« first day (1624 days earlier)      last day (3307 days later) » 

8:56 AM
Unanswered questions now at 5496, so we dealt with around 1% of the backlog: good work
Morning @DavidCarlisle
 
@JosephWright just reading the expl3 review in email:-)
 
@DavidCarlisle Yes: I've looked through quickly but need to take time to reply
@DavidCarlisle Unless anyone else gets in first I plan to split one issue at a time
 
@JosephWright I was planning to let you or Frank or someone answer, I really haven't used it in anger enough recently to comment on the stylistic/usability issues
boo, I'm on -20 this morning:-)
 
@DavidCarlisle Cool
@DavidCarlisle I think it's probably me in the first instance then Frank for a big-picture overview
@DavidCarlisle You though did do a lot of the work on this that we still rely on :-)
 
9:14 AM
@JosephWright yes sure but half the argument type letters have changed since then:-)
 
@DavidCarlisle Indeed
@DavidCarlisle One of the things I'm going to mention (how we got to where we are: about the time I got involved)
 
9:25 AM
@DavidCarlisle That's good news! The system realized you have too much.
 
@egreg I doubt it was you that voted for me too quickly and triggered the reduction.
 
@DavidCarlisle I do vote for you, sometimes.
 
@egreg the sometimes being the important point, that doesn't trigger "-60 serial voting was reversed" I also vote for you occasionally (when you make a good answer).
 
@DavidCarlisle That is, almost never. ;-) Do I write good answers?
@DavidCarlisle So Oxford won the boat race.
 
@egreg sometimes, in between all the % and ooalign and dont use \left\right ones:-)
@egreg yes, although there are conflicting loyalties in the house, my wife being born in Cambridge....
 
9:34 AM
@DavidCarlisle This will give M many troubles.
 
@DavidCarlisle Am off out: will tackle e-mail later on
 
@JosephWright OK
@JosephWright and the ctan zips?
 
@DavidCarlisle Yes, this evening I think
@DavidCarlisle Karl asked about them
 
9:53 AM
I may not be able to officially but unofficially I am voting to reopen...
The curly S from that picture is taken probably from Adams - Calculus, a Complete Course or a similar mathematics text book and it's a genuine symbol. Been looking for it for ages and it certainly doesn't want to be found.
I have a large preference for this rendition of the curly S over the mathrsfs version of it....
 
@1010011010 There's no inherent meaning in the shape of the character; it's just what the used font has.
 
@1010011010 google suggests that the pdf for that book is (probably dubiously) available online I choose not to download it but if you did you could see what font it is using.
 
10:42 AM
@DavidCarlisle Nice suggestion. Unfortunately none of these are proper ebooks, as in, they're just black-and-white copies of the actual book. (I downloaded all versions.) I've been skimming through several hundred script fonts and none of them had anything like that curly S. I guess the font is not for the public, or whatever.
 
A one page document (1 page, 22331 bytes) needs 10 seconds on my machine using LuaTeX. This cannot be normal, right?
 
@Johannes_B rebuilding the font cache?
 
@DavidCarlisle The first time took longer today. But now every run. Even pdflatex is slow
real	0m3.133s
user	0m1.788s
sys	0m0.176s
 
@Johannes_B I remember when 15minutes per page was quite common so initial reaction to 10sec per page is "quite fast" but I suppose that's not the reaction you wanted:-)
 
Sounds similar to an issue I had. I fixed it by deleting all the files that luatex created through fontspec. Maybe it helps for you.
 
10:48 AM
@DavidCarlisle I know that LuaTeX is quite slow. but right now, it is pretty annoying to wait 10 seconds. A minute would be ok though, could do something. But a 10 seconds interval is bad.
 
@Johannes_B luatex can go off and index system fonts but pdftex never does, if that is slow it is almost always filesystem related, eg your texinputs is not pointing to a kpathsea hashed directory tree
 
@DavidCarlisle My whole system is quite slow. Sometimes a chat ping here makes the system freeze or half a second. I think i should start at that end :-)
 
11:13 AM
@Johannes_B @Johannes @Johannes @Johannes @Johannes @Johannes @Johannes @Johannes @Johannes @Johannes @Johannes @Johannes @Johannes @Johannes @Johannes @Johannes @Johannes @Johannes @Johannes @Johannes @Johannes @Johannes @Johannes @Johannes @Johannes @Johannes @Johannes @Johannes @Johannes @Johannes @Johannes @Johannes @Johannes @Johannes @Johannes @Johannes @Johannes @Johannes @Johannes @Johannes
^^ 20 seconds. :)
 
@PauloCereda I have never seen my name this often before. :-)
 
 
3 hours later…
2:37 PM
@DavidCarlisle Back from my ride: going to look at that e-mail now
@DavidCarlisle Upload still not right here: suspect something at the BT end as download is fine and my router reports good S/N on the line itself
@DavidCarlisle Will do upload from Norwich at about 9:30 pm this evening
 
@JosephWright have you got the files back from ctan to see what's changed?
 
