« first day (3149 days earlier)      last day (69 days later) » 

12:06 AM
@marmot You mean it's not fun?
@AlanMunn It might be just me but one could get the impression that the removal and an almost removal of users is correlated with recent activities in comments on meta.
@marmot Well given that I don't know who the other person is, it's hard to tell.
@marmot On thing that always puzzles me is why brand new users with very little connection to the site post on meta to begin with.
@marmot We've had this happen before. I find the behaviour really odd.
@AlanMunn Multiple posts of the removed user were mentioned in a comment just a bit before the user was removed, but this may of course just be coincidence. It's just two coincidences within 24 hours, which could be of course also just a coincidence.
@AlanMunn I guess that it is rather normal if users retire, i.e. are no longer posting answers, and a bit less so if they delete their account. I have no idea about all this, and am not a member here for too long but I do feel that things are a bit strange in that now the frequency of such events seems to be unusually high.
@marmot Yes, it does. I don't quite see the point of deleting, to be honest, even though I can see why people want to stop participating. It's a kind of weird protest really.
@AlanMunn I deleted my account on TeXWelt.de just a few days ago. And I seriously consider deleting my physics account, too, if that is possible.
12:22 AM
@marmot Although I did delete a bunch of answers of mine on ELL because they were downvoting perfectly good answers if they thought the question wasn't good, or if there was a duplicate.
@marmot But what's the difference between deleting it and just stopping participating?
@AlanMunn I didn't get downvoted. The difference on TeXWelt.de is that you get emails if someone comments on your posts, which you do not get if you quit.
@marmot But can't you turn that off? I get no emails from here. Isn't the system basically the same?
@AlanMunn It is probably possible to turn them off, but the only thing I saw was a button "Turn emails off for this post", which is too cumbersome to do for every post.
@marmot I see. What a strange default setting to have. They should be off unless you turn them on.
@AlanMunn It might very well be my ignorance. I am not very good at these things. In any case, they will do fine without me. ;-)
12:28 AM
SOURCE: http://tug.org/utilities/plain/cseq.html#edef-rp
QUOTE: Not quite everything in the replacement text is expanded at definition time [in \edef]. For example, token lists produced by \the and a token following \noexpand are not expanded.

Since token list \count255 in the example below is produced by \the and such token list (according to TeX reference) is not expanded inside \edef, why is it actually being expanded (to 3)? I would expect the console output produced by \show to be something like \z=macro:->\the\count 255\the\count 255, but it's not (it's 33, so everything in \edef aft
@AlanMunn I should also say that I was first hesitant to join since IMHO it does not make too much sense to have such a site in German if it was only for the language. You exclude many just because they are not able to communicate in the language. But then I saw that this site was very tidy, i.e. the fraction of "not-so-well-posed" questions was much lower, so I thought it might make sense to hang out there.
By the way, what's the difference between \count255=3 and \count255 3? Is it just one of several ways of doing the same thing?
But then there was a meta question, and the winning answer was such that I felt that it would completely annihilate all the reasons I was excited about this site, so I quit.
@bp2017 the token list produced by \the\count255 is 3 and that is not further expanded, of course it's not expandable anyway so it makes no difference fro count registers, that condition of \the only makes any difference for toks registers (as it is naturally true for all other types)
@bp2017 a single = and a single space token are optional in assignments
@marmot I couldn't find meta at all on the site.
:50688521 :D quick save that.
12:38 AM
@AlanMunn Meta question have a "["Meta"]" in the title just without " (grrrrr;-)
@marmot Ah, ok. So there's no separate site. I see.
@DavidCarlisle, oh, I see. So if \the\zOuter is expanded to \zInner while \zInner can farther be expanded, then \edef doesn't expand \zInner? So \the acts as expand once?

As for space and equals sign being optional, some assignments can't be done without them. Doesn't it imply they're mandatory in those assignments (but optional in a sense that not all assignments require them, e.g. \count1\count255 needs no space or equals sign).
@AlanMunn Nope. But this does not mean that users are less engaged, in particular those who say they have no time.
