« first day (2926 days earlier)      last day (813 days later) » 

2:55 AM
@DavidCarlisle But as soon as I add \usepackage{geometry}, I get a US Letter PDF. So I assume that geometry is interfering with whatever sets A4 as my default paper size. I then Google to figure out how to get an A4 PDF out of geometry by default, and find that I can use geometry.cfg. I modify this to include \ExecuteOptions{a4paper} and now I get an A4 PDF not realising that this has changed the layout as well as the paper size.
@DavidCarlisle, I pondered more about what tripped me up with paper sizes. As you point out the media size (not layout) is set to A4 for me (presumably this is the same as using tlmgr paper, right?). It didn't occur to me that this might be only media size and not layout. So a simple article based document produces an A4 PDF. So far so good.
3:07 AM
@DavidCarlisle, perhaps as you think about making a warning not about using geometry.cfg you could clarify this. The other thing I have only just realised is that without geometry I get an A4 page even when I explicitly set [letterpaper] on a basic article based document with no other packages loaded. I never tried this before (because I never want Letter paper).
@DavidCarlisle, I realise now that these things are obvious to long time users of LaTeX (although I have used LaTeX for nearly 20 years now—so I'm not that new…), but perhaps the above comments explain why I was not understanding rightly.
3 hours later…
6:35 AM
@JosephWright @DavidCarlisle @egreg You are team members -> latex.org/forum/viewtopic.php?f=30&t=31946
I am not even sure I understand what the user is asking.
2 hours later…
8:08 AM
@Johannes_B and what about me? ;-(. If the user wants to include the content at definition time he will have to get it first with e.g. catchfile, and if he wants to include it at execution time he should use macros to transport the arguments. That input is equivalent to typing is only more or less correct for normal use in documents.
@UlrikeFischer Sorry. I didn't know you are now a team member. I wasn't very active recently.
@UlrikeFischer When did you join?
@UlrikeFischer Just read that. Congratulations :-)
8:44 AM
@UlrikeFischer ooh
Hi @Johannes_B! How are you, mr. penguin?
@Skillmon ooh ducks
@PauloCereda Fine, how are you?
@Johannes_B fine too. :)
9:11 AM
@DavidPurton well the core way (without contributed packages) would be to use graphics or graphicx, hyperref also works. The [a4paper] or [usletter] options set the internam \paperheight and \paperwidth lengths, and then other layout parameters are based off those, but it has no way to set the media size. these days it posibly could as it knows whether it is running in pdftex or luatex and with tex it can make a reasonable guess of dvips,
@DavidPurton but the instruction to specify the media size is system-specific, latex+dvips isn't the same as latex+dvisvg or latex+dvipdfm or back then latex+dvi2ps or latex+dvipsone etc. So the system specific code is in driver files dvips.def pdftex.def etc which are loaded once a driver is specified, typically but not necessarily by the color or graphics packages. These days that is usually automatically chosen by graphics.cfg but....
A: graphicx (or graphics) package distorts format

David CarlisleClassically (La)TeX has no knowledge of the physical paper size, it just sets up a text block area of a give width and height. The "page margins" were just an artifact of where the dvi driver positioned that text block on the page. PDF and pdftex change that a bit as pdftex is essentially its ow...