@DavidCarlisle No, only looked at the checksums, etc. that Petra gave. I get something different to her.
@DavidCarlisle I've checked the files at my end using the Mac native tools, Ubuntu and Windows, and all look OK
 
@JosephWright yes but interesting (slightly:-) to know if the file was truncated or if it's changed internally somehow
 
@DavidCarlisle File sizes agree with Petra's ls output
 
@JosephWright odd, but I guess if it's OK this evening, can just pretend it never happened.
 
2:44 PM
@DavidCarlisle Yes: there's been no change at my end so I think it's a net issue. I'm going to include MD5 values in the upload text!
 
3:24 PM
Running tlmgr update --self --all --reinstall-forcibly-removed for probably the last st of TL2014 updates
Indeed
TeX Live 2014 is frozen forever and will no
longer be updated.  This happens in preparation for a new release.

If you're interested in helping to pretest the new release (when
pretests are available), please read tug.org/texlive/pretest.html.
Otherwise, just wait, and the new release will be ready in due time.
> The pretest for TL'15 will be available soonish.
 
3:59 PM
We are back to 94% answered on stackexchange.com/sites
 
@JosephWright Only ELU has a better performance, among the sites with more than 50K questions (but they close questions rather quickly, I believe).
 
yo'
@egreg yes, they've got the larger answer rate and also the larger closure rate :)
 
4:35 PM
@egreg Indeed
@egreg We're also still top ten for questions per day :-)
 
 
1 hour later…
5:47 PM
@egreg My 'How many mods do people have?' sweep shows we are still doing well there: all of the other sites anything like as big as TeX-sx have at least four nowadays.
 
5:57 PM
@JosephWright Better community. ;-)
 
@egreg Indeed
 
@JosephWright And better mods.
2
 
yo'
@JosephWright we've got only two mods active now quite :) You do a good job in letting the community be involved as much as possible. It's a long-term investment :)
 
6:37 PM
Does anyone know how to reduce the space between the title and the text body in a KOMA article (scrartcl)?
 
6:54 PM
Funny, the only mention of psuedo-lengths in KOMA is wrt scrlttr2.
 
@FaheemMitha looks like \@maketitle in scrartcl.cls has fixed lengths it ends with \vskip 2em for example so you just need to redefine that command
 
@DavidCarlisle Shall I ask a question? I was in the process of doing so, anyway.
 
@FaheemMitha no harm in asking but looking at the source, that's the answer:-) (but I don't know the koma classes at all, so it's possible that I missed something)
 
@DavidCarlisle ok. So I could redefine and adjust the \vskip?
 
@FaheemMitha that's what I'd do, copy the definition into preamble between \makeatletter \makeatother, then adjust to whatever you want
 
7:04 PM
@DavidCarlisle Ok. I'll try. Thanks.
Funny, scrlttr2 has all sorts of adjustable parameters, but the other classes don't seem to have that.
 
7:22 PM
\documentclass{scrartcl}

\usepackage{xpatch}

\makeatletter
\patchcmd{\@maketitle}{\par \vskip 2em}{\par \vskip 0.5em}{}{}
\makeatother

\title{Foo}
\author{Bar}

\begin{document}

\maketitle

Hello world.

\end{document}
 
@PauloCereda can you trust that package?
 
@DavidCarlisle ooh spooky. :)
 
@PauloCereda post it as an answer, you might get some votes from the author of that package:-)
 
People sending HTML e-mails deserve no sympathy. :)
@DavidCarlisle Should I? :) It's a quick fix. :)
 
@PauloCereda as I said to @FaheemMitha above, patching or redefining \@maketitle is the answer...
 
7:30 PM
@PauloCereda I've no idea how that works. Is it magic?
 
@DavidCarlisle Indeed. But my patch seems incomplete, as I do not consider a situation with no author.
 
@PauloCereda blame @egreg, the patch command is obviously defective
 
I don't know this xpatch thingy. So, why don't the other HOMA classes have adjustable parameters? It seems discriminatory.
 
