I noticed something very strange, may be someone has an idea: I run pdfltex first time, I get no error, but message saying to re-run again Package longtable Warning: Table widths have changed. Rerun LaTeX Now, I type pdflatex again, and now I get an error about capacity ! TeX capacity exceeded, sorry [main memory size=5000000]. \eq@repacka ...bal \setbox #1\hbox {\unhcopy \tw@ \unskip \count@ -\lastpena... l.452302 \end{dmath*}
! ==> Fatal error occurred, no output PDF file produced!
The above tells me, that second run uses something new that did not exist in the first run to cause this problem. So, I go now to the latex file and remove the offending long equation and try again. But it is strange that first run does not show the error ! very strange.
it seems, in the second pass, Latex is using more memory (since now it is making the TOC), and so it has less memory to process the actual latex document. Hence the error shows up in the second run. Since when I open the pdf file after first run, the TOC is not there. So may be this explains why I see the error on second run and not the first.
@AlanMunn No. These are what I am loading. Should I do it? \documentclass[10pt,notitlepage]{report} \usepackage[T1]{fontenc} \usepackage[margin={1.5cm}]{geometry}% \usepackage{amsmath} \usepackage{hyperref} \usepackage{fancyvrb} \usepackage{longtable} \usepackage{pdflscape} \usepackage{graphicx}
@DavidCarlisle well, then why would I get the memory error only on second run and not first? Something is taking more memory in second run. I just got it again:
Underfull \hbox (badness 10000) in paragraph at lines 513130--513130
$[] \T1/ntxrx/m/n/10 + [] \OMS/ntxsy/m/n/10 ^^@ []|$
pdfTeX warning (ext4): destination with the same identifier (name{60}) has been
already used, duplicate ignored
<to be read again>
\relax
l.513150 \hypertarget{60}{}
[2954]
! TeX capacity exceeded, sorry [main memory size=5000000].
\eq@capture ...ty -\@Mi \fi \keep@glue \@@endmath
\eq@addpunct \@@par \eq@wd...
@Nasser yes you can edit the limits in texmf.cnf up to a certain extent. luatex uses dynamic memory allocation but for pdftex/xetex memory is a fixed but configurable sized array]
I have a plot with a lot of measurement datapoints. I exported this plot from Matlab2tikz.
When I wanted to compile this, it said that the memory was exceeded. Now I found this, How to expand TeX's “main memory size”? (pgfplots memory overload). It tells me that I can export different tikzpictur...
@Nasser well I suppose it takes a bit (and a bit more with hyperref) but you must be very close to the edge anyway. when it doesn't fail what do the last lines of the log say the ones that looks like
Here is how much of TeX's memory you used:
238 strings out of 493109
2515 string characters out of 6135009
52967 words of memory out of 5000000
3791 multiletter control sequences out of 15000+600000
3640 words of font info for 14 fonts, out of 8000000 for 9000
1141 hyphenation exceptions out of 8191
23i,1n,17p,117b,36s stack positions out of 5000i,500n,10000p,200000b,80000s
@DavidCarlisle will let you know. I am removing the last complained about equation and trying now.
@DavidCarlisle here it is (from first pass, no error)
Package rerunfilecheck Info: Checksums for `report.out':
(rerunfilecheck) Before: EF8D6F0152EF1C4470D75C154BB009E9;427
(rerunfilecheck) After: B5E47786D64AE7CEFE39F7110C7D515A;865.
Package atveryend Info: Empty hook `AtVeryVeryEnd' on input line 903708.
)
Here is how much of TeX's memory you used:
36143 strings out of 493109
563084 string characters out of 6135003
1307268 words of memory out of 5000000
19263 multiletter control sequences out of 15000+600000
117740 words of font info for 212 fonts, out of 8000000 for 9000
If you really absolutely need more capacity, you can ask a wizard to enlarge me.
Here is how much of TeX's memory you used: 28672 strings out of 493109 440573 string characters out of 6135003 5000000 words of memory out of 5000000 19267 multiletter control sequences out of 15000+600000 116424 words of font info for 210 fonts, out of 8000000 for 9000
Why would words used increase from 52967 words in first run to 5000000 in second run? must be some package or something strange. but will go now try to increase latex memory and try again
@Nasser so you used a lot of extra mem. Is your dmath inside longtable? could be that the width changed as happens with lt and dmath just got stuck on linebreaking the new width an got itself into a loop? but I'm done good night..
