@manooooh They are your friend's friends on FB and saw you started your friendship on FB

@CarLaTeX and why people do that? I mean, we did not talk much, we saw each other but not much else. I do not know what they earn with that. And saying that maybe they want to talk again is not justifying. Did it ever happen to you on fb? (remember, the abbreviations with lower case!! :D)

@CarLaTeX btw, how are you?

@manooooh Someone competes to have more friends on fb. Ops, I forgot the lower case! I'm well, thanks

@CarLaTeX :)

@CarLaTeX hey I hear about Leslie Lamport (I saw him in the 6th HLF). Does he participate in the site?

@manooooh I don't know. Did it talk at your university?

@CarLaTeX nono, hopefully! He gave a lecture in Germany

@manooooh :)

I saw him in that video and I remember his name, I thought he was helping in this site (he was one of the developers of LaTeX, according to Wiki)

@manooooh Someone says "La" of "LaTeX" stands for Lamport. Knuth wrote TeX, Lamport wrote LaTeX, but there are rumors he now works for the enemy (MS) :):):)

@CarLaTeX buuuu!! With the good looks he gave me he looked like this guy (30 years younger)

@manooooh lol

@CarLaTeX that channel is awesome <3<3. I force you (attention!) to watch a coding challenge video

@CarLaTeX time to sleep here. Have an excellent day!!

@manooooh Good night!

[6/7, 00:15/00:17] update: luaotfload [923k] (47444 -> 48771) ... done

@UlrikeFischer ^^

@DavidCarlisle I'm holding my breath. But it isn't everywhere yet, I had to force the dante repository.

@JosephWright how can one run l3build tests without copying the sourcefiles to the test folder?

@UlrikeFischer I think that was a special request from @egreg

@UlrikeFischer I've removed the path setting to my clone of your luaotfload and re-running the latex tests

@DavidCarlisle Ah good. I'm just trying to figure out a config-file which would allow me to run the tests on the default texmf tree.

@UlrikeFischer @JosephWright should I just check this in or should we try to normalise .luc files out, it seems to be only failure left on travis

*** ../build/test-config-TU/tu-composites01.luatex.tlg	2018-09-26 16:28:42.795107998 +0000
--- ../build/test-config-TU/tu-composites01.luatex.log	2018-09-26 16:28:43.387111449 +0000
***************
*** 18,24 ****
  TESTING \^
  TESTING \~
  TESTING \textcommabelow
! [1
  Missing character: There is no ̧ (U+0327) in font [lmroman10-regular]:+tlig;!
  Missing character: There is no ̧ (U+0327) in font [lmroman10-regular]:+tlig;!
  Missing character: There is no ̧ (U+0327) in font [lmroman10-regular]:+tlig;!

@UlrikeFischer You can't: one of the design aims was to run tests in 'isolation'

@JosephWright But if there were only tests in the repo l3build couldn't copy anything beside the tests and would have to run them as is. So shouldn't it work if I set the variables for the source files to empty or redirect the sourcefile dir in a config file?

@UlrikeFischer yes, so long as you don't need the sources to build whatever it is you are testing (which normally you do)

@DavidCarlisle as I want to test the installation in texlive I don't want anything to build - that't the point of the test ;-).

@UlrikeFischer That's very good, er, bad news!

@UlrikeFischer you should be able to have texlive-build.lua and then something like l3build check -c texlive-build (we have two such build files in latex2e/base)

@JosephWright with a stock texlive but l3build from github l3build check in 2e/base passes here but is out by a (font) normalisation in tu-tl2e7 on travis, do the config-TU tests pass for you?

@DavidCarlisle @JosephWright @MarcelKrüger I would like to merge the development branches into master before continuing (the next version is already lurking mail-archive.com/ntg-context@ntg.nl/msg89104.html). But they contain quite a number of dirty commits. Should the history be squashed or not (in view of a later move to the latex github).

@UlrikeFischer squash I'd guess, see what @JosephWright says

@DavidCarlisle the search problem in miktex is not due to openin: github.com/MiKTeX/miktex/issues/199#issuecomment-424843476

@UlrikeFischer MiKTeX always search the disk, even if a file name database exists. eek.