@PauloCereda You're not really using xpatch, because \patchcmd is in etoolbox. So I'm not responsible for any malfunction. :P
 
@FaheemMitha I used a command that searches for a pattern and replaces with something else. I found some potential patch spot in the \@maketitle thingy from the scrartcl def and tried to reduce the current \vskip value. :)
 
7:33 PM
@PauloCereda what happens if it finds multiple matches? Does it replace all of them?
 
@egreg Oh! :) I could take the lazier route and put the x prefix. :P
@FaheemMitha First occurrence.
 
@PauloCereda Oh.
Maybe I should ask my adjustable parameters question as a separate question. Cos nobody seems to care.
 
@FaheemMitha the question of why there are not more parameters? that isn't answerable
 
@DavidCarlisle Well, God might know.
 
@FaheemMitha probably not
 
7:37 PM
@DavidCarlisle Well, he/she probably wouldn't tell us, regardless.
 
@Faheem: do you need only \title in your document?
 
If he/she did, it would be a TeX miracle.
 
@FaheemMitha Letters often require more layout tweaking, kind of doing a layout on a page. Normal documents don't, and there are title pages packages and styles.
 
@PauloCereda correct. just \title.
@StefanKottwitz oh
I like having adjustable parameters.
 
@FaheemMitha How about an alternate solution? Forget \maketitle at all and write your own \mytitle command which basically has some centering, font stuff and your vertical spacing. :) I think it would be easier than trying to "fix" something like \maketitle which technically requires an author and stuff.
 
7:39 PM
@FaheemMitha Letters have addresses, dates, subject lines, various fields, which may be places, so everbody (and institutions) make their own layout
@FaheemMitha Yes, that's what KOMA-Script classes are famous for, many more adjustable parameters
 
@PauloCereda Sounds like a plan.
 
@FaheemMitha The trick is using scrlttr2 for a scrartcl title page. ;-)
 
What is a good rudimentary \maketitle command then? Wrap in a centering environment?
@StefanKottwitz :-)
@StefanKottwitz do you use scrlttr2?
 
@FaheemMitha Yes, I did not often write classical letters but I did always with scrlttr2
 
@StefanKottwitz I see you are using the past tense. :-)
Are your letter-writing days done? :-)
 
7:45 PM
@FaheemMitha used scrlttr2 for 5 job application letters, and I got the job each time. :-)
 
@StefanKottwitz you da man.
 
@FaheemMitha I don't need to do that now :-) Very satisfied with the job.
 
And you've never written letters for any other reason? Not even to the electricity board?
@StefanKottwitz That's nice.
@StefanKottwitz still cruise ships?
 
@FaheemMitha I depends if the occasion would deserve it. For canceling contract I simply took scrartcl.
@FaheemMitha Yes, still cruise ships, but from bird's eye view
 
@StefanKottwitz Ah, you cheaped out?
@StefanKottwitz not sure what that means.
 
7:48 PM
@FaheemMitha I'm employed by an airline, but still do cruise ship networks on behalf of them
 
@StefanKottwitz Oh, I see.
 
@FaheemMitha Life speeded up a bit :-) but very nice projects for great ships in various countries
 
@StefanKottwitz Sysadmin or coding?
 
@FaheemMitha Network design and implementation, so switches, routers, firewalls, ship's networks are as complicated as corporate networks and finely detailed and secured
 
@StefanKottwitz Ugh. Not my cup of tea.
 
7:53 PM
@FaheemMitha Compared with sysadmin things networks are more determined and predictable
 
@StefanKottwitz If you say so.
 
@FaheemMitha I still do sysadmin for web servers and such
 
@StefanKottwitz I see. And TeX on the side? But not in the day job.
 
@FaheemMitha All those windows and application crashes and problems ... and all pretty "closed", with networks configs are pretty open and readable and I can use terminal commands for all :-)
@FaheemMitha Yes, not on the job, just for some network drawings in TikZ where I was too lazy to click that in Visio
 
@StefanKottwitz Funny, all you experts who don't actually use TeX. It's like a paradox of sorts.
 
8:02 PM
@FaheemMitha It's for being compatible with clients and coworkers
 
Hello, a tricky question. How do I check for “next characters” which are unicode (not 8byte), while in pdfTeX. I.e., \@ifnextchar{<em-dash>}{True}{False}?
 
yo'
@Manuel why that? It's very problematic and surely fragile.
 
