Patches item #663, was opened at 2006-10-21 14:29 You can respond by visiting: http://sarovar.org/tracker/?func=detail&atid=495&aid=663&group_id=106 Category: hz Group: v1.40.0 Status: Open Resolution: Accepted Priority: 5 Submitted By: The Thanh Han (hanthethanh) Assigned to: Hartmut Henkel (hhenkel) Summary: fix for bugs #481--483 Initial Comment: This is an attempt to fix bug #481--483 - This patch applies to the pristine pdftex-1.40.0-beta-20061017 sources only; if you have applied another patch(es) to the sources, please discard them and start from the fresh ones. - How to apply: ,-------- | cd /path/to/pdftex-1.40.0-beta-20061017/src | cat /path/to/the/patch | patch -p1 | ./configure | cd texk/web2c | make pdftex `-------- ----------------------------------------------------------------------

Comment By: Martin Schröder (oneiros) Date: 2006-11-06 23:59

Message: Logged In: YES user_id=421 A more general note: If possible please split continue the discussion at the affected bugs. ---------------------------------------------------------------------- Comment By: Martin Schröder (oneiros) Date: 2006-11-06 23:57 Message: Logged In: YES user_id=421 I can't reproduce the crash with the latest version of the patch. ---------------------------------------------------------------------- Comment By: Robert (schlcht) Date: 2006-11-06 22:36 Message: Logged In: YES user_id=2217 Hi, I've tested this patch and found some problems with it: (I haven't had a look at the first patch.) % ---------------------------------------------------- (1) The kerning amount is wrong/has changed. %...\kern2.32 % instead of (with previous versions) %...\kern 10.00002 % ---------------------------------------------------- (2) There's no kerning before a character if it starts a word. \hsize 9cm \tracingoutput1 \showboxdepth\maxdimen \showboxbreadth\maxdimen \knaccode\font`x=1000 \knbccode\font`x=1000 \pdfappendkern1 \pdfprependkern1 xyxyx xyxyx % ---------------------------------------------------- (3) Kerning doesn't work in lines composed with expanded fonts. \pdfprependkern1 \knbccode\font`:=500 \pdfadjustspacing2 \pdffontexpand\font 20 20 5 autoexpand yyyy: yyyy: yyyy: yyyy: yyyy: yyyy: yyyy: yyyy: yyyy: yyyy: yyyy: % ---------------------------------------------------- (4) Kerning now has this unreliable kind of paragraph scope, so that whether or not it will be applied depends on whether tex needs to hyphenate the paragraph, and whether a word contains a discretionary. \pdfadjustspacing0 \knaccode\font`s=500 \knbccode\font`s=500 \pdfappendkern1 \pdfprependkern1 tasty testing tasty testing tasty testing tasty testing tasty testing \pdfappendkern0 \pdfprependkern0 tasty testing tasty testing tasty testing tasty testing tasty testing \bye % ---------------------------------------------------- % (5) finally, this document gives a segfault % (ok without the patches): \documentclass[10pt,a4paper]{ltxdoc} \def\rmdefault{bch} \begin{document} \begin{multicols}{2} A (\textit{Bcdrea} \end{multicols} \end{document} ---------------------------------------------------------------------- Comment By: Robert (schlcht) Date: 2006-11-06 22:34 Message: Logged In: YES user_id=2217 Hi, I've tested this patch and found some problems with it: (I haven't had a look at the first patch.) % ---------------------------------------------------- (1) The kerning amount is wrong/has changed. %...\kern2.32 % instead of (with previous versions) %...\kern 10.00002 % ---------------------------------------------------- (2) There's no kerning before a character if it starts a word. \hsize 9cm \tracingoutput1 \showboxdepth\maxdimen \showboxbreadth\maxdimen \knaccode\font`x=1000 \knbccode\font`x=1000 \pdfappendkern1 \pdfprependkern1 xyxyx xyxyx % ---------------------------------------------------- (3) Kerning doesn't work in lines composed with expanded fonts. \pdfprependkern1 \knbccode\font`:=500 \pdfadjustspacing2 \pdffontexpand\font 20 20 5 autoexpand yyyy: yyyy: yyyy: yyyy: yyyy: yyyy: yyyy: yyyy: yyyy: yyyy: yyyy: % ---------------------------------------------------- (4) Kerning now has this unreliable kind of paragraph scope, so that whether or not it will be applied depends on whether tex needs to hyphenate the paragraph, and whether a word contains a discretionary. \pdfadjustspacing0 \knaccode\font`s=500 \knbccode\font`s=500 \pdfappendkern1 \pdfprependkern1 tasty testing tasty testing tasty testing tasty testing tasty testing \pdfappendkern0 \pdfprependkern0 tasty testing tasty testing tasty testing tasty testing tasty testing \bye % ---------------------------------------------------- % (5) finally, this document gives a segfault % (ok without the patches): \documentclass[10pt,a4paper]{ltxdoc} \def\rmdefault{bch} \begin{document} \begin{multicols}{2} A (\textit{Bcdrea} \end{multicols} \end{document} ---------------------------------------------------------------------- Comment By: The Thanh Han (hanthethanh) Date: 2006-11-02 09:44 Message: Logged In: YES user_id=710 Thank you Martin for looking at the patch and blessing it. I was struggling with TeX The Program more than a week before I came up with this patch and still I am not sure it doesn't break anything (I didn't claim it fixes bugs #481-484 but said this is an attempt to fix that). ---------------------------------------------------------------------- Comment By: Martin Schröder (oneiros) Date: 2006-11-01 18:37 Message: Logged In: YES user_id=421 Thanh, I looked at it, and as it does what you claim (fix #481-484) and doesn't seem to brake anything, it's in. :-) Although I'd like Hartmut to also have a look at it. hz isn't exactly my specialty. ---------------------------------------------------------------------- Comment By: The Thanh Han (hanthethanh) Date: 2006-10-28 18:01 Message: Logged In: YES user_id=710 Martin, have you looked at this patch? What do you think of it? ---------------------------------------------------------------------- Comment By: The Thanh Han (hanthethanh) Date: 2006-10-22 23:38 Message: Logged In: YES user_id=710 This is an update of the previous patch, which attemps to solve problem #484 as well (I would not call #484 a bug but a limitation -- a bug would mean that the behaviour is inconsistent with the specification). This patch is *very different* from the previous one, since now kerns inserted by the new primitives are treated as implicit kerns (they were treated as explicit kerns before this patch). This can lead to a very different behaviour, so be prepared :) ---------------------------------------------------------------------- Comment By: Martin Schröder (oneiros) Date: 2006-10-21 15:54 Message: Logged In: YES user_id=421 Thanks! Any idea for 484? :-) ---------------------------------------------------------------------- You can respond by visiting: http://sarovar.org/tracker/?func=detail&atid=495&aid=663&group_id=106