- ntg-pdftex - NTG Mailing Lists

[ pdftex-Patches-735 ] patch for simplified horizontal positional tracking
by noreply＠sarovar.org 25 Feb '07

25 Feb '07

Patches item #735, was opened at 2007-02-19 22:00 You can respond by visiting: http://sarovar.org/tracker/?func=detail&atid=495&aid=735&group_id=106 Category: Positioning Group: v1.40.0 Status: Open Resolution: None Priority: 5 Submitted By: Hartmut Henkel (hhenkel) Assigned to: Nobody (None) Summary: patch for simplified horizontal positional tracking Initial Comment: 1. simpler calculations for horizontal position update when inside []TJ arrays, now by pure additions within a fontsize/10000 raster; no rounding errors can accumulate. Gives about 7 % pdftex speed gain when typesetting fiction-type stuff. 2. writet3.c now also with standard /FontMatrix [0.001 0 0 0.001 0 0], tidied up a little (.pgc part untested). No need anymore (hope it's right) to handle PK fonts specially by adv_char_width() in pdftex.web. 3. wrong use of printf() at a few places corrected by Martin. Work in progess, not safe for production. Needs more testing, particularly with PK fonts. Regards, Hartmut ---------------------------------------------------------------------- >Comment By: Hartmut Henkel (hhenkel) Date: 2007-02-25 23:51 Message: Logged In: YES user_id=929 The previous patch 735 is now extended to general text and graphics positioning on the PDF page. pdf_set_origin() (now pdf_move_to()), pdf_set_textmatrix(), and pdf_begin_string() are rewritten. In low placement resolution (e. g. \pdfdecimaldigits=0) now the glyph placement is corrected within the []TJ array by a leading number, like [7(a)...]TJ, if required, so that always full TeX precision is kept (very early pdftex versions might have used this also). Any accumulation of rounding errors should be avoided now by the used calculation principle. Needs lots of testing. Regards, Hartmut ---------------------------------------------------------------------- You can respond by visiting: http://sarovar.org/tracker/?func=detail&atid=495&aid=735&group_id=106

1 0

[ pdftex-Bugs-737 ] Missing pfbs lead to invalid pdf
by noreply＠sarovar.org 25 Feb '07

25 Feb '07

Bugs item #737, was opened at 2007-02-25 16:03 You can respond by visiting: http://sarovar.org/tracker/?func=detail&atid=493&aid=737&group_id=106 Category: Fonts Group: v1.40.1 Status: Open Resolution: None Priority: 5 Submitted By: Martin Schröder (oneiros) Assigned to: Hartmut Henkel (hhenkel) Summary: Missing pfbs lead to invalid pdf Initial Comment: Try this file: ----------- \pdfmapline{} \nopagenumbers \font\ugqb ugqb8r at 10pt \ugqb foo \bye ----------- with ugqb8r.tfm in your texmf, but without the pfb. Here is the log from pdftex: ----------- This is pdfTeX, Version 3.141592-1.50.0-alpha-20070219 (Web2C 7.5.6) entering extended mode (./fonts20070225.tex [1]kpathsea: Running mktexpk --mfmode / --bdpi 1200 --mag 1+0/1200 --dpi 1200 ugqb8r mktexpk: don't know how to create bitmap font for ugqb8r. kpathsea: Appending font creation commands to missfont.log. ) pdfTeX warning: pdftex (file ugqb8r): Font ugqb8r at 1200 not found Output written on fonts20070225.pdf (1 page, 1018 bytes). Transcript written on fonts20070225.log. ----------- So a pdf is written, but it's invalid: > pdffonts fonts20070225.pdf ----------- name type emb sub uni object ID ------------------------------------ ------------ --- --- --- --------- Error: font resource is not a dictionary ----------- indeed the font object (#4) doesn't even exist. pdfTeX shall not generate invalid pdfs, so in this case it should abort. ---------------------------------------------------------------------- >Comment By: Martin Schröder (oneiros) Date: 2007-02-25 16:18 Message: Logged In: YES user_id=421 This all started in a thread on texhax: http://tug.org/pipermail/texhax/2007-February/007916.html ---------------------------------------------------------------------- You can respond by visiting: http://sarovar.org/tracker/?func=detail&atid=493&aid=737&group_id=106

