@Dr.ManuelKuehner I really don't care if the question is too broad or not. I am worried about the current state. We have Nicolas nice books, we have lshort and the unofficial ref manual, there is a wikibook that needs some attention. Sharelatex wants to improve its own LaTeX documentation, generating a possible new layer of confusion. A starter is left in a flood of documentation and misleading information.
Consider a student attending a LaTeX introduction at a university getting his code reviewed at the monthly regulars table and being told that almost everything is rubbish. Stuff like that happens if the guy giving the introduction never have systematically learned LaTeX and just read a few passages of the wikibook and other suspicious sources. That is what i am worried about, not only because i saw just that case.
@Johannes_B Read the documentation: there is a prepressor named mal involved which transliterates the "roman" input in something the package can handle (similar to devnag for devanagari).
@Johannes_B but I think that neither in miktex nor in texlive the bash-script is working directly (there is no exe or something like this), actually I think the script is not installed at all so one will have to install it manually.
Anyone have an idea as to what this message during indexing means:
ERROR: READ: input stream #<CLOSED INPUT BUFFERED FILE-STREAM CHARACTER #P"/var/folders/70/92ft8q5j4gd1phvhrn2767yw0000gn/T/QyoAeDeOlz" @451> ends within a string
As usual, it's an expansion problem; when the write happens, the error is that \bit is not defined, which it isn't unless we're processing \si or \SI.
I'm not sure why the declaration seems to need being required also in the preamble, but you get a better setup if you avoid expanding the index e...
@PauloCereda Seems that if the last parameter of the solution at the above indexing question contains a pair of double quotes causes this problem. Wondering if I should post a new question? There are no other questions on this site that I found with that error message.
As "SFr." is kind of a unit in my eyes, I would like to do a feature request. Would be nice to have this handled correctly. Especially for the 'double-dot-case' I just have shown.
I am using the solution from Indexing with Binary Units using siunitx package. This not only solved the specific problem that I had posted but seemed to resolve a numerous other problems as well. However, it appears that there is some difficulty if the there are double quotes in the unexpanded pa...
@DavidCarlisle Opps -- did I reply to Murray and say that didn't work? (I'm operating on very little sleep today). I thought I tried it (setting those symbols that accept VS1 to \mathalpa) and it didn't work, but I must have missed something.
@WillRobertson my devices are an old 10in tablet where the main site works well enough, probably better than the mobile view and a phone with a screen small enough that it's best to use as a phone and forget it has a keyboard:-)
@yo' Well, I don't like taking photographs with my cell phone and apart from having my calendar and 'tasks' app I use my phone ... well, as a phone ;-)
I always get an empty page between end of one chapter, and start of next chapter, in a report class. Is this common? I can't figure how to get rid of the empty page. Here is a screen shot:
@ChristianHupfer but I'm aiming quite high, I don't like making compromises. However, I may happen to have to, since my provider is not willing to join my 2 SIM tariffs into 1 SIM, and dual-SIM phones are of lesser quality in general.
@Nasser is it always that the chapter starts on an odd page?
@yo' I see that the chapter ends at page 16, then an empty page at page 17, then the next chapter starts at page 18. Here is the latex code which loads the chapters....
\chapter{Handouts} \section{real analysis, Jan 21, 2016} \begin{center} \includegraphics[width=0.95\textwidth]{hand_outs/handout_real_analysis} \end{center} ...
The empty page is between the "class notes chapter" and the "handouts" chapter.
I wish there is a way that Latex will tell the user WHY it is generating an empty page., I guess I have to go look at each of the 10,000 lines printed on the screen to try to guess why :)
It might be a side effect of using the standalone+import pages. The first chapters is separate document (standalone), then I load it using import into the main document to make one book.
@ChristianHupfer exactly. I've got a very old phone tariff which is very cheap but enough for me. And a 2-year-old special offer internet tariff with LTE and 20GB FUP
@Nasser could be the problem, yeah. I'm afraid that this needs either someone with a good knowledge of the import mechanics (I'm not sure there's someone like this in the chat), or a proper question on the site.
@yo' it is ok., Right now, my solution is to go though the PDF each time, look for empty pages, then delete them manually before I print it ! Easier than having to figure what Latex is thinking about when it does this.
