- dev-luatex - Mailing Lists

Wrong placement using \displaylimits and unicode-math
by Jakub Kaczor 08 Dec '21

08 Dec '21

Hello there, there were some reports regarding the placement of the limits in integrals a few years ago <https://tug.org/pipermail/luatex/2015-March/005076.html>. It was with regard to display math and `\limits`. They seem to work fine for me, but I spotted some weird behavior using `\displaylimits` *in in-line math*. In my opinion, the limits of integration are placed way too much on the right. The expected behavior is to have the same placement as with `\nolimits` in in-line math, and as with `\limits` in display math. The problem doesn't occur with XeLaTeX and it seems to be independent of a font. Here is a minimal working example comparing different options: ``` latex %! TeX program = lualatex \documentclass{article} \usepackage{unicode-math} \begin{document} $\int\limits_a^b f(x)\,\mathrm dx$ $\int\nolimits_a^b f(x)\,\mathrm dx$ $\int\displaylimits_a^b f(x)\,\mathrm dx$ \[\int\limits_a^b f(x)\,\mathrm dx\] \[\int\nolimits_a^b f(x)\,\mathrm dx\] \[\int\displaylimits_a^b f(x)\,\mathrm dx\] \end{document} ```` Here, the third line after `\begin{document}` is the faulty one. Best regards Jakub Kaczor

1 0

luahbtex Mac M1 memory
by Oliver Marugg 08 Dec '21

08 Dec '21

Hi Yesterday I experienced some strange memory issues on mac mini running maco 12.0.1 (monterey) with luabhbtex 1.13.2 (installed via texlive 2021 w. latest updates via tlmgr) and texstudio 4.1.1 as frontend. I read about memory leaks with m1 CPUS. I found out that the process luahbtex run with 33GB virtual memory permanently (without having a job to run) and used 99-100% CPU load, CPU was running hot. Instead of investigating, I killed the process and memory as well CPU load went back to normal. What could I do for you next time I experience such an effect? Many thanks. Ollie Marugg

1 0

Trouble with generating warning messages in Unicode
by निरंजन 30 Nov '21

30 Nov '21

Hello, I am having trouble with generating a warning message with unicode characters and LuaLaTeX. My code is as follows: \documentclass{article} \begin{filecontents}[overwrite]{quack.sty} \ProvidesPackage{quack}[2021/11/30 v0.1 quack quack] \newcommand{\quack}{% \PackageWarningNoLine{quack}{% ककाकिकी खखाखिखी गगागिगी घघाघिघी ङङाङिङी\MessageBreak चचाचिची छछाछिछी जजाजिजी झझाझिझी ञञाञिञी\MessageBreak टटाटिटी ठठाठिठी डडाडिडी ढढाढिढी णणाणिणी% }% } \end{filecontents} \usepackage{quack} \begin{document} \quack \end{document} If I run this with xelatex, the warning message is printed as expected. With pdflatex and lualatex it isn't. In a GitHub issue[1] I was told that pdftex and luatex count in bytes and I need to reset it with max_print_line in order to print this warning correctly. As it has to be set in the .cnf or .ini files in the local installations or per compile, I asked if it can be set in the TeX code. They told me to report this to luatex developers and therefore I am mailing this issue here. There are two questions: 1. If I set the max_print_line to, say 300, it prints the log in an unusual way. Everything in the log is affected. So can luatex be requested to count the strings the way XeLaTeX counts them? What implication will it have? 2. If the above is not possible/worth/advisable; can there be a feature to set the max_print_line in the TeX code? David Carlisle says[2] in LuaTeX it can be checked, but not changed. [1]: https://github.com/lualatex/luatexbase/issues/22 [2]: https://github.com/lualatex/luatexbase/issues/22#issuecomment-982553035