1 0

[ pdftex-Bugs-737 ] Missing pfbs lead to invalid pdf
by noreply＠sarovar.org 25 Feb '07

25 Feb '07

Bugs item #737, was opened at 2007-02-25 16:03 You can respond by visiting: http://sarovar.org/tracker/?func=detail&atid=493&aid=737&group_id=106 Category: Fonts Group: v1.40.1 Status: Open Resolution: None Priority: 5 Submitted By: Martin Schröder (oneiros) Assigned to: Hartmut Henkel (hhenkel) Summary: Missing pfbs lead to invalid pdf Initial Comment: Try this file: ----------- \pdfmapline{} \nopagenumbers \font\ugqb ugqb8r at 10pt \ugqb foo \bye ----------- with ugqb8r.tfm in your texmf, but without the pfb. Here is the log from pdftex: ----------- This is pdfTeX, Version 3.141592-1.50.0-alpha-20070219 (Web2C 7.5.6) entering extended mode (./fonts20070225.tex [1]kpathsea: Running mktexpk --mfmode / --bdpi 1200 --mag 1+0/1200 --dpi 1200 ugqb8r mktexpk: don't know how to create bitmap font for ugqb8r. kpathsea: Appending font creation commands to missfont.log. ) pdfTeX warning: pdftex (file ugqb8r): Font ugqb8r at 1200 not found Output written on fonts20070225.pdf (1 page, 1018 bytes). Transcript written on fonts20070225.log. ----------- So a pdf is written, but it's invalid: > pdffonts fonts20070225.pdf ----------- name type emb sub uni object ID ------------------------------------ ------------ --- --- --- --------- Error: font resource is not a dictionary ----------- indeed the font object (#4) doesn't even exist. pdfTeX shall not generate invalid pdfs, so in this case it should abort. ---------------------------------------------------------------------- You can respond by visiting: http://sarovar.org/tracker/?func=detail&atid=493&aid=737&group_id=106

1 0

[CTAN] qstest
by Martin Schröder 23 Feb '07

23 Feb '07

---------- Forwarded message ---------- From: CTAN Announcements <ctan-ann(a)dante.de> Date: 22.02.2007 22:54 Subject: CTAN has a new bundle: qstest To: ctan-ann(a)dante.de On Wed, 21 Feb 2007, David Kastrup submitted the qstest bundle to CTAN. Location on CTAN: /macros/latex/contrib/qstest Summary description: Bundle for unit tests and pattern matching License type: lppl Announcement text: ---------------------------------------------------------------------- This is the public release of the qstest bundle (written for DocScape Publisher <URL:http://www.docscape.de>) (C) 2006, 2007 QuinScape GmbH <URL:http://www.quinscape.de> The bundle contains the packages `makematch' for matching patterns to targets (with a generalization in the form of pattern lists and keyword lists), and `qstest' for performing unit tests, allowing to run a number of logged tests ensuring the consistency of values, properties and call sequences during execution of test code. Both packages make extensive use of qstest in their package documentation, provided illustrated examples that are automatically verified to work as expected. Check the README file for details. David Kastrup <dak(a)gnu.org> QuinScape GmbH http://www.QuinScape.de ---------------------------------------------------------------------- This bundle is located at http://dante.ctan.org/tex-archive/macros/latex/contrib/qstest . More information is at http://tug.ctan.org/info/?id=qstest (if the package is new it may take a day for that information to appear). We are supported by the TeX Users Group http://www.tug.org . Please join a users group; see http://www.tug.org/usergroups.html . _______________________________________________ Thanks for the upload. For the CTAN Team Rainer Schöpf

1 0

[ pdftex-Bugs-736 ] font expansion leads to wrong log output of \kerns.
by noreply＠sarovar.org 22 Feb '07