@yo': It would be extraordinarily expensive to have a mobile rate of 20GB with Deutsche Telekom (but they've got the best connection quality, in my point of view)
by accident I figured why the empty page was showing up ! This is amazing. But may be this is how Latex works. The first page in the next chapter was an image. The image was too large. So Latex made an empty page and pushed the chapter and the image to the next page for some reason. Here is before:
Here is after:
\chapter{Class notes}
\subimport*{class_notes/}{class_notes}
\chapter{Handouts}
\section{real analysis, Jan 21, 2016}
\begin{center}
\includegraphics[width=0.75\textwidth]{hand_outs/handout_real_analysis}
\end{center}
@DavidCarlisle it is report class. Not book class. But I do not understand why making the figure smaller in the first page of the next chapter solved the empty page.
@DavidCarlisle The log has many screams. I never look at it, it is pages and pages and pages of screen texts. I only look at errors, not warnings.
@Nasser that's probably a mistake, certainly no point in complaining that you have to "Right now, my solution is to go though the PDF each time, look for empty pages" if latex is in fact reporting every page that is bad.
@Nasser LaTeX tries to place something (in your case the chapter title and the image) on the current page, sees that it doesn't fit and push everything to the next page in the hope that it will work there. When this doesn't work either it gives up, places everything and complains ...
@UlrikeFischer "sees that it doesn't fit and push everything to the next page in the hope that it will work there. When this doesn't work either it gives up" Ok, thanks. Now at least I understand the logic! But if it gave up, why not restore things back? Instead of keeping an empty page?
This strange. Does not Latex uses temporary variables? it can make a page, keeping the old one. If it does not work, it keeps the old one. some of you experts should fix this in Latex please !
@Nasser but even if you don't mind text sticking into the margin by 2cm then it is absolutely wrong to ignore warnings, TeX only warns if boxes are worse than limits that you set. If lines that are 2cm wider than their specified area are OK, set \hfuzz=2cm and then TeX will only warn of lines that are worse than that.
@DavidCarlisle Could you arrange to hold the content of each box255 in memory and only really shipout at the end of the run (or at some pre-determined point)?
@JosephWright well in some cases (I posted an answer here the other day that interleaved odd and even pages from two input files by holding al the boxes and shipping out at the end, but that's easier as you are not changing the page breaking just the page shipout to reconsider page breaking you need to undiscard discaraded items and preferably re-consider line breaking at the paragraphs and....
@JosephWright can you take a mod view on what to do with this question? I don't want to answer it as if you need to ask the answer can only be don't do it but OP keeps coming back in comments
so your lualatex.fmt file does not match your luatex binary presumably you made a new binary and not remade the format. I can only suggest you revert to a clean texlive 2015 and use fmtutil-sys --all to remake all the formats, the current test luatex releases are absolutely not in a state where it is sensible to describe how to build them on a general forum like this. luatex 0.8x releases are for testing only (0.88 is already obsolete, I'm running 0.89.1 for example) If you do make a format, the latex packages currently known not to work include graphics, hyperref, amsmath, ifluatex, — David Carlisle5 mins ago
@DavidCarlisle Hmm, as a general question it's OK I guess: the problem is that the issue here is not just 'how to build formats' but 'how to build them for a very different engine to the one the rest of the set up is for'
@JosephWright yes I don't think it was unreasonable to ask, but I'm not sure what a reasonable answer is:-) I feel a bit responsible as he got into that mess as I commented on a luaotfload github issue saying something worked for me (using 0.88) but he saw that and built the engine and....
@DavidCarlisle Perhaps one could do something in this special case where an a full page is pushed and empty page is produced? (one the other side as people ignore log-warnings it is perhaps good that an empty page show them that there is a problem.
@UlrikeFischer not without some pain, it's not like floats that are more that textheight high were latex is in control and can warn and force the issue by resizing the box, here latex isn't explicitly deciding to ship out an empty page it is just a consequence of a box/glue/penalty calculation
@PeterGrill fair enough. Interesting indeed. I will go this way.
@barbarabeeton thanks for pointing that out. This works and seems to be the correct interpretation of that problem. As Joseph stated afterwords, the dot is already in uses as "inter unit separator".
@JosephWright If it is able to use \DeclareSIUnit{\SFr}{SFr.}, I guess it should be able to escape the inter-unit-product behaviour for a single usage... but well, does not matter. The problem is solved for me.
@DavidCarlisle @JosephWright Adding new formats to fmtutil.cnf is not so difficult. But as the new luatex needs also some other new files I would (if I had the time which I don't have so I'm waiting for the 2016 pretest) copy the texlive2015 to e.g. texliveexp, exchange the binaries and all the files that needs to be exchanged and then change the PATH.