I thought that my job should be to find which are the actual characters which form the em-dash, and check for the first one (which is not robust, but it should work to me).
 
yo'
@Manuel that is possible, to check for two next characters
 
Problem, is, hoy do I know? I thought, with \detokenize{<em-dash>} I got â as the first character. But it seems not to be that one.
 
8:04 PM
@Manuel you mean the dash has been entered as a utf8 encoded character, not using --- ligature?
 
@StefanKottwitz @JosephWright Can you zap this tag as it would cause a lot of headache later?
 
@yo' Well, I need to check wether the next character is an em-dash or not :)
 
@Manuel in utf8 encoding that is the three bytes E2 80 94
 
@yo' Not necessarily to check for the three, since the first one (I think) is enough.
 
@Manuel It is three characters to pdftex not one
 
8:06 PM
@DavidCarlisle Yep, that's it.
@DavidCarlisle That's what I mean, how did you kno that E2 80 94 are the three ones?
 
@FaheemMitha I wrote an answer. Use a manual backing up. The \@maketitle macro of scrartcl is very poorly written.
 
@Manuel well I knew it was three bytes and could have worked it out but actually I just pasted — into "mixed input" box at people.w3.org/rishida/tools/conversion and read off the utf8 code points
 
@DavidCarlisle Secret society website. Thanks.
@DavidCarlisle So, there's no more “robust” way of checking that the next char is an em-dash than \@ifnextcharâ{True\@gobbletwo}{False}? (I mean, without much trouble.)
 
@Manuel don't do that!
 
There's no problem for me, because I don't use â in my document, but hmm… (what if other unicode “starts with â” hehe).
@DavidCarlisle What do I do then :)
 
8:12 PM
@Manuel there are over 1000 characters with a three byte utf8 encoding starting with that
 
yo'
@Manuel many do start with the same first char of course, utf8 is a prefix code
 
@Manuel don't gobble, look at the next two characters and check they are 80 and 94, and if not put all three bytes back into the input
 
@DavidCarlisle Too much trouble.
 
@Manuel If you did have a â in the document it would not appear as that byte but as the two bytes C3 A2
@Manuel using utf8 input in pdftex is a lot of trouble, you should thank whoever it was who wrote the utf8 decoding macros:-)
 
@DavidCarlisle What? Then â is not the “firs character”? Then what is it?
 
8:23 PM
@Manuel For pdftex, â is three characters (active, by the way), when utf8 is used.
 
@Manuel the first byte of the utf8 encoding of emdash is hex e2 but that is not related to â other than the unrelated fact that in latin1 encoding â is encoded by the single byte e2.
 
Jesus christ.
Okey, I believe you, it's a lot of trouble.
But I don' exacly understand what did you propose before. Definitive question: What would be the best way of checking if the next character is an <em-dash> (the unicode —)?
 
@Manuel in utf8 encoding only characters below 127 (ie ascii, more or less) are encoded as a single byte as "themselves" all other characters are multi-byte sequences in which every byte is above 127.
 
yo'
@DavidCarlisle one of the cleverest concepts ever :D
 
@Manuel It depends on the context, easiest is if you can delay the test until inputenc has decoded the utf8 to an emdash, if that is not possible you need to work harder
@yo' the famous napkin
 
8:28 PM
@DavidCarlisle Yep, I don't need nothing at all. I can delay forever (I think). I have a macro \foo before a paragraph, and I want to check wether the first “character” is the em-dash, or the (two byte) ». That's all.
 
@manuel Look at this
\documentclass{article}
\usepackage[utf8]{inputenc}

\makeatletter
\newcommand{\@ifnextemdash}[2]{%
  \def\@nextemdash{#1}\def\@nextnoemdash{#2}%
  \futurelet\@let@token\@lookforemdash
}
\def\@lookforemdash{%
  \ifx\@let@token^^e2%
    \expandafter\@firstoftwo
  \else
    \expandafter\@secondoftwo
  \fi
  {\@@lookforemdash}%
  {\@nextnoemdash}%
}
\def\@@lookforemdash#1#2#3{%
  \ifnum\pdfstrcmp{\detokenize{#2#3}}{\detokenize{^^80^^94}}=\z@
    \expandafter\@nextemdash
  \else
    \expandafter\@nextnoemdash
 
@egreg I was just doing one that would do that and » :-)
 
@DavidCarlisle Saved some work
 
@StefanKottwitz How so?
@egreg I like your answer. But if I met Marcus, I would be very polite to him.
:-)
 
@FaheemMitha That macro is silly. But I'll fix the answer.
 