22 Feb '07

Bugs item #736, was opened at 2007-02-22 03:56 You can respond by visiting: http://sarovar.org/tracker/?func=detail&atid=493&aid=736&group_id=106 Category: hz Group: v1.40.1 Status: Open >Resolution: Accepted >Priority: 3 Submitted By: Robert (schlcht) Assigned to: The Thanh Han (hanthethanh) Summary: font expansion leads to wrong log output of \kerns. Initial Comment: When applying font expansion, implicit kerns will be output *with* a space in the log file, which makes them appear like explicit kerns. The following \tracingoutput1 \pdfadjustspacing2 \pdffontexpand\font 20 20 5 autoexpand bye\bye will yield ...\tenrm b ...\kern -0.27779 ...\tenrm y ...\kern -0.27779 ...etc. instead of ...\tenrm b ...\kern-0.27779 ...\tenrm y ...\kern-0.27779 ...etc. ---------------------------------------------------------------------- Comment By: The Thanh Han (hanthethanh) Date: 2007-02-22 11:46 Message: Logged In: YES user_id=710 I confirm the bug. Will try to fix it soon. ---------------------------------------------------------------------- You can respond by visiting: http://sarovar.org/tracker/?func=detail&atid=493&aid=736&group_id=106

1 0

[ pdftex-Bugs-736 ] font expansion leads to wrong log output of \kerns.
by noreply＠sarovar.org 22 Feb '07

22 Feb '07

Bugs item #736, was opened at 2007-02-22 02:56 You can respond by visiting: http://sarovar.org/tracker/?func=detail&atid=493&aid=736&group_id=106 Category: hz Group: v1.40.1 Status: Open Resolution: None Priority: 5 Submitted By: Robert (schlcht) Assigned to: The Thanh Han (hanthethanh) Summary: font expansion leads to wrong log output of \kerns. Initial Comment: When applying font expansion, implicit kerns will be output *with* a space in the log file, which makes them appear like explicit kerns. The following \tracingoutput1 \pdfadjustspacing2 \pdffontexpand\font 20 20 5 autoexpand bye\bye will yield ...\tenrm b ...\kern -0.27779 ...\tenrm y ...\kern -0.27779 ...etc. instead of ...\tenrm b ...\kern-0.27779 ...\tenrm y ...\kern-0.27779 ...etc. ---------------------------------------------------------------------- >Comment By: The Thanh Han (hanthethanh) Date: 2007-02-22 10:46 Message: Logged In: YES user_id=710 I confirm the bug. Will try to fix it soon. ---------------------------------------------------------------------- You can respond by visiting: http://sarovar.org/tracker/?func=detail&atid=493&aid=736&group_id=106

1 0

[ pdftex-Bugs-712 ] problems with \letterspacefont
by noreply＠sarovar.org 22 Feb '07