@Paul Gessler I get an error trying to compile the following code. It indicates that there is an error in the command that I issue to plot the line y = x.
@DavidCarlisle @AlanMunn I switched to using lualatex and now I no longer see the problem with ram in second runs. So will just use lualatex from now on to compile to pdf.
StackExchange continues to support TUG with corporate membership (many thanks). With that come eight individual memberships that can be given out to members of the community. The time has come to select our representatives for 2015. (See TUG Membership: Names for 2014 for last year's selection.)
...
I have a pdf file about 340 KB and I used a Ghostscript command copied off somewhere and compressed it to about 260 KB. There is an image file of about 100 KB in it. Can I compress more?
@GaoWeiwei Backticks do inline code, four-space indenting does block code but if you paste in multiple lines you should see a 'fixed font' button next to 'Upload'
@JosephWright In MS Word, I remember there is an option to embed only those characters used, which greatly reduces size. Is there something similar that I can do with PDFs?
@PauloCereda I may possibly have inserted a minor typing error into a docstrip guard and caused a large chunk of the latex sources to be skipped while making the format.....
@subhamsoni Many questions on TeX.SX are about changing the layout. But the journal has a fixed layout. So the only valid answer/advice would be: Don't do it.
@PauloCereda I guess you mean into Danish. My immediate thought is that there a much better qualified person on the floor below, namely Lars Madsen (daleif). He has the advantage of being a native speaker.
@subhamsoni using a journal class when not submitting to the journal is weird:-) You can not report problems to the class maintainers (as you are not using their submission system) and the usual open source routes are likely to be slower because people are less familiar with them and less inclined to fix them as they can not alter the sources and the publishers have people paid to maintain them....
@subhamsoni The standard classes are called standard classes because they define very very basic set of macros. Hournals use those basic macros to define a template for their own needs. You now want to whack the journals defs back to the original ...
@subhamsoni But this is a basic misunderstanding of many many many users. They are searching for a template to use. But you cannot start any LaTeX document using one of the basic templates article, report or book.
@subhamsoni but that is another class designed to use for a specific journal publisher and supported by that publisher solely for submissions to IEEE Journals.
@subhamsoni I once hat a user with some whacky template by a university s/he wasn't even studying at. It defined section titles differently. The user wanted to revert it in a way, that was exactly the standard of article. And that was all the template was doing, redefining the titles. Kicking the package in the trash bin was the simple solution.
@Johannes_B: I'll have another appointment... for two reasons: 1st. I don't like Solyanka and 2nd: I moved recently and am not inclined to help anybody else to move ... period!!!!
@ChristianHupfer Oh, puh. äh, ich hab keine Zeit, ich muss noch Altglas wegschaffen. hab noch Wurst im Auto muss Sand schippen bekomm an dem Tag ne Tonne Kies.
@ChristianHupfer Ausreden für jede Gelegenheit ;-)
I am writing my thesis. To write math formulas I created the command \eq
as follows.
\newcommand{\eq}[1]{
\begin{equation}
\begin{split}
#1
\end{split}
\end{equation}}
However, now I want to use cleveref together with autonum.
I think, however, that this is not possible due to the wa...
@JosephWright I have sent a private message to the pgfplots user informing him about local user groups (giving him links as well). He might feel more comfortable if there is one in his native language. I think it really is a matter of language we are dealing here.
@HenriMenke I edited my comment to try to cover this: XeTeX shouldn't break anything with input that would also work with Knuth's TeX or e-TeX. Once you start using new features there are different design decisions. Kunth's TeX can't load OTF fonts, for example.
@ChristianHupfer I was just thinking back when i was at an actual MacDonalds. I had coupons and i decided to learn what a Latte Macchiato is. Got to the counter, said: A latte macchiato please. The woman looked at me and asked: Which taste? I bet i was never in my whole life more confused than in this second. After one or two seconds of pure puzzled staring i asked Kaffee?