8:40 PM
@egreg It's certainly very complicated.
 
@egreg Well, you were faster. I was working with
 
And I agree that leaving space for stuff that isn't there is suboptimal.
 
\NewDocumentCommand \foo { m m m }
 {
   \str_if_eq:ffTF { #1#2#3 } { — } { Em-dash #1#2#3 }
   { \str_if_eq:ffTF { #1#2 } { » } { Guillemot #1#2#3 } { Nothing #1#2#3 } }
 }
 
I like all the adjustable thingys. We should have more of them.
Though I've got mixed feelings about the scrlttr2 ones.
 
@Manuel The problem is that you are assuming that there are three tokens and absorbing {...} groups along the way. No, it won't work.
 
8:42 PM
It would be more intuitive to have relative lengths rather than absolute lengths.
However, scrlttr2 has absolute lengths.
 
@FaheemMitha The lengths are relative (in em units), but, as I said, there are too many of them and out of control.
 
@egreg No, I meant something different (the pseudo-lengths). And relative is probably the wrong term.
The pseudo-lengths in scrlttr2, that is.
 
@egreg True. For me it works (since I'm working in plain text with (almost) no groups), but I know it's not perfect. Yours might be, but I need to spend a few minutes understanding.
 
@Manuel I check whether the next token is ^^e2; if it isn't use the first argument and stop the search. If it is, we know that two other characters follow, so a three argument macro can be used to check the next two. If they are the required ones, then use the second argument or use the first. Finally, deliver the three characters back in the input stream.
 
@egreg Understood. It would be nice a \@ifnextunicodechar in the lines of \newunicodechar :P
 
8:48 PM
\documentclass{article}

\usepackage[utf8]{inputenc}
\usepackage[T1]{fontenc}

\makeatletter
\def\foo{\expandafter\zfoo}
\def\zfoo#1{%
\ifx\UTFviii@two@octets#1%
\expandafter\ztwo
\else
\ifx\UTFviii@three@octets#1%
\expandafter\expandafter\expandafter\zthree
\fi
\fi
#1}


\def\ztwo#1#2#3{%
\edef\tmp{\detokenize{#1#2#3}}%
\edef\ztmp{\detokenize{#1»}}%
\ifx\tmp\ztmp
was a french quotey thing
\else
\expandafter\tmp
\fi}



\def\zthree#1#2#3#4{%
\edef\tmp{\detokenize{#1#2#3#4}}%
\edef\ztmp{\detokenize{#1―}}%
@Manuel ^^^ :-)
@Manuel your version fails if your first word has less than three letters as you drop spaces looking for m arguments
@Manuel why not use xetex then a unicode character is a single token
 
@DavidCarlisle Well, I have a 600 pages document with pdfTeX which I don't really want to accommodate for (Xe|Lua)TeX. Moreover, (an old question of mine in this site), the breaking of words with em-dashes doesn't work correctly.
 
@Manuel that's me:-)
 
Is there an in-line way of dividing an unicodecharacter into the two or three bytes it's composed?
 
@Manuel not sure what you mean. The problem is that it is already multiple bytes and the trick is to combine them.
 
I think I see more future to your solution (the generalization). But egreg's is also straightforward but one needs to know the ^^aa before writing the macros.
 
8:55 PM
@egreg can you suggest a simple \maketitle command without all that extra stuff?
 
@Manuel yes I force the first stage of the utf8 decoding expansion so the first token is a csname that tells you how many bytes are involved, if there is no csname then 1 otherwise 2 3 or 4 and then you can pick up that many bytes, the first byte will always be active
 
@egreg you forgot to say " after the date, present or not". :-)
 
@FaheemMitha I see
 
@egreg your comments imply that it is possible to do better. Would you add adjustable lengths in there or not?
 
@FaheemMitha Of course.
 
9:08 PM
@egreg Ok. I wonder how one would sent wishlist bugs to KOMA.
Hmm, looks like a Sourceforge project.
Surprised anyone uses Sourceforge these days.
 
@FaheemMitha do not say "wishlist bugs" because they are not bugs. :)
 
@PauloCereda that's Debian terminology.
In Debian, everything is a bug. :-)
 
@FaheemMitha which is insanely wrong. :)
 
@PauloCereda Well, nobody's perfect.
 
@FaheemMitha specially those who hear "bug" when it's not a bug. :)
 
9:11 PM
@PauloCereda Debian uses the term "bug" generically, as meaning issue.
 