@Johannes_B we needed someone to take any blame for lualatex font issues.
9:26 AM
@DavidCarlisle ooh
9:45 AM
@PauloCereda, @DavidCarlisle Still nothing on the committee list ...
Everyone must be on holiday!
@JosephWright spooky!
@JosephWright what did you ask them, who wants to be chair? or why are we here?
@PauloCereda No, pretty representative of the usual pattern
@DavidCarlisle Expect a mail
@JosephWright oh
@DavidCarlisle dungeon and sandwiches?
@PauloCereda last two years we have had pizza:-)
9:49 AM
@DavidCarlisle This is the plan this year too
@DavidCarlisle ooh
@PauloCereda pineapples are in season
@DavidCarlisle oh
10:15 AM
typical Brackley car parked at last night's fireworks
user image
@DavidCarlisle ooh
@DavidCarlisle Its owner has no problem with driving on either side of the road.
@egreg true, although as you see it's designed for sunny British weather, no roof needed.
11:10 AM
Hi folks. I'm unclear what the argument to \textlsmeans (for spacing characters), but it seems to be the case that the spaces won't increase beyond a certain amount, regardless what number I put in for spacing. I've looked at the microtype documentation, but it didn't say anything about that.
12:03 PM
12:14 PM
@JosephWright :-)
@DavidCarlisle Pondering it ... do I really need comments, for example
@JosephWright cuts out a lot of worrying about spam if you dont
@DavidCarlisle Exactly
when I converted from google sites originally I took google's "export" of the original site, but that would have taken a bit of work to make usable, so instead I actually just uploaded a wget mirror of the original site. Nice thing about gh-pages is if you don't have the --- data it just leaves it as-is so in fact I had (almost) completely working site within minutes. It gave some confidence and I've gradually cleaned it up, but kept a working site at all times.
@DavidCarlisle Yes, that's how I'm thinking about it too
@DavidCarlisle Don't really fancy the charge for my current host, and to be honest I'm not sure I need the dynamic stuff at all
@DavidCarlisle For a single-person blog, it's probably easier to go static
@DavidCarlisle I'll start experimenting :)
12:23 PM
@JosephWright now it all goes through jekyll but still with a mixture of markdown and html pages. (most of the old ones html, most of the newer ones md) but using same external styling.
@DavidCarlisle Right, makes sense
@JosephWright how many pages do you have?
@DavidCarlisle Think I am going to go this way
@DavidCarlisle About 400
@JosephWright about same as me then I had 360 html I think and a couple of hundred pdf
@DavidCarlisle Ah, cool
@DavidCarlisle GitHub link? Would like to see the history
12:27 PM
@JosephWright you will love my expansive commit log notes, but there is one tag (at the point we switched the dns so it went officially live) github.com/souldern
12:42 PM
@DavidCarlisle I see that Jekyll export actually looks quite easy
@DavidCarlisle By the end of the day, I suspect I'll have something working ...
1:00 PM
@JosephWright You are a developer, so people can contact you through GH issues. :)
@egreg and @UlrikeFischer - congratulations.
1:22 PM
Could someone look at my \textls question? Or should I post it to the site? Is this a known feature, or am I missing something?
@FaheemMitha you didn't really give any indication of the problem, I have never used the command but the image on page 17 of the microtype manual shows how it is supposed to work in combination with \SetTracking
@FaheemMitha Package microtype Warning: Maximum for option letterspace' is 1000`


  \expandafter\textls\expandafter[\the\numexpr\count255*10]{ABC DEF}\endgraf
  \advance\count255 1

