« first day (2164 days earlier)      last day (2770 days later) » 

5:33 AM
@AlanMunn ha ha i agree with you :)
 
 
1 hour later…
6:54 AM
@JosephWright (@egreg) did we change \font_suppress_not_found_error: some code that I say I got from @egreg was reported as not working this morning, I just changed to set \suppressfontnotfounderror directly. The \font_suppress_not_found_error: is in l3trial/xfont/l3font but did it used to work as posted?
5
A: XeLaTeX: Set an alternative font in case it is not found

David CarlisleThis works in both xelatex and lualatex and sets the document in Pagella for me: \documentclass[11 pt, a4paper]{article} \usepackage[T1]{fontenc} \usepackage{fontspec} \ExplSyntaxOn \NewDocumentCommand{\onlywarniffontnotfound}{} { \suppressfontnotfounderror = \c_one %\font_suppress_not_fou...

 
@DavidCarlisle @egreg oh my, a duck!
 
7:13 AM
@DavidCarlisle Probably not! I've added all of the primitives to \(whatever)tex_... but that's different
 
7:56 AM
@JosephWright odd thing is OP reported it working at the time and I thought I had tested egreg's addition but I couldn't see any relevant changes in the log. Oh well it works now:-)
 
8:28 AM
Is there a way to set different fonts to different scripts so that my document can contain mixed Japanese, English, Czech, without font switching? XeLaTeX or LuaLaTeX.
IOW, what browsers do.
 
@wilx In XeTeX you can use interchartoks, in LuaTeX it is doable from Lua, but there's a lot of concern about the idea: you can't get hyphenation right that way, for example. For typesetting you do need language mark-up.
 
@JosephWright I see. OK.
 
@wilx in particular while you can probably reliably switch between say english and japanese, it's much harder to switch between english and french so while the fonts may be OK, hyphenation (which browsers, by default, don't do) will all be wrong
 
8:47 AM
@JosephWright ah used to be in fontspec github.com/wspr/fontspec/blob/…
 
 
1 hour later…
yo'
9:54 AM
@JosephWright @DavidCarlisle please, (English question) do you have money in your account, on your account or at your account? I suppose it's in but I'm never sure...
 
10:05 AM
@yo' in (also on but that means something different)
@yo' I remember a long time ago when I was a phd student a german visitor asked a similar question, but whether money was in or on a bank, I replied that of course it was in, because a bank was a building that you walked in to, but he correctly pointed out that we stole the german word for bench/counter without understanding it, and you put money on a counter. I decided it was best not to know too many foreign languages, just leads to confusion:-)
 
10:23 AM
@DavidCarlisle Of course the German visitor was wrong: the word bank in the modern sense he was referring to comes from Italian; we stole the word from German.
 
Max
@DavidCarlisle Gday. I have more issues with longtable. The problem is that hacking that line into the output routine leads to problems with a never converging Table widths have changed cycle upon large tables.
I can force widths of the columns in which case I get loads of overfull and underful boxes, which would be acceptable for the guy I am helping. However the last column doesn't let me specify width and orientation. This breaks in \LT@hline for some reason
Hmm it appears that it doesn't recognize the last column specifier. If I remove the \hlines then it complains about extra alignment tabs
 
10:43 AM
Is there any sort of LaTeX cheat sheet that I can print? I want something that will remind me of things like different kinds of space, \@ vs. \␣, etc.
 
10:55 AM
@egreg but you probably understood it:-)
@Max sorry: too much information not enough detail to process it:-)
 
yo'
@DavidCarlisle Thanks!
 
Max
@DavidCarlisle http://tehsuck.de/tmp/ltbl.tex
Compile it and you'll see the error, it goes away when you remove the last raggedright in line 94
if you replace the column specifiers with {llrrr} you will see the nonconverging behaviour
 
@Max ! Package array Error: >{..} at wrong position: token ignored.
 
Max
11:11 AM
That's weird, mine doesn't say that. Plus why would it be wrong? I just saw the spurious | but removing it doesn't change anything
 
@Max it has >{\raggedleft}|p with > applied to |
 
Max
yeah I noticed, but it fails with the same error for me even after removing it
 
@Max sorry no time to debug that but the chunk logic looks suspicious you are injecting a new chunk but I'm not sure that the chunk counter is reliably set (and LT needs to know which chunk the widest entry in each column is from, so it can detect changes.
 
yo'
11:31 AM
@David @Joseph I'm afraid you'll start to hate me for my English questions, but is 740-meter-long train the proper equivalent of 740\,m long train (it's for article titles where I prefer to omit SI shorthands for units)
 