@FaheemMitha not my call of course but if it were me I'd reject a suggestion to add extra parameters to a well used class. It doesn't add any new functionality, and it adds an incompatibility to previous versions.
 
@DavidCarlisle Sure it wouldn't be incompatible if it had the same defaults?
 
@FaheemMitha my experience of maintaining tex code is that anything breaks something. If your document had that patch command that @PauloCereda suggested earlier, and the class changed to not have \vskip2em but instead have \vskip\aftertitleskip then the patch fails. if you have a new install and have \setlength\aftertitleskip{1em} in your document but send it to somewhere with an older installation the document fails. ....
 
@DavidCarlisle Yes, I see.
I wonder anyone changes anything in TeX code, in that case.
 
@FaheemMitha that is why plain tex essentially never changes and the latex format hardly ever changes, and why longtable hasn't changed for 20 years even though there have been loads of good suggestions (and real reported bugs)
 
9:18 PM
@DavidCarlisle Maybe time for a longtable2?
 
@DavidCarlisle Okey, this is what I've done, and I think it works perfectly
\NewDocumentCommand \foo { }
  { \exp_after:wN \man_dsh_or_rg:n }
\cs_new:Npn \man_dsh_or_rg:n #1
 {
  \cs_if_eq:NcTF #1 { UTFviii@two@octets }
   { \__man_chk_rg:nnn }
   { \cs_if_eq:NcT #1 { UTFviii@three@octets } { \__man_chk_dsh:nnnn } }
  #1
 }
\cs_new:Npn \__man_chk_rg:nnn #1 #2 #3
  {
   \str_if_eq:nnTF { #1 #2 #3 } { #1 » }
    { GUILLEMOT! ~ » }
    { #1 #2 #3 }
  }
\cs_new:Npn \__man_chk_dsh:nnnn #1 #2 #3 #4
  {
   \str_if_eq:nnTF { #1 #2 #3 #4 } { #1 — }
    { EM-DASH! ~ — }
    { #1 #2 #3 #4 }
 
@FaheemMitha you can find some code for a potential longtable v5 if you search around (your favourite vcs:-) but I don't know if will ever be released and if it is released whether it will be longtable version 5 or have a new name
 
@DavidCarlisle Oh. So, not written by you?
Yikes, the KOMA home page is in German!
 
@FaheemMitha written by me but I don't know what a future me will do
 
@DavidCarlisle oh. By "your favourite vcs" you mean Git?
 
9:21 PM
@FaheemMitha it was svn but google shut down googlecode:-)
 
@DavidCarlisle Ask M to rewrite as homework? :) You can ask Frank and Arno. :)
 
@DavidCarlisle oh. bummer.
 
yo'
@DavidCarlisle I still don't see what you all have against git
 
Well, I'm not going to be filing any bugs against KOMA if I can't figure out how to.
@yo' crummy UI.
"KOMA-Script says the best way to get help with its software is by visiting komascript.de";.
 
yo'
@FaheemMitha what is crummy about add, commit and push?
 
9:22 PM
@yo' for a one person project with a single central repository it has no advantages that I see over svn and considerably more complication
 
If you know German, maybe.
@yo' There is a lot more to a DVCS than those commands.
 
yo'
@DavidCarlisle but no real disadvantages. It's like saying that bash is bad because it has many functions while I only need cd and rm.
 
But I certainly disagree with David's statement above.
 
@yo' the last one is a disadvantage:-)
 
DVCS has advantages over centralized even for a single person. To start with you can commit without a network.
@yo' are you a Git user?
 
9:24 PM
@yo' But note I could have moved my projects to somewhere that hosted svn, but moved them to github, I may mutter and curse about it but not enough to actually not use it.
 
yo'
@FaheemMitha yes
@DavidCarlisle :-)
 
@yo' no you need a shell, but people do use smaller lighter shells than bash 9although personally I use bash)
 
@DavidCarlisle Indeed, for that set up it's much the same as SVN (I do very little with siunitx or achemso using Git that I didn't used to with SVN before BerliOS shut down)
 
@yo' for one thing add commit and push is 50% more complicated than add and commit
 
yo'
@DavidCarlisle well, push has auto-complete, so I only do git pus<tab>o<tab>m<tab><CR>
 
9:30 PM
@DavidCarlisle well, you might want to push to different places. or not push at all. Do you never want to modify or merge your commits after the event?
 