@DavidCarlisle It doesn't stretch beyond a certain amount. And I couldn't find any indication of an upper limit in the documentation, what there was of it. Maybe I didn't look hard enough.
@egreg Oh, that's in the log? I didn't see that. It should really return an error. I didn't find anything in the documentation.
Searching again.
@FaheemMitha do you really want more than 1em of letterspacing?
@DavidCarlisle In this case, it appears I do. It's not me, it's the PDF form I'm trying to fill in.
Suggestions for workarounds? And why is there a specified maximum, anyway?
1:36 PM
@FaheemMitha beyond that you are not really letter spacing text just positioning characters by hand, are you trying to fit letters into some equal spaced boxes?
I did a search of the documentation, but it's not mentioned there as such. I did get a hit on the same log that @egreg quoted.
@DavidCarlisle Yes, indeed I am trying to fit letters into some equal spaced boxes. Though in this case, numbers.
@FaheemMitha There is a limitation in pdftex, although not spelled out in the documentation. Try the following
  \letterspacefont\test \tenrm \numexpr\count255*10\relax
  {\test ABC DEF \the\numexpr\count255*10\relax}\endgraf
  \advance\count255 1
@FaheemMitha This issues no warning, but no more spacing is done after 1000
@egreg Some indication on how to use this would be nice. I don't even know what it does.
@FaheemMitha \def\x#1{\makebox[2em]{#1}} then \x1\x2\x5\x8\x9\x0
@DavidCarlisle Yes, I suppose that's another way to go. Thank you.
1:38 PM
@DavidCarlisle hmmm
@FaheemMitha It's a plain TeX file to be run with pdftex
@DavidCarlisle you, sir, are the living manifestation of my compiler optimization techniques. :)
@egreg If no more spacing is done after 1000 regardless, then what is the point of it?
@PauloCereda is that good or bad?
@DavidCarlisle surprisingly, it's good. :)
Very good!
1:40 PM
@PauloCereda as good as @CarLaTeX's nutella pizza?
@FaheemMitha To test whether the limitation microtype warns about is from the package or is “absolute”.
@egreg Oh.
@DavidCarlisle ooh compiler techniques while eating Nutella pizza is good. :)
@egreg So I should run it, and see what the output is? Or what the log is?
@FaheemMitha Both: the output shows that values beyond 1000 are treated as if they were 1000; the log says nothing about it.
1:42 PM
@PauloCereda Aho, Ullman, and Hopcroft's guide to Pizza toppings.
@DavidCarlisle LOL
@DavidCarlisle aka the dragon pizza. :)
@egreg Yes, I see. Is there some way of getting around this, or should one just do something different?
@FaheemMitha do something different, letterspacing was never designed to do what you want it to do,
@FaheemMitha I'm not sure why one would letterspace more than a certain amount.
@DavidCarlisle Alas.
@egreg As I mentioned earlier, I'm trying to fill in this PDF form...
So this spacing limitation comes from some pdftex limitation?
1:45 PM
@DavidCarlisle Argh, of course, I have to worry about {, like in the FAQ
9 mins ago, by egreg
@FaheemMitha There is a limitation in pdftex, although not spelled out in the documentation. Try the following
@FaheemMitha If you are just trying to over-print a form I would use picture mode and position each text block by coordinate.
@egreg I was just confirming...
@JosephWright oddly enough announcements of Church bring and buy sales don't have many { that need quoting.
@DavidCarlisle That's what I'm doing. But doing each character individually smacked of overkill.
1:47 PM
@FaheemMitha not if the form designer thought it necessary to make a box for each digit
@DavidCarlisle ooh syntatic church nesting
