- ntg-pdftex - NTG Mailing Lists

[ pdftex-Bugs-719 ] few buglets in build.sh
by noreply＠sarovar.org 07 Jan '07

07 Jan '07

Bugs item #719, was opened at 2007-01-02 04:39 You can respond by visiting: http://sarovar.org/tracker/?func=detail&atid=493&aid=719&group_id=106 Category: Build process Group: v1.40.0 >Status: Closed >Resolution: Fixed Priority: 5 Submitted By: Sanjoy Mahajan (sanjoy) Assigned to: Martin Schröder (oneiros) Summary: few buglets in build.sh Initial Comment: Just compiled 1.40.0 using build.sh. I had to make a few changes to get it to run on my system (Ubuntu 6.10, i386). The purpose of the chunks in the attached diff: 1. The JOBS variable wasn't used, so I changed its name be descriptive for parallel builds, and used it (and a related load variable) in the line that appends to the MAKE variable. Also, without the necessary double quotes around that line, the rest of the line after $MAKE is interpreted as a command rather than as stuff to append to MAKE (run build.sh with "--parallel" to see the problem). I used the short option names for make, to keep the line within 80 chars. 2. "-a build" is an (incomplete) AND statement, rather than a test for the "build" directory existing. The test should be "-e build" to check for existence. 3. Brace expansion is suboptimal in sh and not very portable to different reimplementations of sh. According to the bash man page: "Brace expansion introduces a slight incompatibility with historical versions of sh. sh does not treat opening or closing braces specially when they appear as part of a word, and preserves them in the output. Bash removes braces from words as a consequence of brace expansion. For example, a word entered to sh as file{1,2} appears identically in the output." I don't know the full story, but the fake sh on my system (actually 'dash') doesn't do brace expansion at all. And if the bash man page is right, even if it did the expansion, it wouldn't help since build.sh uses braces in the middle of a word. So I rewrote the two uses of brace expansion to avoid it. The resulting "build.sh --parallel" worked fine in my limited tests. It still complained that it couldn't find the texmf tree, but I figure that's okay since pdftex'll eventually (I hope) use kpathsea and texmf.cnf. -Sanjoy ---------------------------------------------------------------------- >Comment By: Martin Schröder (oneiros) Date: 2007-01-07 22:41 Message: Logged In: YES user_id=421 Thanks. I'll include your changes in the next version. ---------------------------------------------------------------------- You can respond by visiting: http://sarovar.org/tracker/?func=detail&atid=493&aid=719&group_id=106

1 0

Re: [NTG-pdftex] [tex-live] \pdffontattr in pdftex-1.40.0-rc4
by Thanh Han The 06 Jan '07