@bp2017 in edef yes, in fact the \unexpanded primitive in etex is (internally) essentially the same as putting its argument in a toks register and then using this form of \the expansion.
@DavidCarlisle, thank you (yet again).
12:50 AM
@bp2017 if you insist on using literal numeric count registers you may need a space to terminate the number, but an additional space and = are optional in assignments however additional spaces are optional for count assignments as leading spaces are ignored in a <number> so all these work:
\count255 \space 3

\count255 \space\space 4
\count255 \space\space\space 5

\count255 = 6

\count255 = \space 7
If you are doing a token assignment \let\foo=\bar then exactly one space token is optional after the =
4 hours later…
4:24 AM
I've heard of the TRIP test for TeX. Has this been done for LuaTeX?
2 hours later…
6:03 AM
Yesterday a user who loved me very much was removed :) ^^^
@CarLaTeX Yes, I lost 2505... ;-)
@marmot So they loved you more than me!
@CarLaTeX Or I wrote more answers ... ? ;-) (Sad that they left...)
@marmot Of course, nobody can reach your amount of answers per day!
6:27 AM
@CarLaTeX The user who left wrote more answers per day, at least in a certain period.
@marmot I'm so sorry he left, first of all because I could have eased his leaving due to a comment of mine :'(
6:46 AM
@CarLaTeX It is not Mico who left.
@marmot Oh no! Another one! I hope not be involved in this case
@CarLaTeX Why would you?
@marmot I don't know, even in the case of @Mico my intentions were good but probably I can't explain myself well in English
@CarLaTeX It is not at all clear that you were involved in this in any way.
@marmot The reason for his leaving is the downvote to this post: tex.meta.stackexchange.com/a/8327/101651 but my comment could have been rude (see my comment under the answer)
@PauloCereda Happy quacking birthday!
7:02 AM
@CarLaTeX If you didn't downvote him, your actions were not the reason: tex.meta.stackexchange.com/questions/8025/…. Anyway, luckily he didn't leave.
@marmot Is he still here! I'm so happy! No, the downvote is not mine, when I downvote I usually leave a comment, unless it is clear without any doubt the reason for downvoting or unless someone else left the same comment I want to leave (in that case I upvote the existing comment).
1 hour later…
8:19 AM
@DavidCarlisle I see you lost points too. Now you can try to hit the same palindrome (509905) again!
@HaraldHanche-Olsen ah good plan:-) do we know who left?
@DavidCarlisle I certainly don't. Of course, you need to correct your reputation mod 10. Should be easy. I can upvote one of your questions. Oh, wait …
@HaraldHanche-Olsen it would also work to get 15 if @egreg accepted one of my answers
8:36 AM
@DavidCarlisle :)
@CarLaTeX thanks <3
samcarter isn't around anymore. I haven't seen her for a while in chat. Is that old news, or is she the recently deleted user? (Sorry, I can't keep up with events all the time.)
@PauloCereda Oh, it's your birthday today? Happy birthday. I'll refrain from eating duck, then. Just for today, mind you.
@HaraldHanche-Olsen it is, thank you. :) And thanks for not eating duck today. :)
@DavidCarlisle if you check your last answers on meta you should be able to see it ...
@PauloCereda Happy Birthday! And here come the reserve Ducks:
@UlrikeFischer thanks, Ulrike. <3 ooh these are different ducks
8:51 AM
@PauloCereda should I do it instead?
@UlrikeFischer ah.
@DavidCarlisle ;-).
@DavidCarlisle oh no
@DavidCarlisle you are mean
May 30 at 9:04, by Paulo Cereda
@DavidCarlisle you are mean
9:08 AM
@DavidCarlisle I tried yesterday a bit the idea to use \mark. While it worked at the end for the simple example I think there are lots of pratical problems: the code must work also in boxes, with all pagestyles and there can be (many) more than one BDC-entry on a page.
@DavidCarlisle oh
@UlrikeFischer get Joseph to add a \nonimmediatepdfpageresources primitive?
