The problem is that several symbols doesn't have Unicode counterparts, so we need to output them as two characters. In some cases it doesn't really fit the original.
SkakNew.htf:
SkakNew 32 121
'' '' 32
'!' '' 33
'”' '' 34
'#' '' 35
'$' '' 36
'%' '' 37
'&' '' 38
'&...
2 hours later…
Anonymous
4:51 AM
Somehow, I can't include the character # (number sign) in some-url in ... -> -> ->
{\color{orange}\underline{\href{some-url}{That's a link}}}
@VincentVerheyen You shouldn't do that. hyperref has some features of marking links with color or a frame (even if the frame only consists of the bottom line).
Anonymous
5:52 AM
@Johannes_B I have looked into to those features, but couldn't get the desired result using those options. Instead, there always was a full frame around the links, and not only a line; which is why I opted for the current option.
Anonymous
@Johannes_B For example, both of the answers in Werner's answer here tex.stackexchange.com/a/26085/67761 don't work for me. The first option yields no line at all, and the second option yields a red frame, instead of just an under-line.
I posted a question, then found I had typo in my ties, which made the question invalid since after fixing my typo it works. is it ok to delete it when someone gave an answer? I do not want to be rude.
@Johannes_B when I tried to delete my question, I got a message saying there is answer and it is not recommended to delete it. The answer I have did not actually work for me. I should delete the question I think.
The way my question stands now, it makes no sense, since what I have actually works. So there is no problem to solve. I had silly typo.
@PedroTamaroff Well, as mentioned a figure is a floating environment, which means that LaTeX will position it in the text according to a specific algorithm (see tex.stackexchange.com/questions/39017 for details on that) in order to avoid bad page breaks, i.e. big empty areas at the bottom of pages. You can control this of course, see e.g. tex.stackexchange.com/questions/2275 and tex.stackexchange.com/questions/8625 in addition to the already mentioned post.
@PedroTamaroff the only reason for wrapping an image inclusion in a figure environment is so that latex can move it to help with page breaking. So that sounds like the expected behaviour, not a problem.
@DavidCarlisle The edited answer was corrected and it now answers my old question with the typo resolved. I'll accept it. It shows a method that can be useful to others.
ohh I suspect that's possibly known Stephen reported {{ problems, changes.txt says * keyval.dtx: add \long on internal parsing macros to allow \par in values, Only strip one {} group not two from values. graphics/3446 but then a few months later 2014-05-08 David Carlisle <david.carlisle@latex-project.org>
* keyval.dtx: Revert change to {} parsing. I can't recall what it broke, but I seem to remember deciding that fixing was harder than not fixing after all that time
@JosephWright have a look at graphics/3446 then:-) I think fixing it broke pstricks or hyperref or some other such package that I didn't want to break and only one person had complained about dropping too many { in 20 years so I put it back, but in new code that wouldn't be an issue of course.
@DavidCarlisle In \keyval_parse:NNn the current behaviour is always to strip exactly one set of surrounding braces, so I just have to not break that whilst trying to find all of the bottlenecks
which works like \keyval_parse:NNn (pass processors rather than look up keys) but avoids the assignments we currently have, etc. Need to address babel, the braces and space stripping (all spaces at each end, not just one), but I think I can get some serious improvement
@JosephWright how much slower is slower (compared say to time taken on package loading? ) do you need thousands or millions of key settings to see a difference?
@DavidCarlisle The bigger issue is the 'higher level' (l3keys) code, where I wrote it all ages ago and it's far too slow
@DavidCarlisle I'm looking at this as despite Frank's template ideas, most people seem to use keyval approaches 'live' (ConTeXt, pgf) with no pre-digestion and so performance is important
@JosephWright you do wonder if people follow Frank's 30-year plan of language layering, and all the real code is using positional arguments and just the top level styling keyval, whether there should be more than a few dozen rather than a few dozen thousand settings per document...
@DavidCarlisle Much of the user interface to TikZ is written in pgfkeys, whilst for ever table someone does using siunitx any per-cell data has to be set for every cell (say table-format = 3.4)
@DavidCarlisle I'll give it a once over: will gain siunitx speed, for example (a good demo of the idea that shared libraries allow improvement across the board)
@JosephWright yes I know really:-) in the initial template ideas we tried to arrange that the keyval stuff was resolved at definition time (more or less) but I don't think that ever worked completely.
@DavidCarlisle It does work (when I re-wrote I spent a long time making sure of that): I see the benefits but I do wonder if it's that useful, as 'in the wild' people want to adjust 'just this one case' a lot
@DavidCarlisle Take a look at xtemplate :-) Of course, the issue there is that it's not really right at a wider level
@JosephWright not so much in the wild, as later that evening when Frank and I tried to make any examples of code we'd written that day, we'd always optimised out the interface we needed:-)
@DavidCarlisle When I looked at this and tried to make some test cases, I got the feeling you end up wanting something more like l3keys so you can fiddle about ad hoc
@DavidCarlisle Ah
@DavidCarlisle Of course, if we don't need multiple keyval interfaces then the \keyval_parse:NNn/l3keys split isn't needed and there's a possible (small) optimisation there
@DavidCarlisle I'll work on what I can solve at the moment: making l3keys faster :-)
@DavidCarlisle All good stuff for the team meeting
@ChristianHupfer I am sorry, moderncv is one of the most hideous things i have ever seen and it doesn't have a manual and the freaking warning is there for i don't know how many months.