1 0

\partokenname, \partokencontext in LuaTeX?
by Petr Olsak 19 Nov '21

19 Nov '21

Hello, may I ask if the primitives \partokenname, \partokencontext will be implemented into LuaTeX? They were implemented in pdftex and XeTeX into TeXlive sources. I would very much welcome their implementation in LuaTeX, I need then on OpTeX macros. Regards Petr Olsak ----------- texk/web2c/NEWS: - new primitives \partokenname and \partokencontext allow overriding the name of the internally-emitted \par token at blank lines, the end of vboxes, etc. ----------- \pdftexprimitive{\Syntax{\Tex{\partokenname} \Whatever{control-sequence}}} \bookmark{\tex{partokenname}} \TeX{} inserts the {\it par-token} at empty lines and at various more places into the input stream. Its meaning is {\it end-paragraph} by default and it can be changed (by \type{\def}, for example). Its name is \type{\par} by default and you can change its name by \type{\partokenname}\Something{control-sequence}. The name of the {\it par-token} is given \Something{control-sequence}. The setting performed by \type{\partokenname} is global. The no-\type{\long} macros throw the error ``runaway argument'' if {\it par-token} is scanned to its parameter regardless its name. The following example releases the name \type{\par} to the ``user name space'', i.e.\ users can define and use \type{\par} as they need without any influence to behavior of internal \TeX{} algorithms. \starttyping \catcode`\_=11 \let\_par=\par \partokenname\_par Hello world.\end \stoptyping The token with name \type{\_par} is auto-emitted at the empty line and before \type{\end} primitive in this example. The no-\type{\long} macros can read parameters with \type{\par} but \type{\_par} is prohibited. \pdftexprimitive{\Syntax{\Tex{\partokencontext} \Whatever{number}}} \bookmark{\tex{partokencontext}} The {\it par-token} (i.e.\ the token with the name \type{\par} or with the name given by \type{\partokenname}) is inserted into the input stream depending on the \type{\partokencontext} value. Its value can be: 0: {\it par-token} is inserted at empty lines (more exactly, when a token category 5 is found in the state~$N$ of reading a line) and before \type{\end}, \type{\vskip}, \type{\hrule}, \type{\unvbox}, and \type{\halign} when \TeX{} is in horizontal mode and at various error recovery cases. 1: {\it par-token} is inserted as above and moreover at the end of \type{\vbox}, \type{\vtop}, and \type{\vcenter} if \TeX{} is in horizontal mode. 2: {\it par-token} is inserted as above and moreover at the end of \type{\insert}, \type{\vadjust}, \type{\output}, \type{\noalign}, and items of \type{\valign} if \TeX{} is in horizontal mode. Default is \type{\partokencontext=0}, i.e.\ the cases 1 and 2 are processed by direct call of {\it end-paragraph} routine without emitted {\it par-token}. This is classical \TeX{} behavior. If \type{\partokencontext=1} then \TeX\ inserts {\it par-token} in additional cases: when vertical boxes are completed but horizotal mode is not finished. The settings \type{\partokencontext=2} deactivates all cases where classical \TeX\ uses direct {\it end-paragraph} routine and {\it par-token} is inserted in such cases. The setting value to the register \type{\partokencontext} is local.

2 1

math.tointeger in l-math
by Robert 18 Nov '21

18 Nov '21

Hi, [sorry, I know it's the wrong list, but I'm not subscribed to ntg-context] I think line 144 in l-math.lua should read function math.ult(m,n) otherwise, the function math.tointeger would be overwritten with older luatex versions. (reference: https://source.contextgarden.net/tex/context/base/mkiv/l-math.lua ) Best, -- Robert

1 0

Synctex bug
by Michael Klooss 21 Oct '21

21 Oct '21

Hello, I hope this is the right place to report lualatex bugs. The lualatex ecosystem seems to have a synctex bug which has been resolved in pdflatex and xelatex some time ago. The problem is, that inverse search sometimes goes to the wrong file. The trigger seems to be the TikZ library shadows. An example and further references (detailing the bug) are below. Best, Michael % From https://sourceforge.net/p/pgf/bugs/333/ % See also https://tug.org/pipermail/tex-live/2017-March/039781.html % See also https://tex.stackexchange.com/questions/300079/goto-source-opens-pgflibrary… % The stackexchange ^ contains a workaround. I haven't tried it yet. % The problem only occurs with lualatex. pdflatex and xelatex work fine. \documentclass{article} \usepackage{tikz} \usetikzlibrary{shadows} \begin{document} Test. \end{document}

2 1

LuaTeX doesn't discard empty paragraphs when textdir is used
by Marcel Krüger 02 Sep '21