@DavidCarlisle Unfortunately form designers often like such things. Perhaps they are just trying to give their lives some purpose and meaning.
@PauloCereda we do have two churches...
@DavidCarlisle ooh context sensitivity
Well, thanks for the help. Dinner time. I should try to remember to actually read the logs. I tend to assume they are just a lot of noise.
1:50 PM
@JosephWright why do we bother:-) ^^^^
@FaheemMitha not that message (which is from microtype) but in general we spend hours worrying about format of log messages. every line of the log has been generated by someone programming something intentionally to make the message, in the hope that it's useful to the reader....
@DavidCarlisle No offense intended. It's just my ignorance, I suppose.
E.g. all those constant overfill warnings which I've got into the habit of ignoring over the years.
@FaheemMitha I don't take offence that easily:-) No more mild amusement (and resignation that no-one reads logs, and with some editor interfaces they are not even aware that a log exists)
@FaheemMitha Warnings from Will that unicode-math is redefining 2000 math commands, you can ignore, but I would never ignore any overfull box warnings.
@DavidCarlisle :)
@DavidCarlisle !!!
@DavidCarlisle Those are gone
@JosephWright I know but since no one looks anyway I thought I'd still use it as an example
@DavidCarlisle :)
@DavidCarlisle Trying a first conversion to Jekyll, will do some clean-up then probably rebase for a sensible log. If it looks OK, I might well move pretty quickly.
2:06 PM
@JosephWright yes I had planned to restart the log with mine too but didn't bother in the end, I might yet investigate retrospective cleanup of it which I've never really done in git
@DavidCarlisle It's quite easy
@DavidCarlisle There's a whole school of thought that clean-up should be a routine part of merging into the release branch
@JosephWright yes but I may experiment on one of my own repos first
@JosephWright isn't that the squash thing? I've done that at the point of merging but not gone back and edited old history. I looked and see it shoudl be easy but I may try on a test repo first:-)
@DavidCarlisle :)
@DavidCarlisle There is that too
@DavidCarlisle You can squash or reword commits during rebasing
@DavidCarlisle Get a feeling I'll spend this afternoon trying to clean up to get to a workable build, so I can then find I don't like it ;)
@JosephWright my main worry (if I get others to contribute to the site) is the editor interface (for people not checking out locally) it's sort of Ok for simple md files but unlike they one they have for wiki and issue posts you can't drag images etc which makes it a bit technical. There are some online editors for github pages available but they all seem to want scary amounts of permissions granting, so I have not tried them.
@DavidCarlisle Totally agreee: for a multi-person site with non-experts, WordPress is a much better bet. But that's not the case here.
2:38 PM
@DavidCarlisle I do look at them. Usually when there is an error. If you want to force someone to look at a log, issue an error.
That isn't a recommendation, by the way. Just an observation.
@DavidCarlisle Why can't one ignore overfull box warnings? I do so all the time.
@FaheemMitha it makes no sense to ignore them:-) there is a user-settable limit on how bad you want the over-run to be (\hfuzz) so if the box is overfull more than hfuzz tex is warning you that the output is worse than you deem acceptable. so ignoring the warning makes no sense, either fix it, or increase hfuzz so it doesn't warn, and they you do see the warnings for the really bad cases that are worse than your new setting.
A: Do I have to care about bad boxes?