@DavidCarlisle I'll check in a bit

@JosephWright thanks

@UlrikeFischer Really down to you: I'd probably squash to have a clean history, unless there is important detail that would be lost

@DavidCarlisle I'll try my Windows box and my VM, and see if I can track it down

@JosephWright thanks just l3build check -c config-TU seems to differ (quicker than whole thing:-)

Hi, Would you recommend adding a non-breaking space '~' after usages like 1) e.g. , 2) viz. and 3) i.e. in a thesis?

Clearly, it doesn't look good if i.e. appears at the end of a line in a paragraph, right?

@UlrikeFischer Personally I am not a big fan of squashing because I think having a merge commit provides an almost as clean history as using squashing without loosing the history of the changes. Especially when looking into LuaTeX (or ConTeXt) commits I would often wish for have access to smaller commits. But this is not that important for me so if David Carlisle prefers squashing I'm OK with that too.

I am asking the nbsp question here because, it is not documented in this highly popular question here tex.stackexchange.com/questions/15547/…

@MarcelKrüger well i don't prefer it but I don't mind it, if the branch has some commit history that was just you and @UlrikeFischer getting used to the luaotfload layout and re-organising it a bit, losing some of that history is fine if you want to lose it (it's probably only of interest to you)

@Krishna yes although in more or less formal text like a theses I'd avoid e.g. anyway ad write for example

Ah....That is now a big change to do David. Thesis submission is next Wednesday

I know its a trivial change with grep, sed /awk , but there is no time to re-read 300 pages again to see if it reads well

I just hope the examiners do not take offence at the use of e.g.

@Krishna if you were to fail an exam for writing e.g. rather than "for example" something would be very wrong with the world, but as you were asking about good style...

:)

I appreciate the suggestion. I think I should consider doing it

So, replace e.g. with "for example"

and "viz" with "namely"

and "i.e." with "that is" ......the last one sounds a bit silly to the ear

@MarcelKrüger with (much) more time at hand I could (perhaps) have done the development in sensible steps with more "dirty" branches and merge them with good commit messages one by one. But in reality a lot of steps happened in parallel, I often had to stop in the middle and to push only as backup, and there are tons of "next try with travis" commits. So the branch is the "dirty branch". continued ...-->

... As @DavidCarlisle and @JosephWright didn't object I will squash but not delete the branch and leave a reference in the commit message.

@CarLaTeX -- all these rumors are true. lamport created latex. the "la" is almost certainly the first two letters of his last name. and he does now work for microsoft. as far as i know, he is no longer active in the tex world.

One more really-quick question, should I use a non-breaking space '~' in front of the \gls commands for acronyms in my document when using the glossaries-extra package?

@UlrikeFischer, @MarcelKrüger I see the use of multiple commits where they have some useful. But it's a question of how much work to do.

@Krishna Better be generous with ~ ties (and possibly consistent). When the text is in final form you can judge, in cases of bad paragraphing, whether to reword or to remove a tie.

It is in the final form!

I am just doing the last minute checks.

@DavidCarlisle It could be a leftover from the intermediate fontloader: I had a set a higher log-level in the luaotfload.conf, and then you get messages about loading and saving the luc in the log. @JosephWright

@DavidCarlisle I mixed about the sides. The line is on travis. That's seems to indicated that it hasn't the new fontloader yet (and now I wonder why the line has disappeared in the new fontloader ...)

@UlrikeFischer ah that could be it.... if that's a user settable thing what setting would give the log file that looked most like a pdftex log file (to reduce the need for .luatex.tlg test results)

takamu.sakura.ne.jp/xml-entity-names-ja/index.html

@DavidCarlisle well there is log-level, but the default is already 0. I had a higher value in the lua-font-pond, which explains why some of the tests there have the luc-line. But I don't know why the behaviour of log-level 0 changed between tl17 and now.

I need to use the pdfpages package and its includepdf command for inserting a pdf. I'd like to redact some personal information from that pdf. What is the best way to do this?

