- dev-context - NTG Mailing Lists

Minor Problems in the LuaMetaTeX Manual
by Sebastian Miele 05 Feb '20

05 Feb '20

"The characters table is a list of character hashes indexed by an integer number": The sentence says that the table is a list. According to 'Programming in Lua' that indeed is the case. But the term 'list' conveys that the order of the items is something that matters, which (most probably) is not the case here. As far as I see, the intent of the table is to just map character codes to character information. Using the term 'list' in that context is a bit misleading. "and tell the engine that œ counts as one character": I think it should be "two characters". "In figure ?? and 7.7 we show what happens with three samples" "(the ones resulting from an \exhyphenchar": Missing closing parentheses. "the case relate codes": "relate" instead of "related". "This word separation is worth mentioning because": Should be "separation of stages" or just "separation" instead of "word separation". Maybe "work separation" was intended. In the table at the top of page 80, the "sub 1" probably should go directly below "POST", and the "sub 2" directly below "REPLACE". Now they occur aligned to the numbering of the terminal nodes of the structure presented in the table. "it does not expect there to be discretionaries inside of discretionaries": As far as I understood the previous text, what is really meant here is "it does expect there not to be discretionaries inside of discretionaries". The Lua(Meta)TeX manuals talk about ghost nodes. I did not find any information on what that is. Up to now, for all other stuff encountered in the LuaMetaTeX manual that I did not know and wanted to have a rough idea of what it is about, I found the relevant information either in the LuaMetaTeX manual itself, in TeX by Topic, or in the ε-TeX manual. I.e. I found it using the LuaMetaTeX manual itself, or the manuals that were referenced at the beginning of chapter 1 of the LuaMetaTeX manual as important additional material. The Lua(Meta)TeX manual mentiones that \leftghost and \rightghost come from Aleph. I searched the web quite a bit about Omega and Aleph on that issue, and tried other things, but found nothing of relevance. There is a manual for Omega, but, after conversion from ps to pdf, and opening the pdf with Evince, the search function did not find the phrase 'ghost' in it, either.

2 3

Example on \mathstyle + \mathchoice + grouping in the LuaMetaTeX Manual
by Sebastian Miele 04 Feb '20

04 Feb '20

It took me longer than I want to admit here to mentally parse the example at the end of section 8.3.1 on \mathstyle, \mathchoice, and grouping. In particular it took a while until I noticed that the b case has no \bgroup ... \egroup. This may very well be because of temporary stupidity on my side. But nonetheless I recommend changing the example either so that the mathchoices between \bgroup ... \egroup receive one more level of indentation, or, even better, to change the example into something like: \def\a#1{ \mathchoice {\bf \scriptstyle (#1:d :\mathstyle)} {\bf \scriptscriptstyle (#1:t :\mathstyle)} {\bf \scriptscriptstyle (#1:s :\mathstyle)} {\bf \scriptscriptstyle (#1:ss:\mathstyle)}} [a:\mathstyle]\quad \bgroup\a x\egroup \quad [b:\mathstyle]\quad \a y \quad [c:\mathstyle]\quad \bgroup\a z\egroup \quad [d:\mathstyle] Apart from being shorter, that totally minimizes the effort required by the reader to mentally parse and compare what is equal and what is different in the example.

1 0

Typos in the LuaMetaTex Manual
by Sebastian Miele 01 Feb '20

01 Feb '20

"separation between font- and backend": "font" instead of "front". "(an UNICODE or index)": "an" instead of "a". (U is a vowel, but its pronounciation starts with a consonant sound. The pronounciation decides whether a(n).) "A node can have a subtype to that it": "to" instead of "so". "with an node": "an" instead of "a". "(here we do that with [a]": Missing closing parenthesis. "but if you need a dozen you might wonder what you’re doing. .": Two consecutive full stops. "and expands it content": "it" instead of "its". "But this kind of expansion it still": Superflous "it". "This primitives is like \ignorespaces": "primitives" instead of "primitive". "These commands are use as": "use" instead of "used". "There is an nice side effect": "an" instead of "a". "in several places where the engine to deal with the dimensions of nodes": Missing verb. "as mentioned ,their": Space and comma transposed. "reference to an image s specification": Space instead of "'". "An implementation probably should accepts": "accepts" instead of "accept". "when no proper nested grouping is used": "proper" instead of "properly". "This is conform the LuaTeX philosophy.": "is conform" instead of e.g. "is conform to" or "conforms to". "can be present in en entry": "en" instead of "an". In section 6.2 (Defining Fonts) the word 'hash' is used strangely. It seems like it is used to mean hash tables: "The parameters is a hash with mixed key types", "The characters table is a list of character hashes indexed by an integer number", "Each character hash itself is a hash", "The kerns table is a hash indexed by character index", "The ligatures table is a hash indexed by character index", "with the values being yet another small hash, with two fields". "They were mostly use for": "use" instead of "used". "At the font level there can be a fonts an (indexed) Lua table."