22 Feb '07

Bugs item #712, was opened at 2006-12-12 02:53 You can respond by visiting: http://sarovar.org/tracker/?func=detail&atid=493&aid=712&group_id=106 Category: hz Group: v1.40.0 Status: Closed Resolution: Fixed Priority: 5 Submitted By: Robert (schlcht) Assigned to: The Thanh Han (hanthethanh) Summary: problems with \letterspacefont Initial Comment: % Testing the new \letterspacefont, I found: % % (1) it's certainly not correct to have the spacing % relative to character widths (look at `minimal'!) % (2) everything's shifted to the right % (3) when trying to expand the letter-spaced font, I % get the warning `font should be expanded before % its first use' (even before its first use), and % then pdftex segfaults. \pdfadjustspacing2 \hsize13cm This is normal text This is normal text This is normal text This is normal text This is normal text This is normal text {\letterspacefont\lspfont=\font 1000 %\pdffontexpand\lspfont 20 20 5 autoexpand % segfault \lspfont This isn't normal text This isn't normal text This isn't normal text This is minimal text} This is normal text This is normal text This is normal text \bye ---------------------------------------------------------------------- >Comment By: The Thanh Han (hanthethanh) Date: 2007-02-22 09:31 Message: Logged In: YES user_id=710 yes it sounds reasonable. Or we can extend the syntax of \letterspacefont so that if an optional keyword like 'nolig' is found then ligatures will be disabled. ---------------------------------------------------------------------- Comment By: Robert (schlcht) Date: 2007-02-22 02:54 Message: Logged In: YES user_id=2217 The current situation with regards to the lack of kerning of letterspaced fonts is not really satisfying. After some consideration, I therefore think it is actually better *not* to disable ligatures, and hence kerning, for letterspaced fonts. So I would propose to remove the line: set_no_ligatures(k); {disable ligatures for letter-spaced fonts} It is easy enough for the high-level interfaces (microtype, context) to (automatically) disable selected ligatures only, so that kerning for the rest of the characters is still available. What do you think? \letterspacefont\lsfont\font 100 \lsfont \tagcode\lsfont`f -1 VALID off -- \bye ---------------------------------------------------------------------- Comment By: The Thanh Han (hanthethanh) Date: 2007-02-02 13:54 Message: Logged In: YES user_id=710 the warning message can be removed. I left it there since I don't know what happens if a virtual font is loaded, used and then expanded and used again with font expansion. It should be alright to do so, but I am not very sure. Anyway, I will change the warning to be issued for non-virtual fonts only. It turned to be a design mistake that the \pdffontexpand primitive takes a base font and turns it to an expanded font. It would be better it was more like \letterspacefont, ie it should create a new font and leave the original font intact. ---------------------------------------------------------------------- Comment By: Taco Hoekwater (taco) Date: 2007-02-02 07:30 Message: Logged In: YES user_id=1608 > I hope luatex will abolish with this nonsensical connexion of ligatures and kerns? Yes. ligatures and kerns are totally separate in luatex. ---------------------------------------------------------------------- Comment By: Robert (schlcht) Date: 2007-02-01 19:02 Message: Logged In: YES user_id=2217 If the warning is misleading, should it not be removed? On a different matter: Since \letterspacefont issues set_no_ligatures, there won't be any kerning, either. There's no way around that, I guess? (I hope luatex will abolish with this nonsensical connexion of ligatures and kerns?) ---------------------------------------------------------------------- Comment By: The Thanh Han (hanthethanh) Date: 2007-01-04 15:59 Message: Logged In: YES user_id=710 The warning is misleading but harmless. The reason of the warning is that pdftex checks whether a font has not been used before expansion (its type has not been examined and is new/unknown yet). \letterspacefont sets the type of \y as a virtual font, thus pdftex thinks the font might have been used before. ---------------------------------------------------------------------- Comment By: Robert (schlcht) Date: 2007-01-04 00:40 Message: Logged In: YES user_id=2217 I'm a bit confused about the status of (3): letterspacing and expansion: When expanding a letterspaced font, pdftex issues the warning: ``font should be expanded before its first use'', but apart from that, I cannot see any problems -- the font is both letterspaced and expanded. Or is there a reason for the warning? \pdfadjustspacing2 \font\x=cmr10 \x \letterspacefont\y=\x 100 \y \pdffontexpand\y 20 20 5 autoexpand ... \bye ---------------------------------------------------------------------- Comment By: Martin Schröder (oneiros) Date: 2006-12-21 16:09 Message: Logged In: YES user_id=421 A patch for this will be included in the next version (1.40). ---------------------------------------------------------------------- Comment By: The Thanh Han (hanthethanh) Date: 2006-12-12 17:47 Message: Logged In: YES user_id=710 oops, I disabled ligature for letter-spaced font, but this should not affect the base font. Will have to rethink. Most likely the syntax will change like follows: \letterspacefont\f=cmr10 at 11pt 100 \f text ---------------------------------------------------------------------- Comment By: Robert (schlcht) Date: 2006-12-12 15:50 Message: Logged In: YES user_id=2217 % and here's % (4) ligatures will be broken up, not only in the % letter-spaced font (which is probably acceptable), % but in all subsequent uses of the base font, too fi\par {\letterspacefont\lspfont=\font 100 \lspfont fi}\par fi \bye ---------------------------------------------------------------------- Comment By: The Thanh Han (hanthethanh) Date: 2006-12-12 13:31 Message: Logged In: YES user_id=710 I fixed (1) and (2); (3) is disabled now, ie you cannot expand a letter-spaced font. Please wait for another release from Martin. ---------------------------------------------------------------------- You can respond by visiting: http://sarovar.org/tracker/?func=detail&atid=493&aid=712&group_id=106