@DavidCarlisle ooh
@UlrikeFischer but once you have a page hook can't you just put the \pdfpageresource call into the hook for the next page, without having to go via a write?
@DavidCarlisle it is not only \pdfpageresource, \pdfpageattr has inherently the same problem.
@DavidCarlisle yes, but the problem is that I don't know how to collect what to put into the argument.
9:19 AM
@UlrikeFischer you could use \write and pick it up on the next run... :-)
@DavidCarlisle But it is a bit of a pity as it seems to be needed only for pdflatex.
@UlrikeFischer non immediate forms of the primitives really would be a possibility....
@UlrikeFischer or I did start an emulate-pdftex-with-luatex format at one time....
@DavidCarlisle yes. I think I will try to make a real example with \latelua to check if the collecting really works, also with more than one BDC on a page and then I can ask @JosephWright.
9:40 AM
@DavidCarlisle You have never given good answers to my questions
@egreg Could it be because you don't ask good questions?
9:58 AM
A short question about the licensing of Computer Modern: The files say e.g. `THIS IS THE OFFICIAL COMPUTER MODERN SOURCE FILE cmbase.mf BY D E KNUTH. IT MUST NOT BE MODIFIED IN ANY WAY UNLESS THE FILE NAME IS CHANGED!`
So would it be ok to distribute changed file when only the extension is changed, e.g. as `cmbase.mp`, or does the "basename" has to change?
@AlanMunn ^^
Also Sue Perkins
@MarcelKrüger that licence is tricky actually (and not accepted by the open source licence rules) (we had to remove the "name change" rule from LPPL) debian legal essentially finesse it by claiming texts elsewhere mean that the files are public domain. The important thing is to make sure any generated tfm have different names, so changing the base name is safest really....
@PauloCereda musical recommendation from egreg I assume?
@DavidCarlisle surely :)
@DavidCarlisle It would be nonfunctional changes anyway, so while using them would generate the same TFM names, it would also not change the TFM file content, so that shouldn't be a problem.
@DavidCarlisle The English, the English, the English are best? :)
10:19 AM
@MarcelKrüger to be honest I wouldn't set up modified sources that generate the same tfm names it just puts an extra burden on the distributors to check the files really are the same. What are you up to anyway? :-)
10:30 AM
@DavidCarlisle I am in the process of making some code ready to publish which is about using MetaPost for dynamically generated fonts in LuaTeX. Here especially cmbase.mf is very useful because it provides the primitives to get characters compatible with the Computer Modern glyphs. Anyway, cmbase.mp uses a lot of constructs like z@0e--z@1e..z@2e which are read as z@0-z@1e..z@2e by MetaPost, leading to ~chaos and destruction~ "interesting" error messages and weird output.
This can easily be fixed by adding some spaces, but the changes are so small that I do not feel like a new name is really appropriate. Especially since the changes might no longer be necessary after MetaPost is fixed...
@MarcelKrüger if it were any other file and any other author, I'd say just drop the author a line and get the changes pushed upstream, but as it is I'd say change the name (unless you do want to ask Knuth, he might agree...)
@DavidCarlisle Thanks, I will probably ask Knuth and push a bit more to get MetaPost changed.
10:48 AM
samcarter is still around, but only on Stack Overflow (profile).
11:32 AM
@Werner I see. But when she deleted her account on tex.se, presumably all her votes there were removed too?
@HaraldHanche-Olsen her account is not deleted: tex.stackexchange.com/users/36296/user36296
11:53 AM
@UlrikeFischer Oh. Good news and bad news there. (Ouch.)
12:06 PM
@HaraldHanche-Olsen No question of mine has received less than 50 upvotes
12:17 PM
Would it be correct to say that a \directlua call cannot be in the preamble if it writes out, namely calls to tex.sprint, for example?
And would it make sense to divide Lua code into code to include the preamble and code to include in the document?
Though actually, if it is calls to tex.sprint in a function, perhaps that is Ok?
12:43 PM
@egreg True. And totally undeserved, in every case.