@yo' well it's metre in english english:-) and no hyphens but yes a 740 metre long train (or a train of length 740 metres reads better)
 
yo'
@DavidCarlisle ok thanks :) I'll just have to decide on the re-er business (basically, look for signs of British vs American English preference elsewhere in the article, which is going to be tough given the authors are Czech :-D )
 
@yo' think of it as deciding between right and wrong, and it'll probably become easier
(I'm sure @barbarabeeton would agree with me)
 
yo'
@DavidCarlisle it's more like "authors won't complain either way so I don't care" :-)
 
@JosephWright see the xetex harfbuzz checkin? (Khaled's back:-)
 
yo'
11:56 AM
damn, is ad-hoc an adjective or is it considered a preposition with a pronoun. I mean, shall it be capitalized in titles?
 
@yo' The dictionary says it's written ad hoc, adjective or adverb.
@DavidCarlisle Yes: we use banco for the counter and banca for the institution or building. Although some banks are called Banco for historical reasons.
 
yo'
@egreg ok thanks
@egreg in depth is an adjective as well and we don't write In Depth
I better asked the professionals:
0
Q: Title-Casing "ad hoc"

yo'Let's say there's an article or a book titled solutions for ad hoc problems in foobar. How is ad hoc title-cased? As a preposition ad with a pronoun hoc or as an adjective? I mean, which of these four is correct? Solutions for Ad Hoc Problems in Foobar Solutions for Ad hoc Problems in Foobar So...

@AlanMunn ^^ any ideas? :-)
 
@egreg since I normalise away trailing vowels along with accents, such details like banco v banca are lost on me
 
@DavidCarlisle No? Link?
 
@yo' I'd write it as lowercase (without doing any grammatical analysis on why I'd do that)
 
yo'
12:10 PM
@DavidCarlisle I'll see what the true English speakers think :-)
@DavidCarlisle also, this seems to be the "let's not capitalize short words" track; hoc is clearly a pronoun an shall be capitalized the same way it is
 
@yo' It's only clearly a pronoun if you choose to think of it that way (which few English speakers would) "ad hoc" is treated as an opaque latin word that happens to have a space in the middle for historical reasons.
 
yo'
@DavidCarlisle yeah, almost everybody writes it as ad-hoc which is wrong :-)
 
@yo' but it's no longer treated as a foreign word that you'd write in italics, so I think going back to the original translation to explain the usage is also wrong, "ad hoc" is just being used as an adjective for problem in your example.
 
yo'
@DavidCarlisle yeah, but not hyphenated, like in-depth (which missed a hyphen in my previous comment)
 
12:36 PM
@yo' Here's a real live example of it as Ad Hoc: ebiquity.umbc.edu/_file_directory_/papers/122.pdf
 
yo'
@AlanMunn "it's on the internet, it must be right" approach? :-)
 
@yo' And this book, in which all the paper titles have 'ad hoc' in them, took the same route in its table of contents: books.google.com/…
@yo' I don't have a strong intuition on this, but I think since it's treated as an adjective, then it makes sense to capitalize it. So then the options are Ad Hoc or Ad hoc. And the links just seem to support the idea that others who may have thought about it went for Ad Hoc
 
yo'
@AlanMunn in that case I certainly prefer Ad Hoc as hoc is a pronoun...
 
@yo' Yes, but as @DavidCarlisle said, whatever it is it's not treated as a preposition+pronoun by most English speakers.
@yo' I would go with the UN on this one: un.org/womenwatch/daw/followup/as2310rev1.pdf :)
 
yo'
@AlanMunn the number of examples seems convincing...
 
12:49 PM
@yo' Yeah, nobody ever seems to either not capitalize it or capitalize just the first word.
 
@DavidCarlisle Depends on the publisher (RSC do all foreign phrases italic, so I do too!)
 
yo'
@AlanMunn right. Thanks!
 
@yo' One more for good measure, from a reputable publisher: wiley.com/WileyCDA/WileyTitle/productCd-1118087283.html
 
@DavidCarlisle One to ask Jonathan about in a couple of weeks
 
@JosephWright yep
 
Max
12:59 PM
@DavidCarlisle That chunk logic is almost verbatim from one of your answers ^^
 
@Max quite possibly. answers here are usually the result of at most a few minutes thought. They are not really fully tested production code.....
 
Max
@DavidCarlisle It's a miracle you have enough details of that garish code in your head that you cann pull this off in a few minutes
 
yo'
@Max he's put a lot of effort to ignore reported features of the code that he knows the code now byte by byte.
 