1 0

[ pdftex-Bugs-736 ] font expansion leads to wrong log output of \kerns.
by noreply＠sarovar.org 22 Feb '07

22 Feb '07

Bugs item #736, was opened at 2007-02-22 02:56 You can respond by visiting: http://sarovar.org/tracker/?func=detail&atid=493&aid=736&group_id=106 Category: hz Group: v1.40.1 Status: Open Resolution: None Priority: 5 Submitted By: Robert (schlcht) Assigned to: The Thanh Han (hanthethanh) Summary: font expansion leads to wrong log output of \kerns. Initial Comment: When applying font expansion, implicit kerns will be output *with* a space in the log file, which makes them appear like explicit kerns. The following \tracingoutput1 \pdfadjustspacing2 \pdffontexpand\font 20 20 5 autoexpand bye\bye will yield ...\tenrm b ...\kern -0.27779 ...\tenrm y ...\kern -0.27779 ...etc. instead of ...\tenrm b ...\kern-0.27779 ...\tenrm y ...\kern-0.27779 ...etc. ---------------------------------------------------------------------- You can respond by visiting: http://sarovar.org/tracker/?func=detail&atid=493&aid=736&group_id=106

1 0

[ pdftex-Bugs-712 ] problems with \letterspacefont
by noreply＠sarovar.org 22 Feb '07