Needless to say I don't have Acrobat Pro

@Krishna Redacting for print or for digital use?

Redacting for use in thesis, the printed copy will be examined by examiners, and the online copy will be seen by the whole world....

So, in summary redacting the online version is sufficient.......

But when the PDF printed, it shouldn't print the redacted content on paper

@TeXnician Forgot to reply directly to you in my earlier message

@Krishna That's very difficult, close to impossible if you want to keep text as text.

@Krishna Redacting for online use is harder (as overlaying does not do anything but "hide" the text behind a layer). You have to remove the underlying layer which is a bit tricky. The easiest would be to use a graphics program like Gimp for that (if it only a few pages that might be feasible).

@TeXnician I was looking for some command-line application that will do this

@Krishna You could also try to edit the pdf in inkscape.

@samcarter I see

@samcarter There are several online services which will do this. But I am not sure about their trustworthiness

@Krishna What? Magically detect which parts of the page are personal information and then remove that bit from the PDF? You can actually do something very simple: Create an uncompressed PDF, use the postscript and just insert an appropriate gap instead of the text.

@UlrikeFischer it's possible I checked in some tlg while using your font pond version, I'll see what others get on the test suite now before changing the test results again

@TeXnician Nope... I am not that crazy.... .Something like pdfcrop, where you specify a rectangular bounding box at the command line

@Krishna it's hard to do from a rectangle as you need to find the text in the pdf structure. (acrobat can do this, but not the free version)

@TeXnician Any tutorial on how to do that? Postscript is too low level for my current knowledge

@Krishna Why online services? Inkscape is available for most operating systems.

@DavidCarlisle Why can acrobat do a lot more than everyone else...I hate this.

@DavidCarlisle Since 2008, PDF has been an open standard, right. ISO32001 or something

@Krishna all you need to know is that postscript strings are in (..) brackets so find the strings and edit.

@Krishna and?

So, if the standard is open.....Readers and editors know how to implement it

@DavidCarlisle and be very careful not to change the length of the strings.

@Krishna there are ISO standards for building bridges as well, but that doesn't mean that if you come to a river you can easily cross without paying.

@UlrikeFischer not so important in postscript (fatal in pdf of course)

Yes... I understand ...I am not talking about competing against a payment-based business model

After all we all got to eat

@DavidCarlisle I thought it is about pdf.

All I am wondering is, how come there is no good open-source competing product that implements a lot of features that are acrobat-exclusive

@UlrikeFischer @TeXnician suggested a few commends back to use pdf2ps, edit the postscript then ps2pdf

@DavidCarlisle When I edit this, how does the black redaction rectangle appear?

@DavidCarlisle I only saw the "Create an uncompressed PDF" bit ;-)

@Krishna because very few people need to redact text in that way, and mostly those that do do not worry about paying a few hundred (or thousand) $ for anything.

I see

@DavidCarlisle It is not just redaction. It's also about the other features

Like animate, media9 etc. which use ocgs?

Everything requires adobe acrobat

@Krishna yes there I think it's just because it's hard:-)

And the funny thing is, acrobat reader doesn't work on Linux

@Krishna it's not that great on windows either

Well....harder than the kernel itself? And yet, they did it isn't it?

What I mean is, there is no acrobat for linux

Anyway, some things about pdf really perplex me... I shall leave it at that

@Krishna yes sure that was a commercial decision, there used to be, but they pulled it

Looks like inkscape is the way to go

@Krishna only "some"?

yes...since acroread 6 or something

@Krishna when I made corrections to my thesis I used scissors and glue

:)

inkscape it is then

I found an interesting project which converts to images, redacts and then back to pdf

github.com/firstlookmedia/pdf-redact-tools

@Krishna doesn't that make all the text in your pdf not text but an image?