David CarlisleIt is worth noting that TeX doesn't make the value judgements here. The user, or more likely, the class file on behalf of the user, has set constraints on the amount by which boxes may overflow, the amount of stretching allowed on short pages etc. TeX only warns if these user-set constraints are...

@DavidCarlisle I see. Thanks for the education. Is increasing the hfuzz limit something people commonly do?
@FaheemMitha I do it in some cases of automatically generated text where I know there are tricky tables etc that might not fit and I choose not to care. If i was writing a book I would be more likely to reduce hfuzz, and fix all the cases by hand.
@DavidCarlisle Reduce vs increase?
I'm also fuzzy on how one fixes overflow. Which I suppose is part of the reason why I have never tried.
@FaheemMitha yes any overfull box is bad, so if I was spending months writing a book using hand written tex markup, spending a couple of days at the end tweaking the line breaking to avoid all issues is time worth spending. If it's an automated production where you want no hand editing in the file, then it is different.
@FaheemMitha it means a word is sticking in to the margin, so you either need to allow it to stretch white space more, or add extra hyphenation point (\-) or if necessary re-word the text.
2:50 PM
@DavidCarlisle Is allowing more white space stretching something to set locally or globally?
@FaheemMitha if doing very final cleanup in a book, locally, perhaps adding \sloppy for one paragraph, or simply adding an extra \hspace around some tricky non-breakable word in the paragraph or ...
@DavidCarlisle So it depends on the situation?
@FaheemMitha yes, if it was automatically fixable tex would have fixed it, it only warns as it could not find a good solution in this case:-) Of course, if you are doing local fixes on every paragraph that is perhaps a hint that the global settings are wrong for this document, but there are no general rules.
@DavidCarlisle I see.
3:18 PM
@DavidCarlisle Working on the conversion ...
@DavidCarlisle josephwright.github.io ... no actual index yet, so you have to find pages directly :(
3:30 PM
@JosephWright I just tried out prose.io (using the machine account so I don't have to worry about permissions on my account) looks OK actually, I may recommend it if anyone wants to edit md files on the site, has usual shortcuts for lists etc, but in particular has a file upload for images which adds the image markdown and checks the image in to the current directory at the same time.
@DavidCarlisle Ooh
@DavidCarlisle Might be handy
@DavidCarlisle I'm working on cleanup of my converted files, then have to look at layout
@DavidCarlisle Think I will go for this: I don't fancy spending the hosting money, and I really don't need WordPress for what I do
@DavidCarlisle Have at least some kind of index now
@JosephWright interesting title positioning:-)
@DavidCarlisle Er, yes: it's a work in progress!
@DavidCarlisle Want to get all of the cleanup done, then will start on layout
@DavidCarlisle I have about a month before I have to pay current hosts: aim to be sorted before then
@JosephWright sounds like a good plan.
3:58 PM
@DavidCarlisle Cleanup going OK :)
4:17 PM
@ShreevatsaR -- the "missing" errata listing for tex+mf (1983) has been scanned and posted. you may already know about this, since two upvotes appeared almost instantaneously.
5:04 PM
@barbarabeeton Wow, thank you... So glad this still exists, and that now I can read it! (For anyone else seeing this comment: the post is here: tex.stackexchange.com/a/458351)
2 hours later…
7:34 PM
% This old errata list from SAIL was dated 10 Mar 83, Sections marked with a $ in this
% contents list have been removed.
% The file has been destroyed after printing this copy.