Of course, you don't have any warning on l-templates, because the old version is still distributed.
@Johannes_B @ChristianHupfer: Not quite sure what you mean but it breaks for me in miktex because xelatex finds the pfb: MiKTeX2.9/fonts/type1/public/fontawesome/FontAwesome.pfb. I have forgotten if xetex shouldn't do this or if one must block the folder.
@UlrikeFischer Actually I don't know what @Johannes_B means (but that's the way it goes concerning the communication between @Johannes_B and me ;-)), but I think he refers to the moderncv xelatex failure due to fontawesome
@Johannes_B Your system looks a bit curious. Why are you using a 2013 tree with a 2015 TL? Beside this xetex is using like in my case a pfb instead of the otf font
Requested font "FontAwesome/I/OT" at 10.0pt
-> /usr/local/texlive/2013/texmf-dist/fonts/type1/public/fontawesome/FontAweso
me.pfb
is texstudio open to change requests? If I see another person confused by mangled error messages I think I'll scream:
@Nasser No. As I say that is not the tex error message, that is your editor showing you the start of a message, and the end of the message, and omitting the important part that says which command is undefined. Look at tex's console output or log file. — David Carlisle1 min ago
@UlrikeFischer still it's not exactly unnatural to assume that an error message display Undefined control sequence. \variable is to be read as saying \variable is undefined.
@DavidCarlisle It tries to show the text on the error line. E.g. a line like "blblub xxxxxxxxx \variable vlvlvld bldldld" would lead to "Undefined control sequence. blblub xxxxxxxxx \variable". With another error it would look like this "Missing = inserted for \ifnum. blblub xxxxxxxxx \variable". Like in real log-messages one has to learn the structure of such messages.
@UlrikeFischer yes but if the editor is only going to show a compressed message it should learn the structure and show the relevant bit. It could detect an undefined control sequence error and show the command that is not defined.
@DavidCarlisle Yes probably. It would naturally much easier if the engines would add the command name in the first line directly. That would be e.g. a useful change in luatex for once ...
I am designing gamecards in LaTeX
Everything is going pretty ok so far but when i try to insert the %-sign for a question about percentages my file keeps giving errors.
I tried putting the percentage between dollarsigns because i know just % will make it a 'remark' and that the dollarsign makes...
@UlrikeFischer Apparently, you can upgrade to a newer TeX Live with tlmgr. The question you linked states, that this i fixed in TL 13. If i recall correctly, i haven't had that issue before. Are you aware of any changes of XeTeX?
% 9830 keyval
% 15925 X with braces retained and all spaces leading removed
% 18087 X with braces and \trimspaces
% 22282 Y (babel safe)
% 55508 current expl3
using \pdfelapsedtime: this is all for \keyval_parse:NN` or emulation thereof
@DavidCarlisle Parsing 3 keys + 2 blanks 10k times
@DavidCarlisle Difficult to compare as I'm only looking at the 'identify a key and value' part at the moment: the model we have separates that out from 'look up a key'. That's where I think the real issue is: expect checkins!
@DavidCarlisle The docs for options discuss the speed requirements, but that is getting as far as finding and executing /some/key/path not just getting to a series of key-value pairs
@DavidCarlisle babel is a pain: costs a lot of time to be 'safe'. I'll see if I can do better on that part
@Johannes_B I saw that. I expected it to be spam. I clicked on the link provided, already expecting nothing healthy and it wants me to enter my cell phone number, together with advertisement of ring tones ;-)
If I have title for a list, how can I keep from breaking the page before the list? \nopagebreak doesn't work. E.g., List of things\nopagebreak\begin{enumerate}\item Thing\end{enumerate}
It is not nice to have a page break right after the colon and have the list start on a new page:
And here, finally, is the list of mission-critical items:
\begin{itemize}
...
\end{itemize}
How can I prevent that?
@AlanMunn Indeed -- this works: \NewDocumentCommand{\KeepTitleWithList}{}{\@beginparpenalty=10000} used inside an environment so that it's contained in a group.
@Johannes_B Saw the spammer. He had 171 messages in the outbox (pending) and 5 sent (read) but as I saw to people who can't be fooled that way. I deleted the spammer profile together with the spam messages. Thanks!
I am trying to use the newfile package to write a multiline file. Something like \def\foo{a\\b}\addtostream{mystream}{\foo} but where a and b are on different lines.
I should have kept searching. I got ^^M and ^^J confused.
Now I need to look for a question that tells me what the difference is.
@Johannes_B No, I don't know of changes, it could also be a fontspec problem. In my case after some fc-cache rounds and similar it worked again. It also worked when I used \newfontfamily{\FA}{FontAwesome.otf}
@DavidCarlisle my pleasure ;-). Did you check the other drivers?
The problem is that several symbols doesn't have Unicode counterparts, so we need to output them as two characters. In some cases it doesn't really fit the original.
SkakNew.htf:
SkakNew 32 121
'' '' 32
'!' '' 33
'”' '' 34
'#' '' 35
'$' '' 36
'%' '' 37
'&' '' 38
'&...
I am designing gamecards in LaTeX
Everything is going pretty ok so far but when i try to insert the %-sign for a question about percentages my file keeps giving errors.
I tried putting the percentage between dollarsigns because i know just % will make it a 'remark' and that the dollarsign makes...
It is not nice to have a page break right after the colon and have the list start on a new page:
And here, finally, is the list of mission-critical items:
\begin{itemize}
...
\end{itemize}
How can I prevent that?
