LMTX, BIdi, full paragraph, and spurious empty line
Hello all, Last October Denis Maier complained about a spurious empty line generated with protrusion and bidi active and a full paragraph. (See https://www.mail-archive.com/ntg-context@ntg.nl/msg95916.html https://www.mail-archive.com/ntg-context@ntg.nl/msg95916.html or similar on your favorite mail archive.) I have recreated the issue without involving protrusion. When a paragraph is very full and bidi is on, an extra line seems to be generated. Without bidi there is no extra line. This is quite sensitive, of course, to line width. I have probably 30 instances of it occurring in a tightly-set bibliography. I would assume that it may occur in any text, but have not noticed it in the body of my book. I have created an example document that shows it. The problem disappears if bidi is turned off (comment out the first non-comment line) but of course then the RTL text is mis-set. In this example, the spurious line appears after the Walt Kelly entry. This was tested under ConTeXt ver: 2021.06.18 23:37 LMTX fmt: 2021.6.21, but is not a new problem with this version. /The problem does not occur with MKIV from the same installation./ %% Demonstration of spurious line insertion on full paragraph with bidi active \setupdirections [bidi=on,method=unicode] \definepapersize [Testing] [width=6.69in] \setuppapersize [Testing] \setuplayout [width=fit, backspace=1.4in, cutspace=1in, leftmargin=0.65in, rightmargin=0.65in] \setupwhitespace [small] \setupbodyfont [libertinus,11pt] \define\SetupBib{ \spaceskip 0.7\interwordspace plus .5\interwordstretch minus \interwordshrink \definedelimitedtext [Exdent] [leftmargin=1em, indenting={yes,-1em}, indentnext=yes]} \starttext \input jojomayer \input klein \blank[2*line] \SetupBib \startExdent Kelly, Cynthia C., ed. 2005. {\em Oppenheimer and the Manhattan Project: Insights into J.~Robert Oppenheimer, ‘Father of the Atomic Bomb’}. Singapore: World Scientific. Kelly, Walt. 1953. {\em The Pogo Papers}. New York, NY: Simon & Schuster. Kuhn, Thomas S. 1962. {\em The Structure of Scientific Revolutions}. Chicago, IL: University of Chicago Press. Weinreich, Max. 1945. “Der YIVO un di problemen fun undzer tsayt (דער ייִוואָ און די פּראָבלעמען פֿון אונדזער צײַט).” {\em YIVO Bleter} 25 (1): 3--18. The title is translated as {\em The YIVO and the problems of our time}, and also as {\em The YIVO faces the post-war world}. \stopExdent \stoptext Of course, there could be some other mechanism at work here than bidi, but changing that clears the problem in this instance. (It may be the adjustment to spaceskip that I use to tighten up the text in the bibliography, and if it is, I welcome an alternative way to do that.) -- Rik
On 6/22/2021 4:52 AM, Rik Kabel wrote:
Hello all,
Last October Denis Maier complained about a spurious empty line generated with protrusion and bidi active and a full paragraph. (See https://www.mail-archive.com/ntg-context@ntg.nl/msg95916.html https://www.mail-archive.com/ntg-context@ntg.nl/msg95916.html or similar on your favorite mail archive.)
I have recreated the issue without involving protrusion. When a paragraph is very full and bidi is on, an extra line seems to be generated. Without bidi there is no extra line. This is quite sensitive, of course, to line width. I have probably 30 instances of it occurring in a tightly-set bibliography. I would assume that it may occur in any text, but have not noticed it in the body of my book.
I have created an example document that shows it. The problem disappears if bidi is turned off (comment out the first non-comment line) but of course then the RTL text is mis-set. In this example, the spurious line appears after the Walt Kelly entry.
This was tested under ConTeXt ver: 2021.06.18 23:37 LMTX fmt: 2021.6.21, but is not a new problem with this version. /The problem does not occur with MKIV from the same installation./
%% Demonstration of spurious line insertion on full paragraph with bidi active
\setupdirections [bidi=on,method=unicode]
\definepapersize [Testing] [width=6.69in] \setuppapersize [Testing]
\setuplayout [width=fit, backspace=1.4in, cutspace=1in, leftmargin=0.65in, rightmargin=0.65in]
\setupwhitespace [small]
\setupbodyfont [libertinus,11pt]
\define\SetupBib{ \spaceskip 0.7\interwordspace plus .5\interwordstretch minus \interwordshrink \definedelimitedtext [Exdent] [leftmargin=1em, indenting={yes,-1em}, indentnext=yes]}
\starttext \input jojomayer
\input klein
\blank[2*line]
\SetupBib \startExdent
Kelly, Cynthia C., ed. 2005. {\em Oppenheimer and the Manhattan Project: Insights into J.~Robert Oppenheimer, ‘Father of the Atomic Bomb’}. Singapore: World Scientific.