@yo' in some git shell? I don't get completion if I type git p... in an xterm?
 
@yo' Does it?
@FaheemMitha Remember @DavidCarlisle and I are used to the SVN model :-)
 
@FaheemMitha as I said above a one person project with a single repository, there is no other place to push
@FaheemMitha no I never want to modify a commit after the event
 
@DavidCarlisle you need command completion / tab completion. Maybe you need to install something. What OS/distribution?
@JosephWright I remember.
 
@FaheemMitha cygwin
 
9:32 PM
@DavidCarlisle There could potentially be multiple places.
 
@FaheemMitha Definitely no joy here, bash on OS X 10.10
 
@DavidCarlisle there might be a bash completion package.
@JosephWright ditto.
 
@FaheemMitha Not really likely, though
 
@FaheemMitha no. It's my project and I can assert with some accuracy there is only one repository with only one branch
 
@DavidCarlisle :-)
 
9:33 PM
@JosephWright I've pushed stuff to as many as three different places. Most of the times I just use two, though.
 
@FaheemMitha Why would one do this?
 
@DavidCarlisle Um, you could still push to multiple places even so.
 
@JosephWright Pure bash or iTerm?
 
yo'
@DavidCarlisle in bash
 
@JosephWright The most obvious reason is multiple backups.
 
9:34 PM
@PauloCereda Mac Terminal so bash
 
@Joseph: You could also try zsh or fish. :)
 
@FaheemMitha I have no idea what use cases you have in mind.
 
@JosephWright no no no get iterm. :)
 
@DavidCarlisle Neither do I :-)
@PauloCereda Nope
 
@JosephWright Get it. :)
 
9:35 PM
The other reason is to have one location that you push temporary commits to, which is sorta private. Then the other could be a more finished/public place.
 
yo'
pushing to multiple places is weird, but pulling from multiple places is a standard thing to do
 
@yo' well i have tab completion in bash but after git p it is still bash command line so how can it know git commands to complete on those?
 
Mercurial has the concept of cset phases: secret/draft/public, if I recall correctly.
mostly draft and public.
@yo' pulling from multiple places won't happen with a 1 person project.
 
@yo' not in a one person project that only has a single repository and you are not accepting pull requests:-)
 
OTOH, pushing to multiple places is a perfectly reasonable thing for one person to do.
 
9:37 PM
@FaheemMitha more likely than pushing to multiple places (as people can make change requests by forking and making a pull request)
 
@DavidCarlisle See below.
@DavidCarlisle Then it is no longer a one-person project. :-)
 
yo'
@DavidCarlisle even there: you can branch your own project if you implement a new feature. It allows you to modify the master branch and still work on the new feature
 
@FaheemMitha yes but if you choose to host on a public site that is out of your control
@yo' as noted in the comment that restarted this permathread I don't add new features:-)
 
yo'
@DavidCarlisle nor you correct any bug, that's right.
 
@yo' I never have bugs to correct
 
9:39 PM
@DavidCarlisle I find that... surprising.
 
@FaheemMitha you have to have faith in your code.
 
@DavidCarlisle Uploads made: fingers-crossed
 
@JosephWright yay!
 
@JosephWright OK... so just gives Karl time to integrate them and push to tl2015?
 
@DavidCarlisle Yes: as soon as CTAN say things have worked I'll let Karl know
@DavidCarlisle If only I had access to the right part of comedy
 
9:44 PM
@JosephWright surely you wouldn't return to the bad old days and just put stuff on ctan;-)
 
@DavidCarlisle Well if we could get hold of Robin ...
 
@JosephWright was a time I had the cambridge node nfs mounted on to my local machine...
 
@DavidCarlisle I am generally faith-deficient. Maybe I need faith vitamins.
 
@yo' @JosephWright seems I am missing this: git-scm.com/book/en/v1/Git-Basics-Tips-and-Tricks
 
@DavidCarlisle Indeed: seems a bit like cheating to me (auto-complete at the terminal is for file names)
 
yo'
9:49 PM
@JosephWright no, it's for "whatever goes there": for examle to me, acroread <tab> offers only directories and .pdf files :)
 
@JosephWright bash auto-complete does command names and environment variable names, but adding command arguments as well is sneaky:-)
 
@DavidCarlisle I'm used to Windows, remember: just file names, thanks
@yo' How can it know? (Same problem as verbatim-in-arguments for TeX)
 
@JosephWright windows without bash seems unthinkable
 
@JosephWright There has a mapping file of sorts. I once read about it. It goes for command parameters as well.
 
yo'
@JosephWright there's way how the software developers can tell to bash
 
10:04 PM
1
Q: Where are the downvotes?