@UlrikeFischer @UlrikeFischer yes but building a format (which requires new ini files) is only the start of his problems, luatex nolonger defines \pdfoutput or most of the other \pdf... commands so by default pretty much nothing works. Joseph and I are trying to get things sorted before texlive 2016 but "don't do that" is only advice that can give for anyone who needs to ask how to get started.
@DavidCarlisle I meant using egreg's solution from tex.stackexchange.com/…, with\AddIndexEntry[Name]{Matrix}{Power}{Power $\mathbf{A}^{\!n}$} results in a \AddIndexEntry[Name]{Matrix}{Power}{Power $\mathbf{A}^{\!n}$}.
As usual, it's an expansion problem; when the write happens, the error is that \bit is not defined, which it isn't unless we're processing \si or \SI.
I'm not sure why the declaration seems to need being required also in the preamble, but you get a better setup if you avoid expanding the index e...
@PeterGrill \mskip-\thinmuskip isn't expandable anyway so makes no difference
@PeterGrill or use a makeindex style that uses a character other than ! as the subindex separator. It is currently assuming a use such as \index{tree!Californian!weird}
@DavidCarlisle Not sure what character I can use that won't cause a problem at some point. I guess since I am auto generating this my script can replace any ! with an "!, that should work? Or are there other special places in LaTeX where a ! is used to mean something else?
@DavidCarlisle: Isn't there a named macro equivalent of \!?
@PeterGrill \"! might work but I didn't suggest it as I'm not totally convinced that everything occurs in the right sequence to be quoted but then be the correct command when needed. I'm pretty sure no one ever tested makeindex-quoting a symbol used as a tex command name.
@PeterGrill no \thinmuskip is a length (actually a muskip register) \! and \foo are \mskip-\thinmuskip like differece between \baselineskip and \hspace{\baselineskip}
@JosephWright fair enough. Thanks for your replies.
@barbarabeeton but "Fr." is the official way. I am writing "10 € are 11 Fr. or EUR 10 are CHF 11"
I am trying to strictly separte between ISO or German currency style. For the old German currency it would have been DEM 10 or 10 DM. For the US currency I would write USD 10 or 10 $"
that's why I do not like "10 SFR". You can see that in some places, but it is not consistent IMHO.
@LaRiFaRi But only in German, if even. For example in Czech, it is wrong to write "EUR 15" or "$ 15", the units have to follow the amount, as in "15 EUR" or "15 $", or "15 SFr."
@DavidCarlisle -- is it possible to tell that a page is totally empty? if so, in the case where it's not an even-numbered page maybe an error message could be added to the underfull warning.
@barbarabeeton it depends how empty it is, the example just showed it was white but in general there can be glue coming from the user or the output routine making it not underfull, or perhaps the user wanted a blank page and did \vspace{10in} to paste in a picture later, hard to say anything is an error with -that_ page the error is the image is making the next page overfull and overfull box warnings are warnings not errors, if users choose to ignore warnings they may ignore errors as well.
@PeterGrill -- applying a negative thin space to improve the appearance of superscripts in some situations is a known issue, but nothing automatic can easily be done about it. see example in the texbook, p.169: $\Gamma_{\!2}+\Delta^{\!2}$.
@ChristianHupfer I don't care. I would leave a comment in structure.tex above the changed part => switch by Christian (last name if you want, title if you want) or a link to the LC post. Whatever you want.
@barbarabeeton That's what I thought @PeterGrill meant originally but he meant that the \! messes up makeindex as ! is a special character there, see following comments.
@yo' Now you are mixing the currency code and the currency symbol. The code is EUR, the symbol is €. So "$ 15" looks really bad in my eyes. I am trying to distinguish between codes and symbols by using the code first. e.g. in tables... but actually I dislike them. But let's say, in Switzerland, you have FR. or SFr. and CHF for the code. I think it would not be logical to use both the same way. That's why I put CHF in front as in DIN 5008
@ChristianHupfer All right. Seems the ping-plonk is malfunctioning once again. by the way, i just commented on the Thesis.cls including wrong headers question.
@yo' for sure. If needed in some document where more currencies are appearing. In normal writing I would say, 10 $, to someone not from the US, I would refer to 10 US$ (which are both correct symbols for the US currency) and in some report about the different kinds of dollars in the world, I would use "USD 10".
"This International Standard does not prescribe a particular style of reference or citation. The examples in this International Standard are not prescriptive as to style and punctuation."
@DavidCarlisle -- just catching up, and now i see that it's interaction with makeindex that was intended with "index". so \muskip-\thinmuskip is what is needed; for text, that would be \negthinskip.