Kelly, Walt. 1953. {\em The Pogo Papers}. New York, NY: Simon & Schuster.
Kuhn, Thomas S. 1962. {\em The Structure of Scientific Revolutions}. Chicago, IL: University of Chicago Press.
Weinreich, Max. 1945. “Der YIVO un di problemen fun undzer tsayt (דער ייִוואָ און די פּראָבלעמען פֿון אונדזער צײַט).” {\em YIVO Bleter} 25 (1): 3--18. The title is translated as {\em The YIVO and the problems of our time}, and also as {\em The YIVO faces the post-war world}.
\stopExdent \stoptext
Of course, there could be some other mechanism at work here than bidi, but changing that clears the problem in this instance. (It may be the adjustment to spaceskip that I use to tighten up the text in the bibliography, and if it is, I welcome an alternative way to do that.) I'll send you a possible fix but it needs a bit of testing. It has to do with (probably spurious) spaces which means that the parbuilder then considers the dir node to be a reasoable breakpoint (well, kin dof okay as one gets what one asks for), but I'll be a bit more agressive in dealing with that.
Hans ----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | www.pragma-ade.nl | www.pragma-pod.nl -----------------------------------------------------------------
On 6/23/2021 18:10, Hans Hagen wrote:
[...]ome an alternative way to do that.) I'll send you a possible fix but it needs a bit of testing. It has to do with (probably spurious) spaces which means that the parbuilder
On 6/22/2021 4:52 AM, Rik Kabel wrote: then considers the dir node to be a reasoable breakpoint (well, kin dof okay as one gets what one asks for), but I'll be a bit more agressive in dealing with that.
Hans
Thank you, Hans, for looking into this so soon. I did try what you sent, but see no difference. I have eliminated the changes to spaceskip as a cause and simplified the example, removing the spaceskip and delimitedtext. New example, in which I get a spurious line after the third and fourth Walt Kelly entries: %% Demonstration of spurious line insertion on full paragraph with bidi \setupdirections [bidi=on,method=unicode] \enabletrackers [typesetters.directions] %\showmakeup [glue,dir] \definepapersize [Testing] [width=6.69in] \setuppapersize [Testing] \setuplayout [width=fit, backspace=1.4in, cutspace=1in, leftmargin=0.65in, rightmargin=0.65in] \setupwhitespace [small] \setupbodyfont [libertinus,11pt] \startbuffer[Test] Kelly, Cynthia C., ed. 2005. {\em Oppenheimer and the Manhattan Project: Insights into J.~Robert Oppenheimer, ‘Father of the Atomic Bomb’}. Singapore: World Scientific. Kelly, Walt. 1953. {\em The Pogo Papers}. New York, NY Simon & Schuster. Kelly, Walt. 1953. {\em The Pogo Papers}. New York NY Simon & Schuster. Kelly, Walt. 1953. {\em The Pogo Papers}. New York, NY: Simon & Schuster. Kelly, Walt. 1953. {\em The Pogo Papers}. New York,, NY: Simon & Schuster. Kelly, Walt. 1953. {\em The Pogo Papers}. New York; NY Simon & Schuster. Kuhn, Thomas S. 1962. {\em The Structure of Scientific Revolutions}. Chicago, IL: University of Chicago Press. Weinreich, Max. 1945. “Der YIVO un di problemen fun undzer tsayt (דער ייִוואָ און די פּראָבלעמען פֿון אונדזער צײַט).” {\em YIVO Bleter} 25 (1): 3--18. The title is translated as {\em The YIVO and the problems of our time}, and also as {\em The YIVO faces the post-war world}. \stopbuffer \starttext \getbuffer[Test] \stoptext -- Rik
On 6/23/2021 19:44, Rik Kabel wrote:
On 6/23/2021 18:10, Hans Hagen wrote:
[...]ome an alternative way to do that.) I'll send you a possible fix but it needs a bit of testing. It has to do with (probably spurious) spaces which means that the parbuilder
On 6/22/2021 4:52 AM, Rik Kabel wrote: then considers the dir node to be a reasoable breakpoint (well, kin dof okay as one gets what one asks for), but I'll be a bit more agressive in dealing with that.
Hans
Thank you, Hans, for looking into this so soon.
[...]
Hans and all, I mistakenly put the new code in the wrong instance. When I put it in the right place and remade ConTeXt, all is well as far as my test examples go and the spurious lines are gone. Thank you for the fix. -- Rik
participants (2)
-
Hans Hagen -
Rik Kabel