1 0

sortedkeys and compare in LMTX l-table.lua
by Sebastian Miele 28 Jan '20

28 Jan '20

On table.sortedkeys the page https://www.contextgarden.net/Table_manipulation says: "Returns a sorted array of all the keys in table t. The comparer used for sorting will treat everything as a string iff the compared values are of mixed type." According to l-table.lua, sortedkeys first checks whether the keys in the given table are (1) all strings, (2) all numbers, or (3): neither (1) or (2). In cases (1) and (2) table.sort is used with the default comparer (Lua's '<'). In case (3) table.sort is used with a custom compare function named 'compare' and defined in l-table.lua. According to the sentence from the garden cited above, 'compare' should return (something equivalent) to tostring(a) < tostring(b). However, it does something more complex that, e.g., makes 10 < "2" < 3 < 10 ('<' meaning infix 'compare' here). I.e. there are strict circles in the strict order. https://www.contextgarden.net/Table_manipulation goes on explaining that, "In general this means that you will be fine as long as you avoid stuffing indices along with number hashes together in t, lest the order will end up confused depending on what type an element is compared with in what order (cf. example [2])." This reflects the way 'compare' actually is at the moment. But it contradicts the previous sentence "The comparer used for sorting will treat everything as a string iff the compared values are of mixed type." If everything is treated as a string when neither (1) nor (2), then everything is well-defined and there is no confusion. Why is 'compare' not just (something equivalent to) tostring(a) < tostring(b), violating the sentence cited above? Are there any meaningful uses of the way 'compare' is now? I suspect that the intent was to create an order that is lexicographic and numeric at the same time. As far as I see it, it just is not possible the way it was tried: Lexicographically, 11<2. Numerically, 2<11. I see three ways out: (1) Disallow mixed types of keys in tables given to sortedkeys. (2) Always use lexicographic order in the case of mixed key types. (3) Combine the orders by an order-theoretic sum. One possiblity: Every value that is neither a number nor a string is smaller than every number. Every number is smaller than every string. Otherwise elements are compared as they are now. Another possiblity: Define an order over all types, define an order for every type. Then sort by type first, and in type second. (1) may not possible without breaking a lot. (2) has sort of a discontinuity: E.g., add one numeric key to a table with only strings, and the order changes globally. (3) contains well defined, very predictable solutions without discontinuities.

2 4

Typos in the LuaMetaTeX Manual
by Sebastian Miele 28 Jan '20

28 Jan '20

"The same is true fir" "More fine-grained primitives control is possible and you can look up the details in section 10.3.15. For simplicity’s sake, this manual assumes that you have executed the \directlua command as given above.": I think the second sentence was true in the LuaTeX manual, but is not true in the LuaMetaTeX manual. "Calling some form of tex.enableprimitives is highly important though, because if you do not, you will end up with a TEX82-lookalike that can run LUA code but not do much else.": Since LuaMetaTeX enables all primitives by default, this is not true anymore. "In the following syntax tables, such as the type of such a userdata object is represented as ⟨node⟩.": Remove "such as". 'The return values of node.values("glue" are:': Missing closing parenthesis. Ditto in the following 'The return values of ... are' lines. "when we issue this command, we get a table.": Lower case "w" at the beginning of the sentence. "where the number don’t relate to importance": "number" instead of "numbers". "This function that will append a node list to TEX’s ‘current list’.": Remove "that". "You mignt need" "This helper makes sure that the node lists is double linked": "lists" instead of "list" "9.3.3.12 find_node": This section somehow has "node.traverse_list" in it instead of "node.find_node". "is a single tokens": "tokens" instead of "token".