@LaRiFaRi -- i'm a bit confused here. where did i refer to "Fr." or something about how to render it? (i may have, but the past few days have been totally disrupted by heavy snow and inability to escape from home.)
@DavidCarlisle -- no, not on either of those lists. i did see the mail to murray and the subsequent fallout. (nice to have you explaining the problems in that forum. and nice to have asmus as arbiter.)
@ChristianHupfer I set the default to true and added in main.tex wile first occuring %\usechapterimagefalse % If you don't want to include an image, use this to toggle images of.
@ChristianHupfer Mainly to not get overflooded with "i need math equation lign on equal" which has nothing to do with a template, and for me it is a convenient way to just clone a repo and not have to face all that buttons for facebook and google and twitter and ...
And of course, i don't have to personally mail him pieces of code with instructions on where to put them.
@Johannes_B Most likely returning to my old work place, no, no new job really, but new assignments (administration of school computer network, for example)
@yo' No compilation problems, but no visible effect. I'm guessing that it doesn't make any difference that I'm actually using the checkboxes environment and not choices, as I modified your code accordingly.
@Gregory but unless you decide to fill your document with inline lua scripting, you don't have to decide, it's quite possible to write a document that works with either (or all three with possibly different fonts)
@barbarabeeton: Yeah I can see why an automated solution may be difficult in general for the superscript. But, what David Carlisle said is correct: The negative thin space \! was creating an issue when during indexing due to the ! being a special character to the index. Solution is to replace it with a macro.
@PeterGrill -- i figured that out after reading further. (my problem was that "index" also, in a math context, can mean "superscript". which is why i always call those alphabetical listings in the back of a book "indexes" instead of "indices". but proper latin plural for "appendices".)
@Gregory In practice, for most things XeLaTeX and LuaLaTeX will give you identical results. XeLaTeX is potentially more stable, however, because LuaTeX is actively being developed. Furthermore, XeLaTeX probably has the advantage for non-Latin scripts.
@WillRobertson @JosephWright still trying to do VS1 variants... so the variant U+2268 is uni2268.ss09 in xits-math. in text I can do \XeTeXglyph\XeTeXglyphindex"uni2268.ss09" but I could just is the two character combination to get that:-) I suppose I could load the font with ss09 enabled into a new \fam and then switch \fam to get the VS1 variants? seems a bit costly to load the entire font again but...?
@DavidCarlisle @yo' @PauloCereda @JosephWright @AlanMunn Thanks for the tips. I've settled on XeLaTeX. Honestly, I don't do anything complicated, barely ever venturing into macros, even. It's just good to know that what I'm doing should be pretty portable between systems.
@DavidCarlisle Costly? I wouldn't say that :) We load the font separately with ss01 (or whatever) to differentiate \mathcal and \mathscr anyway. It would be a shame to have to have font-specific ways to do all this though. Anyway do we need a way to access the variants separately? I guess a user might want one variant but not the other. (The OT feature could use cvXX instead of ssXX to make that a bit easier.)
@WillRobertson well for the proposed alphabets it's probably more reasonable, the existing VS1 variants are just some weird collection of unrelated things, no particular relationship between a union symbol with serifs and a lessthan not equals symbol with a vertical bar. but I suppose if all the gylphs are available via a ss-something then *tex can access it that way so it doesn't matter what input convention Unicode decide, Not sure whether that is quite the answere they want to hear but...
@DavidCarlisle BUT if you use OT features only then you lose out in "plain text" -- a script letter with ss01 applied is still a script letter in the source. So it only really works for us :)
@DavidCarlisle The more I think about it, the more I like this idea of one extra alphabet — you have "script alphabet" and "alternate script alphabet". Doesn't matter if STIX is curlier than Cambria, since as mentioned people interchange all the time. The only big need is that they are differentiated. Only issue I see is that if a user ONLY uses alternate script alphabet for their document because they prefer it, then switches to a font that only has the "normal" one.
@WillRobertson yes but to be honest the number of places where it matters between caligraphic/script is somewhat limited. In luatex it seems the lua table with the glyphs effectively means everything is accessed by some number anyway eg the VS1 variant in my example ends up in PUA as U+f0334 but again that's probably worse from a semantic point of view (saves messing around with separate fonts though:-)
@WillRobertson you could reply to my mail and pick up on that (other bits of my message were picked up on, but everyone ignored that bit:-)
@WillRobertson That's my preferred version (I think:-)