I really hope it OCRs it back

@Krishna eek I wouldn't trust that

I just looked at the README.md file. Nothing fancy here

Just a manual process of converting to image and back

@DavidCarlisle Wait a minute

I tried another one

This one actually worked flawlessly. Kind of unbelievable

www.pdfzorro.com

I don't know how they do that

@all please feel to try it out yourself.....My text is unchanged. Selectable and searchable, and yet I was able to redact out stuff

@Krishna pdfzorro does not remove the information from the pdf and just adds some coloured boxes. You could have the same result with tikz

@barbarabeeton Oooh, maybe he'll come back to LaTeX world when he retires :):):)

@samcarter For all practical purposes that's all I need isn't it

tikz is heavy. There needs to be a lightweight package for PDF redaction. If real redaction is hard, the coloured boxes approach. But without loading the entire tikz package

@Krishna Well if the only thing you need is a black rule above the pdf you can use \rule{3cm}{0.5cm}. But as @samcarter pointed out this only blocks the visual view on the text. Everyone with a bit knowledge can still extract the text behind the rule.

@Krishna tikz was just an easy to use example. You could also add boxes with picture mode, or use a picture mode plane to overlay your content.

@Krishna Knowledge about pdf is not even necessary, just copy and paste from the your "redacted" pdf will show all information.

@UlrikeFischer Thank you Ulrike for the suggestion. How would one specify the co-ordinate for the content to the redacted

@samcarter Ah. I see. Thank you. That's a security issue indeed. Thanks for pointing this out

@Krishna Also pdfzorro also embeds advertisement for their site in your pdf ...

Alas. there is no redact package

@samcarter Really? I don't see this in my pdf

@Krishna something like \includepdf[pagecommand={\vspace*{5cm}\hspace*{4cm}\rule{..}{..}]{...}` should work.

@UlrikeFischer I see, and the content behind will still be selectable and searchable

@Krishna yes.

@Krishna select everything, copy and past and you will see something like Powered by TCPDF (www.tcpdf.org)

Ah.. This is terrible

(maybe no advertisement, but some additional text which should not be there)

Indeed

Really good catch

@Krishna If you want to remove information from a pdf, use some editor and delete it. Covering it with boxes is only good if you print the document.

@Krishna but putting coloured boxes over the text is easy but doesn't hide anything

Cool....got it

@DavidCarlisle separate question for you. In this question tex.stackexchange.com/questions/63585/… you suggest how to use a longtable in landscape mode

This doesn't quite cut it for me. I need a longtable rotated sideways, without disturbing the paper orientation i.e. the pages in the PDF should remain pdf, and the table rotated sideways

@Krishna yes (the lscape package was written for that use)

@Krishna that's what it does

@DavidCarlisle I still don't get it

It simply provides the landscape environment`

Which makes the pdf page landscape in my reader

@Krishna yes but that does not rotate the whole page, the page head and foot stay in portrait orientation, it just rotates the page content

@DavidCarlisle Ah. My experience is otherwise

@Krishna well that depends, it may have a hint (depending) but anyway some readers just detect that most of the page is in landscape orientation and rotate the view. Why does the viewing hint make any difference really?

As you can clearly see, my table's footer is messed up

@Krishna the red bit is the page foot?

Yes.....

The the rule on the right is the page's header

@Krishna so as I said. That is in portrait orientation.

But the PDF viewer turns it

The page shouldn't be turned this way at all

@Krishna they do that anyway just by heuristics to guess which way up it should be. Why do you want to read it sideways?? It makes no difference to printing

the preceding page is correctly oriented in the viewer (obviously). So, I expect my next page in the viewer to show up similarly

Is there a way to disable the viewer heauristics here.... Oh man....there is so much black magic in the world of pdfs

I guess I am asking a viewer-specific question here.

@Krishna that of course depends on the viewer, but do you really expect the reader to read the table sideways? If it doesn't rotate automatically isn't anyone just going to select a rotate option in the viewer for that page?

yes. I want them to do that

When I submit my thesis, I am worried the printer is going to have problems

@Krishna but you could try putting \def\pdfpageattr#1{} just after \begin{landscape}

@Krishna why???

@Krishna No it has no affect on printing

Call me a bit layman-ish.