@FaheemMitha Hmm? So long as it prints tex code that is allowed in the preamble, I don't see why not.
@FaheemMitha As a matter of code organisation, I prefer to have one or more lua files to be loaded in the preamble, in which any complicated behaviour is encoded in lua functions. Then any occurrences of \directlua in the body should basically be simple calls to functions defined previously.
@HaraldHanche-Olsen Yes, that sounds reasonable. I forget, do you use LuaTeX?
Nomenclature question. If I had a style/package file called foo.sty, but it was rewritten in Lua and the interface changed in incompatible ways because of that, should I go with foolua.sty or foo2.sty?
I'm inclining towards foo2.sty.
@FaheemMitha From time to time, yes. Not an expert yet, to put it mildly.
@HaraldHanche-Olsen What do you use it for?
@DavidCarlisle you are so modern ... Joseph will be pleased.
1:00 PM
:50691945 No strong opinions on that. I might want to include lua in the name if using the package required writing lua, but not if the use of lua is merely an implementation issue, with the package interface still being tex commands.
@FaheemMitha My first use was to patch some fonts for a book I was typesetting. Lately, I have been experimenting with using it to draw phase diagrams for systems of differential equations. I wrote a simple Runge–Kutta solver in lua, and wrapped it in a lua function that will produce coordinate pairs to be consumed by pgfplots.
@FaheemMitha Sample output on the last page of math.ntnu.no/emner/TMA4165/eksamen/tma4165eksV19lf.pdf
@HaraldHanche-Olsen Oh. So does that work, then?
And is that research or teaching?
Ah, apparently teaching.
@FaheemMitha Yup. That doesn't mean it can't be used for research.
@HaraldHanche-Olsen Fancy.
That wacky diagram was created by LuaTeX?
@FaheemMitha Indeed.
1:06 PM
@FaheemMitha :) I am kind of pleased with it. But I've gotta run. We're packing the car. Starting vacation early this year, to make up for lost vacation last year. (I was ill and had no vacation to speak of.)
@HaraldHanche-Olsen Ok. Have a good vacation.
Where are you going?
@FaheemMitha Thanks. I'll probably be hanging out in chat now and then, especially on rainy days.
@FaheemMitha A small city called Kragerø, in southern Norway (ask your favourite map). We have a cabin there, not in the city, but on an island outside it.
@HaraldHanche-Olsen Sounds sylvan. I hope you can get a net connection there.
Country vacations in your part of the world make me think of the Moomins, for some reason.
And I know that's Finland, not Sweden.
@FaheemMitha Yeah, called mobile broadband. My previous router died on me recently, so I bought a new one. Electricity is solar, kept on a pair of 12 V lead-acid batteries.
@HaraldHanche-Olsen Nice.
1:11 PM
@FaheemMitha Indeed. My mother always used to read those books for us kids when we were on vacation.
Some of my favorite books, when I was a child.
@FaheemMitha We collect water from the roof, except drinking water. That we haul out from the city.
@FaheemMitha They're fantastic.
@HaraldHanche-Olsen That sounds like a lot of work.
@FaheemMitha Not really. We fill from faucets just at the water's edge, lift 20 liter cans into the boat, and out again at the island. We have to carry the water about 30 meters in total.
@FaheemMitha And the water from the roof goes straight into a cistern. Tap water!
@HaraldHanche-Olsen Ok.
1:14 PM
@FaheemMitha But now I really have to move. So long.
@HaraldHanche-Olsen Take care.
@UlrikeFischer just pandering to him so he feels motivated to add nonimmediatepageresource
@FaheemMitha no
@DavidCarlisle Ok.
@FaheemMitha You can tex.print in the preamble any tex commands that are allowed in the preamble.
@DavidCarlisle Ok, noted.
1:26 PM
@FaheemMitha which is same as anywhere else, you can only tex.print commands in tabular that are allowed in tabular.
@DavidCarlisle Right. So TeX commands that don't produce output.
