have italic correction enabled you have define the corresponding font feature
\definefontf...
@mickep as far as i understand, Hans' solution is that it's not his fault and the real fix is to get the fonts changed, so no solution in practice (but perhaps the luatex font loader has workarounds by now I don't follow its internals that closely)
@mickep what on earth are the a and b doing there?:-) I thought the adjustment basically added the italic correction to the advance width and then set the italic correction to 0, but that looks like the itaic correction is getting applied twice?
If I use an integer array in l3, is it better to use many \int_new:c or many \tl_new:c (so in fact the question is: is using many integers ok -- we speak of around 270 integers -- or would it be better to use token lists)?
@Skillmon assuming you never need to use classic tex, you can probably afford to use up some of the 32 or 64 thousand count registers without worrying too much
@samcarter I have a perhaps stupid question: why is in the paths \duckpathbody, \duckpathbill and \duckpathgrumpybill always a cycle at the end? The last point of each of these paths coincides with the first one. My question is if -- cycle has an effect here.
@DavidCarlisle I'm still somewhat reluctant to move to LuaTeX. I don't even have any valid reason I guess, except that I'm used to pdfTeX and pdfLaTeX is set up in all my makefiles...
@Skillmon to be honest that's probably wise (see above for some "features" of math tyopesetting) but medium term that's probably the future, it's hard to justify a system defaulting to 7bit fonts and only using 8bit fonts this century. It sort of works if you are European but apparently the world is round and most people are not European:-)
@UlrikeFischer Thanks a lot! Very helpful! However, \tikz\draw[thick] (0,0)--(2,0)--(1,1)--cycle; seems to produce the same as the second \tikz command of your code.
@marmot yes closing a path implicitly draws a line segment to the start point then closes the path so if the last segment is a line it's not strictly needed, but of course if you use a curve segment to get back to the start then it's needed. (this is really a PostScript/PDF answer I know nothing about tikz:-)
@marmot You already got all the good tex reasons for the cycle at the end. Now the real life reason: This is just the way the inkscape2tikz exporter outputs the paths :)
@samcarter presumably though they were closed paths already in inkscape, the same issues about mitred joins would apply if you draw a closed shape as opposed to a line that looks closed whether or not you export ti tikz
According to Hans some fonts look weird when italic correction is applied. That is why italic correction is disabled by default in ConTeXt. (In fact, most features are disabled by default in ConTeXt)
To have italic correction enabled you have define the corresponding font feature
\definefontf...