When I do "pagedown" on the pdfs, suddenly there is one page that occupies a different width

it just feels weird

@Krishna if you print it to a book, people will turn the book, same on a tablet, but it's so inconvenient to rotate a desktop monitor or even a laptop.

I'd like the readers to consciously use the rotate feature in their viewer, and not give them sudden disorientations

@DavidCarlisle Hmm...of course. I guess this question is simply boiling down to a personal preference here. I'd rather manually use the rotate feature in my viewer

@Krishna as I say you can disable lscape giving a hint, but many viewers do not need that hint, and I don't think any reader would thank you for displaying the table sideways.

\def\pdfpageattr#1{} gives a hint to the viewer?

@Krishna well it stops the package using \pdfpageattr to give a page view hint

@DavidCarlisle Cool...Thank you very much

As always, all of you here are amazing individuals

And have my gratitude forever

@DavidCarlisle The viewer didn't take the hint :)

@Krishna are you using pdflscape or lscape?

pdflscape

I thought this is the more modern package

"The pack­age adds PDF sup­port to the land­scape en­vi­ron­ment of pack­age lscape, by set­ting the PDF /Ro­tate page at­tribute. Pages with this at­tribute will be dis­played in land­scape ori­en­ta­tion by con­form­ing PDF view­ers."

So, maybe I should just use lscape right?

@Krishna no, pdflscape is ok. I only checked. Try this:

\documentclass{book}
\usepackage{pdflscape}
\begin{document}
\begin{landscape}
\makeatletter \PLS@RemoveRotate
some text
\end{landscape}

\end{document}

@UlrikeFischer thank you. Shall try and update

No longer need \def\pdfpageattr#1{} I guess

@UlrikeFischer yahoo.... all of you are amazing here. This works like a charm

@UlrikeFischer Thank you so much

So much black magic going on :)

I thought other than hyperref, no other package interacted with the viewer so much

@Krishna look at the author of pdflscape ....Also the main point of the package is to interact with the viewer ...

@UlrikeFischer Hieko indeed

@Krishna odd actually it makes a difference as pdflscape mostly just adds the rotation hints for all possible driver options whereas lscape only does it for pdftex, but either way if you disable the roataion teh two packages are essentially using the same code.

@DavidCarlisle lscape adds it for pdftex only if the option pdftex is used, so simply \usepackge{lscape} should work too, if one don't want any rotation.

@UlrikeFischer yes that's what I would have thought (I have seen that code before:-) but @Krishna said it rotated so I guessed that either the option was used or the viewer was using heuristics to rotate anyway, either way I'd expect it to work like pdflscape, but whatever...

@DavidCarlisle you never asked if @Krishna used pdflscape or lscape ;-)

@UlrikeFischer ah:-)

Could someone who is using MikTeX check whether the files ducksay.code.v1.tex, ducksay.code.v2.tex and ducksay.animals.tex are included. miktex.org/Package/Browse/ducksay does seem like that isn't the case.

@Skillmon oh, I forgot. I will check asap.

@Skillmon no, it doesn't work. I will make a bug report.

@UlrikeFischer thank you very much.

@UlrikeFischer and no problem :) I forgot, too, just found the still open tab of MikTeX by chance and remembered.

@DavidCarlisle Everything works. Thanks to you and Ulrike

yeh test suite just passed on Travis (@UlrikeFischer)

@Krishna I don't think your readers will appreciate it, but at least you are happy:-)

Final question for the day from my end. I am wrapping up the thesis. The pdflscape/longtable combo works great except for the last page

the table in the last page is flushed left (top, depending on how you view it)

Is there any way to centre the contents of the last page.? The centering macro does not seem to help

@Krishna which is surely what you want

@DavidCarlisle No... It makes sense in portrait mode. Not in landscape

@Krishna possibly there is but I can't see any logic in that, if you have paragraphs of text breaking over a page you would expect the continuation text to start after the break not be vertically centred. a page breaking table is no different

In my understanding things are like this, isn't it?