22 Feb '07

Bugs item #712, was opened at 2006-12-12 02:53 You can respond by visiting: http://sarovar.org/tracker/?func=detail&atid=493&aid=712&group_id=106 Category: hz Group: v1.40.0 Status: Closed Resolution: Fixed Priority: 5 Submitted By: Robert (schlcht) Assigned to: The Thanh Han (hanthethanh) Summary: problems with \letterspacefont Initial Comment: % Testing the new \letterspacefont, I found: % % (1) it's certainly not correct to have the spacing % relative to character widths (look at `minimal'!) % (2) everything's shifted to the right % (3) when trying to expand the letter-spaced font, I % get the warning `font should be expanded before % its first use' (even before its first use), and % then pdftex segfaults. \pdfadjustspacing2 \hsize13cm This is normal text This is normal text This is normal text This is normal text This is normal text This is normal text {\letterspacefont\lspfont=\font 1000 %\pdffontexpand\lspfont 20 20 5 autoexpand % segfault \lspfont This isn't normal text This isn't normal text This isn't normal text This is minimal text} This is normal text This is normal text This is normal text \bye ---------------------------------------------------------------------- >Comment By: Robert (schlcht) Date: 2007-02-22 02:54 Message: Logged In: YES user_id=2217 The current situation with regards to the lack of kerning of letterspaced fonts is not really satisfying. After some consideration, I therefore think it is actually better *not* to disable ligatures, and hence kerning, for letterspaced fonts. So I would propose to remove the line: set_no_ligatures(k); {disable ligatures for letter-spaced fonts} It is easy enough for the high-level interfaces (microtype, context) to (automatically) disable selected ligatures only, so that kerning for the rest of the characters is still available. What do you think? \letterspacefont\lsfont\font 100 \lsfont \tagcode\lsfont`f -1 VALID off -- \bye ---------------------------------------------------------------------- Comment By: The Thanh Han (hanthethanh) Date: 2007-02-02 13:54 Message: Logged In: YES user_id=710 the warning message can be removed. I left it there since I don't know what happens if a virtual font is loaded, used and then expanded and used again with font expansion. It should be alright to do so, but I am not very sure. Anyway, I will change the warning to be issued for non-virtual fonts only. It turned to be a design mistake that the \pdffontexpand primitive takes a base font and turns it to an expanded font. It would be better it was more like \letterspacefont, ie it should create a new font and leave the original font intact. ---------------------------------------------------------------------- Comment By: Taco Hoekwater (taco) Date: 2007-02-02 07:30 Message: Logged In: YES user_id=1608 > I hope luatex will abolish with this nonsensical connexion of ligatures and kerns? Yes. ligatures and kerns are totally separate in luatex. ---------------------------------------------------------------------- Comment By: Robert (schlcht) Date: 2007-02-01 19:02 Message: Logged In: YES user_id=2217 If the warning is misleading, should it not be removed? On a different matter: Since \letterspacefont issues set_no_ligatures, there won't be any kerning, either. There's no way around that, I guess? (I hope luatex will abolish with this nonsensical connexion of ligatures and kerns?) ---------------------------------------------------------------------- Comment By: The Thanh Han (hanthethanh) Date: 2007-01-04 15:59 Message: Logged In: YES user_id=710 The warning is misleading but harmless. The reason of the warning is that pdftex checks whether a font has not been used before expansion (its type has not been examined and is new/unknown yet). \letterspacefont sets the type of \y as a virtual font, thus pdftex thinks the font might have been used before. ---------------------------------------------------------------------- Comment By: Robert (schlcht) Date: 2007-01-04 00:40 Message: Logged In: YES user_id=2217 I'm a bit confused about the status of (3): letterspacing and expansion: When expanding a letterspaced font, pdftex issues the warning: ``font should be expanded before its first use'', but apart from that, I cannot see any problems -- the font is both letterspaced and expanded. Or is there a reason for the warning? \pdfadjustspacing2 \font\x=cmr10 \x \letterspacefont\y=\x 100 \y \pdffontexpand\y 20 20 5 autoexpand ... \bye ---------------------------------------------------------------------- Comment By: Martin Schröder (oneiros) Date: 2006-12-21 16:09 Message: Logged In: YES user_id=421 A patch for this will be included in the next version (1.40). ---------------------------------------------------------------------- Comment By: The Thanh Han (hanthethanh) Date: 2006-12-12 17:47 Message: Logged In: YES user_id=710 oops, I disabled ligature for letter-spaced font, but this should not affect the base font. Will have to rethink. Most likely the syntax will change like follows: \letterspacefont\f=cmr10 at 11pt 100 \f text ---------------------------------------------------------------------- Comment By: Robert (schlcht) Date: 2006-12-12 15:50 Message: Logged In: YES user_id=2217 % and here's % (4) ligatures will be broken up, not only in the % letter-spaced font (which is probably acceptable), % but in all subsequent uses of the base font, too fi\par {\letterspacefont\lspfont=\font 100 \lspfont fi}\par fi \bye ---------------------------------------------------------------------- Comment By: The Thanh Han (hanthethanh) Date: 2006-12-12 13:31 Message: Logged In: YES user_id=710 I fixed (1) and (2); (3) is disabled now, ie you cannot expand a letter-spaced font. Please wait for another release from Martin. ---------------------------------------------------------------------- You can respond by visiting: http://sarovar.org/tracker/?func=detail&atid=493&aid=712&group_id=106

1 0

[ pdftex-Patches-735 ] patch for simplified horizontal positional tracking
by noreply＠sarovar.org 19 Feb '07

19 Feb '07

Patches item #735, was opened at 2007-02-19 22:00 You can respond by visiting: http://sarovar.org/tracker/?func=detail&atid=495&aid=735&group_id=106 Category: Positioning Group: v1.40.0 Status: Open Resolution: None Priority: 5 Submitted By: Hartmut Henkel (hhenkel) Assigned to: Nobody (None) Summary: patch for simplified horizontal positional tracking Initial Comment: 1. simpler calculations for horizontal position update when inside []TJ arrays, now by pure additions within a fontsize/10000 raster; no rounding errors can accumulate. Gives about 7 % pdftex speed gain when typesetting fiction-type stuff. 2. writet3.c now also with standard /FontMatrix [0.001 0 0 0.001 0 0], tidied up a little (.pgc part untested). No need anymore (hope it's right) to handle PK fonts specially by adv_char_width() in pdftex.web. 3. wrong use of printf() at a few places corrected by Martin. Work in progess, not safe for production. Needs more testing, particularly with PK fonts. Regards, Hartmut ---------------------------------------------------------------------- You can respond by visiting: http://sarovar.org/tracker/?func=detail&atid=495&aid=735&group_id=106

1 0