00002 This is a list of all errors in the September 1978 TEX user manual that
00003 Extensions to TEX made since the November printing of the manual:
00004 Corrections noted since the November 1978 printing:
00005 Important changes made to TEX on February 25, 1979:
00006 New extensions to TEX subsequent to the April printing of the manual.
@ShreevatsaR @barbarabeeton ^^^^
@DavidCarlisle -- did you manage to read/copy that from the posted scan, or did you have a copy of the actual file? anyhow, can you send it to me (or post it somewhere i can pick it up reliably -- cut-and-paste isn't totally reliable for me). then i'll have karl post it in a more readable form than what's posted now. (and thanks.)
@barbarabeeton I ocr'ed the scan (google docs OCR) I'll mail it to you; it may want a proof read:-)
@DavidCarlisle -- great! thanks! (proofread is to be expected. one thing i'm sure will be needed is the insertion of ctrl-L's in appropriate places.)
@barbarabeeton a response of "I'm sure your typing will be perfect" would have been better than "proofread is expected" :-)
@DavidCarlisle -- well, i know the font is reasonable, compared to common sans serifs, but we do want to avoid "keming".
7:43 PM
@barbarabeeton actually google docs added ^L but I took them out (control characters in text files are a pain these days) I also re-arranged it as single column (mostly for ease of OCR avoiding getting interleaved text)
@barbarabeeton you should have mail
@DavidCarlisle -- thanks for single-columning it. all i had was the printout that's posted as a scan. the page breaks are significant in this (i did read it cursorily), so some way should be found to indicate them, short of making it into a tex file. suggestions?
@barbarabeeton I'd insert ================ or **************** or some such.
@DavidCarlisle -- okay. reasonable. (my usual convention is a string of 72 % signs, but i'm not sure everyone else would appreciate that.) equal signs seems more conventional.
2 hours later…
10:08 PM
@DavidCarlisle I found the setting that changes the precision in the pdf: \pdfdecimaldigits. If I set it to 4 I get your output. What is your default value? On my system \the\pdfdecimaldigits outputs 3. @JosephWright.
@UlrikeFischer Ah, that might be it
@UlrikeFischer Didn't we set that to something sensible?
@JosephWright Don't know. I found an older discussion (tug.org/pipermail/tex-live/2016-January/037673.html), and it sounds as if 3 (as I have) should be the default, so the question is why @DavidCarlisle seems to have 4.
Say, you have an optional printout of a counter and its setting to 0, in the optional arg of a command.
Let this counter get incremented with each call of the command by 1 or as some function of the last arg of the cmd.
Would it be possible, with this configuration (order of commands?) to get the printout including all increments of the last arg?
In other words, does everything get executed the moment it's read, or does the whole command get parsed, and then executed in some order? Are optional commands implicitly always first?
that is, is it better practice to rewrite the whole command rather then trying to figure out how this would work
10:34 PM
@UlrikeFischer doesn't miktex use the standard (@JosephWright's) tex-ini-files collection? pdftexconfig.dat has:
% Low-level settings unlikely every to need to change
compresslevel    = 9
decimaldigits    = 3
pkresolution     = 600
horigin          = 1 true in
vorigin          = 1 true in
@UlrikeFischer oh you said I had 4, let me see...
$ pdflatex '\showthe\pdfdecimaldigits\stop'
This is pdfTeX, Version 3.14159265-2.6-1.40.19 (TeX Live 2018) (preloaded format=pdflatex)
 restricted \write18 enabled.
entering extended mode
LaTeX2e <2018-12-01> pre-release-6
> 4.
<*> \showthe\pdfdecimaldigits
@UlrikeFischer ^^^ hmmm
$ ls -l  /home/davidc/texmf/tex/latex/config/pdftexconfig.dat
-rwxr-xr-x 1 davidc Domain Users 545 Jan 19  2016 /home/davidc/texmf/tex/latex/config/pdftexconfig.dat
@UlrikeFischer I may have been testing something back then..... Sorry for teh false alarm I'll rebuild my formats
@vlg sorry I can't parse that sentence, optional arguments don't have to be first eg \usepackage[leqn]{amsmath}[2018/01/01] has one option at the start and one at the end
10:53 PM
@UlrikeFischer oh that discussion has the same date as my non standard config, I really should stop testing answers:-)
@DavidCarlisle Oops
$ find . -name \*diff
@UlrikeFischer ^^ the hycolor ones are a space before the (...aux) line, pagegrid xetx did you not update the tlg (it is showing tikz output still)
embedfile3 is this a windows line end diff? ! /Length 66033 <> ! /Length 63842 >>
@UlrikeFischer @JosephWright and the soulutf8 ones are also white space at the start of log line differences. (I disabled the new utf8enc.dfu to avoid diffs from that)
@DavidCarlisle Hmm
@JosephWright Ulrike got space before the page counts in her tlg:
*** 531,537 ****
  Package soulutf8-test Info: * Version: soul + patch on input line ....
  Package soulutf8-test Info: * Version: original soul on input line ....
  Package qstest Info: Passed: string-L-driver-soulH on input line ....
!  [24]
  LaTeX Font Info:    External font `lmex10' loaded for size
  (Font)              <14.4> on input line ....
  Package hyperref Warning: Token not allowed in a PDF string (Unicode):
--- 531,537 ----
  Package soulutf8-test Info: * Version: soul + patch on input line ....
@DavidCarlisle Strange
11:10 PM
@JosephWright I don't have space before [24] even in the raw log
Package soulutf8-test Info: * Version: original soul on input line 315.

Package qstest Info: Passed: string-L-driver-sy on input line 315.
Package soulutf8-test Info: * Driver soulH on input line 315.
Package soulutf8-test Info: * Version: soulutf8 on input line 315.
Package soulutf8-test Info: * Version: soul + patch on input line 315.
Package soulutf8-test Info: * Version: original soul on input line 315.
Package qstest Info: Passed: string-L-driver-soulH on input line 315.

LaTeX Font Info:    External font `lmex10' loaded for size
@DavidCarlisle Mumble ...

« first day (2926 days earlier)      last day (813 days later) »