When latex constructs a float page, it always centers its contents

@Krishna \centering is for horizontal centering, you would want a vertical centering and latex has no standard mechanism for that other than float pages

This "continuation" text is really continuous only for long sets of numerical data

I don't have continuity

@Krishna but that is completely different, that is an insert on a specially inserted page not text carrying over from the previous page

They are in a sense connected, since I am listing the attributions for third-party copyrighted material

So, have to remain in a table.....landscape because Imperial requires way too many columns and long because I have many.......and I need them all to remain in one table

@Krishna yes but if the list takes more than one page they should all be top aligned.

@DavidCarlisle Alright...that's the normal scenario...and I understand

But now, can I force the last page to a rubber-length \vfill or something?

Should I ask on the main forum as a stand-alone question

@Krishna possibly (using lscape) you could replace \setbox\@outputbox\vbox{\hbox{\rotatebox{90}{\box\@outputbox}}}} by \setbox\@outputbox\vbox{\vfil\hbox{\rotatebox{90}{\box\@outputbox}}}\vfil} which would vertically centre all the rotated pages (perhaps) but it seems wrong to me, I wouldn't add an option to the package to do that

are u asking me to hack the sty file of the package?

@Krishna a copy of it, yes, you are wanting to change latex's output routine and there is no built in option to do that (as it is a weird thing to do) so modifying the low level page layout code is expected.

Hmm....See, just when you'd like to do something out of the box, you hit a wall

I thought with the flexibility of luatex some of that can change.

I asked the question here

@Krishna luatex lets you go lower and change the algorithms in tex the program, it doesn't add any easy user customisations that are not in tex

hmm

This is a genuinely valid case...not just for the specific example I am talking about

People might want to present tabular information (not necessarily data)

that is long to fit in one page, and long to fit in portrait mode

@Krishna I don't think so, and it's my code, so there is no user option to do that, but it is only 26 lines of code in lscape so even if you totally re-wrote it it would not be that much customisation clde

@DavidCarlisle Alright

@Krishna but the code i suggested above probably works:-)

Let me try \setbox\@outputbox\vbox{\vfil\hbox{\rotatebox{90}{\box\@outputbox}}}\vfil}

I am using pdflscape though

@DavidCarlisle Does lscape work with luatex?

This is the situation I am talking about. When presenting non-numeric data, I prefer the table to be vertically centered

@Krishna yes

@Krishna it only "looks" more logical to change the last page because you are presenting the table in a form that no one can read, sideways. If you presented the table in a form that people could read it adding a large vertical space mid table at a page break would be very distracting.

@DavidCarlisle hehe

Well...I'd like to try your suggestion....working on it

I made the change to a local copy of lscape.sty in my project repo's root

@Krishna if you think about it as two table then make two tables and centering will be rather easy.

Yes...but need one single number

And, I don't want to manually decide the splitting point

The page split should be automatic

@Krishna then they are not two tables. Btw: I would try to make this table in portrait.

@UlrikeFischer too many columns to fit in portrait mode

When using a local copy of lscape.sty I get this error

@Krishna I don't think so, imho one could manage. Did you actually try?

Yes....first tried portrait, before doing landscape

Anyway...I spent too much time on this todya.....Kind of tired now

@Krishna reducing the font size to \small and the size of \tabcolsep as well as leaving out the outer most \tabcolseps one might get decent results in portrait.

@Skillmon Yes....I have the two outermost \tabcolsep set to @{} ...That's what you meant, right?

The font is currently scriptsize !

@Skillmon I just tried with \setlength{\tabcolsep}{0pt} in portrait mode

no good

Still overflowing

@Krishna you naturally would have to fine tune the column widths. The main column can be smaller and some of the other would need some optimizing too.

@Krishna you'd have to reduce the column width of the third column. And if it's already \scriptsize then one shouldn't reduce the font size further :)

@Skillmon Agree.....

@UlrikeFischer Hyphenation becomes horrible with narrow column widths

even with microtype turned on, this doesn't work....especially with the url in the citation

for the source column

And breaking of urls is not preferable

Another option would be to ask biblatex/biber to suppress urls temporarily

How can I achieve this with biblatex/biber?

This person's question exactly tex.stackexchange.com/questions/113039/…