1:49 PM
Hi from SP! :)
 
2:07 PM
Is there a reason why XeLaTeX (and other implementations as well) do not recognize Unicode Character 'ZERO WIDTH SPACE' (U+200B) as a line breaking point?
 
2:20 PM
@wilx \catcode`\^^^^200B=13 \def^^^^200B{\hspace{0pt}} ?
 
@DavidCarlisle Thank you. I wonder if there is a market for a module that would define some of these Unicode fancy characters to something useful like the above. :)
 
@DavidCarlisle \newunicodechar{^^^^200B}{\hspace{0pt}}
@wilx Something similar should be done for non breaking space and so on.
 
@wilx you usually need to do the same with soft hyphen to \let it equal to \- arguably the format ought to do these things but it's a bit hard to gradually introduce luatex support, the format doesn't even default to a unicode font with luatex and xetex
@egreg same thing:-)
 
@DavidCarlisle I add \protected and use less tokens. :P
 
I take it as a yes. This will be my mini-project for today/tomorrow. :)
 
@wilx The team are thinking hard about a range of issues in this area!
@DavidCarlisle All a bit daft, one might argue (though yes I know it's a lot harder than it sounds)
 
@DavidCarlisle That looks quite complex.
 
@JosephWright yes also came up in discussion with someone the other day switching from xetex to luatex (as a trial) complained about lots of hyphens appearing, turned out the source has "unknown" soft hyphens but xetex doesn't show them (but i guess they are characters so in fact break hyphenation) whereas luatex shows them as visible hyphens (which looks worse, but actually gives better indication of what's happened) having the format make it active by default would avoid a lot of confusion....
 
Oh no, it's raining
 
@wilx but you can ignore the actual line breaking algorithm itself (just defer to tex) but the tables list for example that zero width space should allow breaking, but zero width space non joiner should not. So the tables of character properties can be mapped easily to tex definitions of the form above
@PauloCereda I had ice on my car window this morning
 
2:47 PM
@DavidCarlisle oh no
 
@PauloCereda 3.5°C according to the car's reading
 
@DavidCarlisle what do cars know about weather :)
 
@PauloCereda oh perhaps nothing and it was all a mistake and really it was 40°C outside (it was bright sunshine)
 
@DavidCarlisle :)
 
@DavidCarlisle I guess I could add to the Unicode data: despite the font business, the format is setting up a lot of things on a 'expect Unicode' basis
 
2:54 PM
@JosephWright unicode.org/Public/UCD/latest/ucd/LineBreak.txt (I expect we'll change the font eventually as well)
 
@DavidCarlisle Yes, so do I (particular based on recent internal discussion)
@DavidCarlisle I'll look at the line break data later on today
 
3:20 PM
@DavidCarlisle it is good that format doesn't default to Unicode fonts, as users can use alternative fontloader with LuaTeX (Harfbuzz), or suppress them completely (tex4ht).
btw, we recently fixed for XeTeX in tex4ht
and I updated github.com/michal-h21/luatex-harfbuzz-shaper to support LuaTeX 0.95, it was easier than I thought. although text direction and hyphenation are still broken
 
@michal.h21 using harfbuzz by default would be even better:-), but the uses that you mention are specialist ones that could easily be provided with some switch to get teh right behaviour, but the current situation is that 99% of people who try (lua)(la)tex for the first time do so using 8bit fonts and so completely broken hyphenation as the format as only loaded unicode tables (which is fixable in luatex, but not currently fixed in the default setup)
@michal.h21 did you see harfbuzz is getting math support?
 
@DavidCarlisle I only saw some tweets, I am not really following it at the moment. I am also curious about variable fonts, Hans seems to be not too much excited :)
@DavidCarlisle I think I've done some tests with hyphenation for 8-bit fonts. but of course it would be easier for casual users to just use Fontspec by default in the background
 
@michal.h21 luatex you can of course load the 8bit patterns later but xetex you can not and it is currently built with unicode patterns so arguably defaulting to 8 (actually 7) bit fonts is just wrong.
 
@DavidCarlisle ah, I forgot about XeTeX :(
Anyway, it woud be nice if there will be simple way to suppress that default font loading :)
 
3:48 PM
@michal.h21 I really should see if I can build that in my cygwin environment.
 