1 0

No Official ConTeXt LMTX Repository
by Sebastian Miele 27 Jan '20

27 Jan '20

In the past months I worked through The TeXbook and The METAFONTbook. About a week ago I started to work through the LuaMetaTeX manual. For several reasons I want to learn ConTeXt from bottom to top. The main page of the ConTeXt Garden says "There is no offical repository there the development is going." ("there" in that sentence probably should be "where". I seem to not be allowed to edit the garden main page.) The LuaMetaTeX manual contains typos and other issues. In such cases I usually checkout the most recent development branch of the official repository and create continually rebased commits on top of it, that finally find their way into patches send over a mailing list, or pull requests on e.g. GitHub. This works really well. It would be really cool if ConTeXt LMTX had a/the official repository. For now, I will just somehow post the typos to dev-context(a)ntg.nl. I will use ConTeXt LMTX (from http://www.pragma-ade.nl/install.htm) as a reference point. I will update the installation it at least daily.

6 8

Possible Problem with Mailing List Setup
by Sebastian Miele 27 Jan '20

27 Jan '20

About a week ago I sent four mails to this mailing list without being subscribed. For each mail I got a reply that the message awaits moderator approval. After that I subscribed to the list, because it is low volume, and I want to spare the moderator some approval work. Shortly after that my subscription to the list got approved. But to date the messages sent before did not appear on the list. After waiting for a week, I revised the four messages and sent them again. So the not approved mails are obsolete by now. Since I assume that most moderators would not approve me subscribing to the list after I sent four disapproved messages, the most probably thing is that the messages got lost. I.e. sending mail to this list without being subscribed currently is not possible. Best wishes Sebastian

3 3

document.showprimitives in luametatex-tex.tex
by Sebastian Miele 26 Jan '20

26 Jan '20

luametatex-tex.tex defines and shortly thereafter uses a Lua function document.showprimitives. The definition of the function uses: table.sortedpairs(tex.extraprimitives(tag)) tex.extraprimitives(tag) returns a sequence. table.sortedpairs sorts on the keys of the sequence. In the case of a sequence, table.sortedpairs is equivalent to ipairs, and probably not what is intended here. See the results of the current uses of document.showprimitives in the PDF. They are not sorted alphabetically. There probably should be a sort on the values of the sequence returned by tex.extraprimitives(tag). A few weeks ago, this problem was present in the LuaTeX manual, too.

1 0

Typos in the LuaMetaTeX Manual
by Sebastian Miele 25 Jan '20

25 Jan '20

The following typos in the LuaMetaTeX Manual are from an about one week old ConTeXt LMTX (http://www.pragma-ade.nl/install.htm) In the future I will be more timely, as stated in the previous mail. A problem of not having an official repository for LMTX is: If I see a typo or problem in a section that I already read in the past, then, without further lookup, I do not know whether it already is reported or not. If I had a clone of an official repository and the manual were build from that, including the continually rebased and not-yet published fixes, that problem would not exist. I will try to deal gracefully with that problem. Depending on the problem, time, and mood, I will check for duplicates, completely ignore the typo/problem, or report the same issue more than once. "have are" in "All options are accepted but only some have are understood by" Missing closing parenthesis in "LUAs tostring function (and string.format may return values in scientific notation" "This mainly means that lpeg.S cannot be used with UTF 8 characters encoded in more than two bytes": I do not know any details of LPeg. But from a glance I suspect that "more than two bytes" should be something like "more than one byte" in that sentence. "There are all kind of subtle differences in the implementation": "kind" instead of "kinds". "Of course, of you use low level features": "of" instead of "if". "more of less long term stable": "of" instead of "or". "In chapter 5 we discussed some new primitives, here we will cover most of the adapted ones.": This sentence appears in chapter 3. So either "5" is wrong, or "discussed" should be "discuss". "The documentation, which is actually comes". "See chapter ?? for many small changes related to paragraph building".

1 0

CGModules update: Letter module version 2020.01.20 published.
by modules＠contextgarden.net 20 Jan '20

20 Jan '20

Wolfgang Schuster submitted an update to the Letter module package. Version number: 2020.01.20 License type: GNU General Public License Wants to be in ConTeXt distribution: no Wants to be synced with CTAN: no Summary description: Writing letters with ConTeXt. Announcement text: ---------------------------------------------------------------------- README was missing in the previous release. ----------------------------------------------------------------------

1 0