Michael KjörlingIt seems we have a problem. Specifically, a downvoting problem. Early in the private beta, I asked Should this be a high-voting site? which received (relatively speaking) massive community support. A passage from the accepted answer on that question seems to have gone by largely unnoticed, howev...

Erm, people are being sensible?
 
yo'
@JosephWright I couldn't resist:
Content evaluation can be based on upvotes only. Why only all the people believe this SE "be rude" brainwashing? — yo' 15 secs ago
 
@JosephWright I don't see a problem here. Though I'll admit my bias - I don't like downvotes. Except for spam or gibberish.
 
@FaheemMitha Rather my point: my answer to such a question would be 'just vote up for good stuff and leave downvotes for out-and-out issues'
 
If you don't like it, don't upvote it. We only need the right hand side of the number line. :-)
 
yo'
10:06 PM
@JosephWright TeX wins among the older sites -- what a surprise :)
 
@JosephWright Then I don't understand why you posted that question? If you already know the answer. :-)
@yo' Wins in what sense?
 
@yo' We more or less win outright (I think woodworking is a clear outlier)
@FaheemMitha Just pointing up something I spotted
 
@JosephWright there's a woodworking site?
 
@FaheemMitha Apparently
 
@FaheemMitha Joseph is not the author.
 
10:07 PM
@JosephWright You mean that people here don't like to downvote?
It's pretty obvious, I'd say.
 
yo'
@FaheemMitha in all possible "niceness" senses :)
 
@FaheemMitha We've had a feeling from very earlier on that voting should be a positive thing: you get 40 votes per day, so if people vote often most 'good' questions will have a strong positive score
 
@yo' nice is generally good. If possible.
 
@yo' <3
 
@JosephWright sure. Though you don't have to vote if you don't have to.
 
10:09 PM
@JosephWright 40 is not enough. :)
 
yo'
@FaheemMitha note that we're also the best site in size : mod-number ratio, just a coincidence? I doubt it :)
 
Oh, I just realised that that post was on a different site. :-)
@JosephWright apologies for my customary lack of observation.
 
@PauloCereda Noted
@FaheemMitha No problem: one thing you need as a mod is to take comments calmly :-)
 
@yo' Probably not a coincidence. Apparently TeX users are very well behaved. Totalitarian dictatorships would love us.
 
@FaheemMitha I think we are not well behaved. :) Scores are indicators, I guess. And almost all of us does not care for reputation points. :)
 
10:12 PM
@PauloCereda hey, I'm very well behaved. :-)
 
@FaheemMitha Naughty of you to think like that. :P
 
yo'
10:42 PM
gotta go home. Good night, everybody!
 
11:01 PM
hi
can i ask a little question?
in a counter like: \foreach \x in {1,...,10}
 
@EnthusiasticStudent Hello! Sure, ask it!
 
how can I set the pace for counter values?
for instance a pace of two: 1, 3, 5, 7, 9
 
@EnthusiasticStudent You mean, the step?
 
@EnthusiasticStudent \foreach \x in {1,3,...,9}
 
11:04 PM
I am using the counter in a pgfplots legend
 
@EnthusiasticStudent depends, for example \renewcommand\thesection{\the\numexpr2*\value{section}\relax}
 
the counter shows the first two numbers correctly
in the legend
 
@EnthusiasticStudent the default step is 1, but is otherwise determined by the difference between the two numbers immediately preceding the ....
 
for instance: \foreach \x in {0.1, 0.2, .... , 1.0}
it shows in the legend values like: 0.1, 0.2, 0.29999, 0.39998, 0.49997...
 
@EnthusiasticStudent that's a separate issue: floating-point rounding. You can get round it with some number format key settings.
 
11:06 PM
@PaulGessler how?
I thought because the step is not defined in the loop, that mater happens.
 
@EnthusiasticStudent as always, an example of what you're doing would be much easier for me to show you where to make the settings.
 
@PaulGessler so let me work a little and if I do not succeed to solve my problem with the floating point, I will ask you again....
I did not know the problem may be becaue of that floating thing
 
@EnthusiasticStudent not may be, it is because of the floating thing. :-) Try creating a pattern like that with integers and you'll see that it works fine.
 
Thanks... I will work on it...
 

« first day (1624 days earlier)      last day (3307 days later) »