4:00 PM
@michal.h21 I don't see how that makes much difference: surely you end up simply overwriting stuff anyway
@michal.h21 There's currently no release HarfBuzz/LuaTeX combination, is there, so this is very much a developer thing (as @DavidCarlisle says, if we could have HarfBuzz + LuaTeX this would be excellent)
@michal.h21 BTW, any move to non-TFM fonts would of course mean not loading them as part of the format (you can't build OTF files in)
 
@DavidCarlisle -- actually, i agree partially. what i insist on is consistency. if an author shows a tendency or preference for one or the other (uk or us), i'll settle on that. if there's a mixture, and the author is clearly european, i'll probably settle on uk, but if the author is from the western hemisphere, i'm more likely to settle on us. or -- when in serious doubt, ask; doesn't usually take that long, and authors generally appreciate that you care.
 
4:38 PM
@PauloCereda -- cars do have thermometers, and some even have specialized ones that can determine whether the road surface ahead is at a low enough temperature to be possibly ice-glazed, hence a potential danger. but even better, our car has a radio "weather band" that tunes into the national weather service, so that we can almost always (when we're at home, and mount washington is on line) learn what the weather is on mount washington. almost always makes us feel better about the local weather.
 
5:30 PM
@barbarabeeton Oh my, I had no idea!
 
@PauloCereda -- and it isn't even a new car (2001 model year). but saabs come (originally) from an area where it's advantageous to know when the road surface is freezing, but not apparent by just looking at it.
 
@barbarabeeton That's really clever!
 
@PauloCereda -- trolls know about such things.
 
@barbarabeeton :)
 
@Johannes_B Well known user from the Nordic countries on a question spree
 
5:46 PM
@PauloCereda mine isn't that smart but it has a little snowflake light which comes on below 4C
 
@DavidCarlisle Mine hasn't. :)
 
@PauloCereda perhaps it has, but you have never seen it light up
 
yo'
6:06 PM
@DavidCarlisle you haz a point :-)
My car is clever just in the right way. But my parents' car is a different matter, it's much more clever than me, in a very stupid way.
 
 
1 hour later…
7:36 PM
RIP Neville Marriner
 
@egreg Indeed
 
@JosephWright His last concert was in Padova on September 29. Unfortunately there were no more tickets. :(
 
@egreg I saw it was very recent
@egreg Radio 3 will have something on: I'll of course listen
 
 
2 hours later…
9:23 PM
@barbarabeeton What about us Canadians? :) Our spelling is a mixture. -re and -our but -ize. (but spelt 'i zed e' ).
 
@AlanMunn -- yes, you are definitely a challenge. if i know you're canadian, and i have a question, i will ask. and i have easy access to a former tugboat proceedings and associate editor who is canadian, and is quite happy to set me straight when i goof. (maybe i've just been lucky that the mixture you refer too hasn't happened very often in practice. i suppose now you'll want to put that to the test ...)
 
@barbarabeeton Mainly joking, although I'm sure it must be tricky. Also, I suspect that since Canadians are so bombarded with US English as well, they may be more variable on their spellings. Personally it's one of the things that I try to hang on to, despite being in the US. So my students have dissertation defences. ('c' vs. 's' in nouns/verbs is another difference).
 
9:41 PM
@AlanMunn -- well, i know a couple of canadian residents who were raised in the u.s., and they tend to hang onto u.s. spelling. i think you're the only native canadian i know at the moment who's active with tex and living/working in the u.s.; pierre mackay was canadian by upbringing, but he pretty much adopted u.s. usage. but either situation is easier to handle than trying to make sense of something written by someone whose native language is not indo-european. that can be a real challenge!
 
10:26 PM
Well, the hardest part is to come up with some sensible text for testing the features. The above is not complete yet.
 
@wilx why don't you define them for pdftex?
 
@DavidCarlisle My test shows that they work as intended when I try the test document with pdflatex.
 
@wilx ah utf8 inputenc has set them up already, Ok I suppose, or you could define them the same for all engines.
 
@DavidCarlisle Do you mean for consistency? I added the condition when I noticed because I somehow thought the existing definition might somehow be better. :)
 
@wilx you can see the existing definintions in utf8enc.dfu in latex/base for example \DeclareUnicodeCharacter{00AD}{\-}
 
10:52 PM
I am hungry. :(
 
@PauloCereda Don't treat eating like writing your thesis then.
 
@AlanMunn You are mean. :)
 
@PauloCereda <3
 
@AlanMunn <3
@Alan: I am staying in SP tonight and will return home tomorrow. I will ask mum to prepare a nice pizza. :)
 
@PauloCereda Well there's plenty of good food to be had in SP.
 
10:55 PM
@AlanMunn Indeed, but the quicker I leave this town, the better I will be. :)
 

« first day (2164 days earlier)      last day (2770 days later) »