@HenriMenke Yes, I would too but the trick here picks up based on engine features, so would work with any other new engines that work with Unicode input and extend the ^ syntax. My point is that there are changes that automatically lead to issues
@egreg: Yes, I remember, it was something about etextools... only a few days ago... those bugs are like the poison running through your veins, to cite Alice Cooper
@HenriMenke 30 is what in TeX would be represented by ^^^; I need it to be category code 12 because LaTeX makes it category code 15. ``\catcode\!=12 is just being a bit paranoic. ;-) It is ASCII 0x21.
Is there a question somewhere which shows how to pass unknown keys, parsed using expl3 stuff, to a command which uses keyval syntax? I'm trying to rewrite a command to use expl3 keys but I can't figure out how to pass unknown keys to \includegraphics[unknown keys]{image}`.. I can get them into a token list and typeset them verbatim in my document, but I can't get \includegraphics to take any notice - it doesn't complain, but it just ignores them.
@cfr For a start you need to make sure you expand the variable you pass to \includegraphics: \expandafter\includegraphics\expandafter[\g_cfr_imgopts_tl] or similar
@JosephWright Thank you! (I didn't know you could use \expandafter in the way the second one is used i.e. where it has a bracket after it rather than another command.
@egreg Is that better? I only ask because it uses New Things I Don't Understand. I only used the g because the example in the documentation for \tl_new did and I though I might need to since I defined it outside the \group_begin: stuff. (As you can tell, I have no idea what I'm doing.)
Is it better to turn expl3 on locally or for the whole class or package?
@cfr I was talking about \NewDocumentCommand\clawr{ O{} g }, not the \g_cfr bit. That token list should be local, because it's set and used in the same scope and not needed elsewhere.
@cfr By the way, \keys_set_known:nnN does a local assignment.
@egreg Oh, right. I don't usually use g. But I think it is better here because I don't want to break existing documents more than I can help. (And I will never remember to put the image file name in square rather than curly brackets.)
@HenriMenke From the fontspec docs: "Knuth’s CM fonts fonts are all squashed together, combining letters, accents, text symbols and maths symbols all in the one font, cmr, plus other things in other fonts. Because we are changing the roman font in the document, we need to redefine all of the maths glyphs in LATEX’s operators maths font to still go back to the legacy cmr font for all these random glyphs, unless a separate maths font package has been loaded instead."
@AlanMunn Thank you for pointing that out, but honestly I don't see any difference between math and no-math, even if I change the roman font to, say, Libertine.
@cfr in expl3 code ~ is a space token not a non breaks space command so it does nothing after a command name, just like the space in \mbox {..} does nothing other than terminate the command
@HenriMenke the classic math setup (if not using unicode-math) assumes for example that the roman font has upper case greek in the first few slots eg Gamma in position 0, this is not true of a typical system font
@HenriMenke Reading the comments in the package, it's using the legacymaths business to load accents for a small number of maths font packages so they come from the 'correct' place
@HenriMenke There are also a small set of other chars, but it's quite small
@HenriMenke because the math fonts are using the old scheme as it says in the text you quoted, or because you are using unicode-math in which case the math fonts are using Unicode layout but the definition of \Gamma has changed not to refer to slot 0
@HenriMenke just looking at your {} question, as noted in comments Joseph's answer is extensive, to see an excessive answer just look at (any of) @egreg's answers:-)
@PauloCereda I visited the burial place of both of them; different but equally moving. Bach has just his name on the ground, in the church were he worked; in the monument above his grave, Händel is represented composing the music for “I know that my Redeemer liveth”.
@PauloCereda Händel is in Westminster Abbey, London, where also Newton is buried (along with many others important people). Bach is in the Thomaskirche, Leipzig. Galileo Galilei (who died this day in 1642) is at Santa Croce in Florence. Let's see if @yo' knows the name of a great scientist buried in the Kostel Matky Boží před Týnem.
@PauloCereda On the contrary, Italy's greatest poet Dante is buried in a small chapel in Ravenna. The second one, Petrarca, is buried in the village near Padova where he lived his last years
@cfr: Are you kidding? Lol! I am a member for just one and a half year and my contribution to this site is still very limited, so it's not a good idea to promote myself. ... and no, this is not fishing for compliments
I am writing my thesis. To write math formulas I created the command \eq
as follows.
\newcommand{\eq}[1]{
\begin{equation}
\begin{split}
#1
\end{split}
\end{equation}}
However, now I want to use cleveref together with autonum.
I think, however, that this is not possible due to the wa...