@FaheemMitha if you think of it that way. In any case there is nothing special about the interaction of \directlua and the preamble.
@DavidCarlisle Ok.
2 hours later…
3:36 PM
Anybody know why \textsubscript{123} adds vert space, whereas $_{123}$ doesn't? This was measured with screencaps and overlaying in ms paint. The funny part is that the math font's nums are actually bigger in size than the def roman.
@vlg we would know if you provided an example.... (it's not generally true, it depends on the fonts in use)
well \linespread{0.8} is more or less requesting that....
Why would it not apply to\textsubscript as it does for math's subscript?
@vlg just checking now, you don't need images, the log shows the effect
3:49 PM
You were right that is was the linespread, however I don't know why. If I knew were to look for the textsubscript command, that could provide insight or more confusion

shows that you get
...\hbox(6.5625+2.91159)x345.0, glue set 289.5127fil []
...\penalty 150
...\glue(\lineskip) 1.0
...\hbox(7.07031+2.39256)x345.0, glue set 0.62704 []
That is, the depth of the line with the subscript is 2.9pt and the height of the following line is 7pt so the closest these can get is 9.9pt but because of the linespread you have specified an unachievable baselineskip of 9.6pt
@vlg textsubscript is defined in latex.ltx but the definition isn't that important it doesn't itself add vertical space, tex's linebreaker adds the vertical space just because the lines do not fit
presumably with the math version it fits, let me see...
The linespread is to make it more condensed, for a personal dictionary of sorts and I wanted uniformity in all rows.
yes with the math version the depth is slightly less and you get lucky:
...\hbox(6.5625+2.39256)x345.0, glue set 287.234fil []
...\penalty 150
...\glue(\baselineskip) 0.13716
...\hbox(7.07031+2.39256)x345.0, glue set 0.62704 []
sheer luck
@vlg sure but you can not have uniformity if you specify a baseline spacing that is smaller than the text (well actually you can have that, if you set \lineskiplimit to -10pt it will over-print such clashing lines giving even baseline spacing at the expense of readability.
4:01 PM
Other than \linespread and \lineskiplimit is any other way of reducing white space between lines? Or a preferred one?
@vlg well that's OK (although better really just define \normalsize to use whatever size you actually want rather than just scaling the amount by a random factor) but how you change it is not the issue, it is what value to change it to that matters:-) You need to look at some typical text including subscripts and accented capital letters (if you have those) and anything else that you want inline and set the baselineskip such that it is bigger than the height+depth of that text.
@DavidCarlisle I'm already using \footnotesize but I thought that the distance between lines should stay proportional, i.e., would still appear to be too much if using a short and shout font.
@vlg that is, \sbox0{\'{A}BC()ygfoo\textsubscript{123} l' } \showthe\dimexpr\ht0+\dp0 \relax you want you baselineskip to be more than the value shown
@vlg it is perfectly reasonable to set some fonts with tighter baseline than others (I haven't actually looked at the output of this one:-) but it's not reasonable to make it so tight that the text does not fit in the baseline-to-baseline space.
@vlg if you add that test line to your sample document then you get > 10.99753pt.as output which means that if you make the baselineskip less than 11pt you are likely to get inconsistent spacing. The default is 12pt so you can make it a bit less but 9.6pt is rather extreme
8.6787pt is what I get. If I recall correctly, the at 10pt normal in article footnotesize is 8pt, so whatever val .8 linespread is, it should be more
4:42 PM
@DavidCarlisle I have been dump, I don't need to know the content of the resources when a page is shipped out. I can simply reserve an object for every page and fill them at end document (when the aux is read in again or something like that).
5:06 PM
@UlrikeFischer Interesting... I haven't been active for a bit - focused on something different - so it's odd to see these things happen. Is there any more detail?
5 hours later…
10:08 PM
Good news everyone, user Mico (tex.meta.stackexchange.com/users/5001) will not delete his profile! --> see comments on tex.meta.stackexchange.com/questions/8324
@CarLaTeX and others

« first day (3149 days earlier)      last day (69 days later) »