02 Sep '21

Hi, consider the following plain LuaTeX document: \textdir TRT \noindent\par \bye without the `\textdir TRT` line or with `\textdir TLT`, this would lead to "warning (pdf backend): no pages of output.", but with the non default "\textdir", it creates an empty (except for the page number) page instead. This can be avoided by setting \pardir too: \textdir TRT\pardir TRT \noindent\par \bye again produces no output, but adding a group around it reintroduces the issue: \begingroup \textdir TRT\pardir TRT \noindent\par \endgroup \bye leads to an empty page. Of course similar things happen not only for otherwise empty documents: Instead of empty paragraphs disappearing, they add empty lines. Together, this is not only inconsistent with other engines which always discard empty paragraphs, but also leads to hard to predict behavior (especially for users who are not familiar with the implementation of LuaTeX's directional system), so I think it would be great if LuaTeX could always remove such empty paragraphs. Best, Marcel

5 9

Issues with settex() and glue
by Michal Vlasák 28 Aug '21

28 Aug '21

Hello, as part of [1] I also noticed that, the "settex" function (used as metamethod for the "tex" table and for "tex.set") doesn't work for glue specified by (up to) five values case. There are multiple issues. Firstly this is the Lua stack when this function is called (i = lua_gettop()): tex.set("global", "parskip", 1, 2, 3, 4, 5) -> i-6 i-5 i-4 i-3 i-2 i-1 i "global" "parskip" 1 2 3 4 5 The function essentially doesn't do anything for the five number glue case, because of the outer most check: if (lua_type(L, i - 1) == LUA_TSTRING) { Even if it did, it wouldn't respect the "global" setting, because it is checked like this: if (i == 3 && (lua_type(L,1) == LUA_TSTRING)) { const char *s = lua_tostring(L, 1); if (lua_key_eq(s,global)) And even then, it would try to use values beyond the top of the stack: width(value) = lua_roundnumber(L,i); if (i > 1) { stretch(value) = lua_roundnumber(L,i+1); } if (i > 3) { shrink(value) = lua_roundnumber(L,i+2); } if (i > 4) { stretch_order(value) = lua_tointeger(L,i+3); } if (i > 5) { shrink_order(value) = lua_tointeger(L,i+4); } The manual also mentions the 5 value case twice in the same paragraph: --- a/manual/luatex-tex.tex +++ b/manual/luatex-tex.tex @@ -271,9 +271,7 @@ \type {tex.get} you get the width of the glue and when you pass \type {true} you get all five values. Otherwise you get a node which is a copy of the internal value so you are responsible for its freeing at the \LUA\ end. When you set a -glue quantity you can either pass a \nod {glue_spec} or upto five numbers. If -you pass \type {true} to \type {get} you get 5 values returned for a glue and -when you pass \type {false} you only get the width returned. +glue quantity you can either pass a \nod {glue_spec} or upto five numbers. And then the contradicts itself: 10.3.2.4 Glue parameters The glue parameters accept and return a userdata object that represents a glue_spec node. I tried to fix the issues, but the "settex" function is also complicated by the fact, that it also is a __newindex metamethod of the tex table, which means that its stack can also look like this: tex.parskip = n -> i-2 i-1 i tex "parskip" n So one can't easily just index from the the bottom. Also the line: int a = isglobal; in the glue case would deserve a comment, since it is subtle, but required for define() to work properly (it looks at the variable called "a"). I would prefer something like: int a = isglobal; /* for define() */ In the end I am not even sure how to handle the issues. I leave it to the more experienced. Kind regards, Michal Vlasák [1]: https://mailman.ntg.nl/pipermail/dev-luatex/2021-August/006542.html

1 0

lastnodetype, again
by Robert 22 Aug '21

22 Aug '21

Hi, while the last problem with \lastnodetype (cf. https://mailman.ntg.nl/pipermail/dev-luatex/2020-April/006414.html) has been fixed, there are still incompatibilities if it is preceded by \relax. The following outputs 7,0,8,7,7,7 with pdftex and etex, but 7,0,0,0,0,0 with luatex: %\input luaotfload.sty %\font\x=[lmroman12-regular.otf]:script=latn;+trep;+tlig;%mode=base; %\x % no difference \setbox0\hbox{--} \unhbox0 \the\lastnodetype % OK --\the\lastnodetype % OK --\relax\the\lastnodetype \setbox0\hbox{--\relax\the\lastnodetype} \unhbox0 \toks0={--} \setbox0\hbox{\the\toks0\relax \the\lastnodetype} \unhbox0 \setbox0\vbox{\hbox{--\relax\the\lastnodetype}} \unvbox0 \bye Best, -- Robert

2 4

harfbuzz 2.8.2
by luigi scarso 18 Aug '21

18 Aug '21

I have pushed harfbuzz 2.8.2 into experimental, if it's ok I will sync with the texlive repo. -- luigi

1 0