06 Jan '07

Hi, sorry for the late response. The patch by Hartmut should solve the problem. to Hartmut: > > > pdfTeX warning (\pdffontattr): font \^^@ has conflicting attributes > is this a nullfont? how does one get it? this font has been loaded from a virtual font, hence it doesn't get any control sequence. I made another patch, which is more or less the same as what Hartmut did, plus some changes to make the warning more meaningful. Don't be afraid by the length of the patch, it just turns the module @<Print the font identifier for |font(p)|@> to a function named print_font_identifier() so it can be called with different parameters. Thanh --- /home/thanh/s2/pdftex.web 2006-12-20 16:16:42.000000000 +0100 +++ pdftex.web 2007-01-04 18:26:42.000000000 +0100 @@ -4062,7 +4062,33 @@ @ Boxes, rules, inserts, whatsits, marks, and things in general that are sort of ``complicated'' are indicated only by printing `\.{[]}'. -@p procedure short_display(@!p:integer); {prints highlights of list |p|} +@p +procedure print_font_identifier(f: internal_font_number); +begin + if pdf_font_blink[f] = null_font then + print_esc(font_id_text(f)) + else + print_esc(font_id_text(pdf_font_blink[f])); + if pdf_tracing_fonts > 0 then begin + print(" ("); + print(font_name[f]); + if font_size[f] <> font_dsize[f] then begin + print("@@"); + print_scaled(font_size[f]); + print("pt"); + end; + print(")"); + end else + if pdf_font_expand_ratio[f] <> 0 then begin + print(" ("); + if pdf_font_expand_ratio[f] > 0 then + print("+"); + print_int(pdf_font_expand_ratio[f]); + print(")"); + end; +end; + +procedure short_display(@!p:integer); {prints highlights of list |p|} var n:integer; {for replacement counts} begin while p>mem_min do begin if is_char_node(p) then @@ -4071,7 +4097,7 @@ begin if (font(p)<font_base)or(font(p)>font_max) then print_char("*") @.*\relax@> - else @<Print the font identifier for |font(p)|@>; + else print_font_identifier(font(p)); print_char(" "); font_in_short_display:=font(p); end; print_ASCII(qo(character(p))); @@ -4110,7 +4136,7 @@ begin if p>mem_end then print_esc("CLOBBERED.") else begin if (font(p)<font_base)or(font(p)>font_max) then print_char("*") @.*\relax@> - else @<Print the font identifier for |font(p)|@>; + else print_font_identifier(font(p)); print_char(" "); print_ASCII(qo(character(p))); end; end; @@ -6771,31 +6797,6 @@ Meanwhile, this is a convenient place to catch up on something we were unable to do before the hash table was defined: -@<Print the font identifier for |font(p)|@>= -begin - if pdf_font_blink[font(p)] = null_font then - print_esc(font_id_text(font(p))) - else - print_esc(font_id_text(pdf_font_blink[font(p)])); - if pdf_tracing_fonts > 0 then begin - print(" ("); - print(font_name[font(p)]); - if font_size[font(p)] <> font_dsize[font(p)] then begin - print("@@"); - print_scaled(font_size[font(p)]); - print("pt"); - end; - print(")"); - end else - if pdf_font_expand_ratio[font(p)] <> 0 then begin - print(" ("); - if pdf_font_expand_ratio[font(p)] > 0 then - print("+"); - print_int(pdf_font_expand_ratio[font(p)]); - print(")"); - end; -end - @* \[19] Saving and restoring equivalents. The nested structure provided by `$\.{\char'173}\ldots\.{\char'175}$' groups in \TeX\ means that |eqtb| entries valid in outer groups should be saved @@ -14943,7 +14944,7 @@ begin if (font(p)<font_base)or(font(p)>font_max) then print_char("*") @.*\relax@> - else @<Print the font identifier for |font(p)|@>; + else print_font_identifier(font(p)); print_char(" "); font_in_short_display:=font(p); end; print_ASCII(qo(character(p))); @@ -19340,11 +19341,15 @@ if pdf_char_marked(k, j) then pdf_mark_char(i, j); if (length(pdf_font_attr[i]) = 0) and (length(pdf_font_attr[k]) <> 0) then - pdf_font_attr[i] := pdf_font_attr[k]; - if not ((pdf_font_attr[i] = pdf_font_attr[k]) or - str_eq_str(pdf_font_attr[i], pdf_font_attr[k])) then begin - pdf_warning("\pdffontattr", "font ", false, false); - print_esc(font_id_text(k)); + pdf_font_attr[i] := pdf_font_attr[k] + else if (length(pdf_font_attr[k]) = 0) and (length(pdf_font_attr[i]) <> 0) then + pdf_font_attr[k] := pdf_font_attr[i] + else if (length(pdf_font_attr[i]) <> 0) and (length(pdf_font_attr[k]) <> 0) and + not str_eq_str(pdf_font_attr[i], pdf_font_attr[k]) then begin + pdf_warning("\pdffontattr", "fonts ", false, false); + print_font_identifier(i); + print(" and "); + print_font_identifier(k); print(" has conflicting attributes"); print_ln; print_ln; end; end; On Thu, Dec 28, 2006 at 06:38:51AM +0100, Hartmut Henkel wrote: > On Thu, 28 Dec 2006, Hartmut Henkel wrote: > > > On Wed, 27 Dec 2006, Reinhard Kotucha wrote: > > > > > I'm trying to run the vntex testfiles with the current pdftex > > > release recently uploaded to the TeXLive repository by Martin. > > > > > > Can anybody tell me what the warnings mean? > > > > > > pdfTeX warning (\pdffontattr): font \^^@ has conflicting attributes > > is this a nullfont? how does one get it? > > > in the intended cases it means, that it was tried to add two different > > font attributes to the same font dictionary. But here is a case where > > this happens even without: > > > > \pdfcompresslevel=0 > > \nopagenumbers > > \font\aa cmr10 at 10pt > > \pdffontattr\aa {foo} > > \aa abc > > \font\bb cmr10 at 11pt > > %\pdffontattr\bb {bar} > > \bb abc > > \bye > > > > Both TeX fonts refer to the same font dictionary. So there is only one > > font, and this gets the attribute "foo". > > > > > I get a lot of these messages. The (pdf) output seems to be ok. > > > > The above source creates the right PDF. Maybe we need to get rid of > > the warning here. > > Here is the try of a patch, which also shows what happens with the above > file. Reinhard, does it remove the warnings in your case? > > --- pdftex.web.orig 2006-12-21 17:19:20.000000000 +0100 > +++ pdftex.web 2006-12-28 06:07:11.771455920 +0100 > @@ -19341,6 +19341,8 @@ > pdf_mark_char(i, j); > if (length(pdf_font_attr[i]) = 0) and (length(pdf_font_attr[k]) <> 0) then > pdf_font_attr[i] := pdf_font_attr[k]; > + if (length(pdf_font_attr[k]) = 0) and (length(pdf_font_attr[i]) <> 0) then > + pdf_font_attr[k] := pdf_font_attr[i]; > if not ((pdf_font_attr[i] = pdf_font_attr[k]) or > str_eq_str(pdf_font_attr[i], pdf_font_attr[k])) then begin > pdf_warning("\pdffontattr", "font ", false, false); > > An alternative (more complicated) might be that one keeps font > dictionaries separate when the user wants different attributes. No idea > what the typical use of these attributes is. > > Regards, Hartmut

2 1

Re: [NTG-pdftex] [Fwd: [NTG-context] crash with pdftex-1.40]
by Thanh Han The 05 Jan '07

05 Jan '07

I can reproduce the problem, but have no idea how to fix it. pdftex crashes when a pclose is called in aclose(), file openclose.c. But I don't know anything about pipes in pdftex -- it was done by other people. Thanh On Thu, Jan 04, 2007 at 11:08:50PM +0100, Martin Schroeder wrote: > 2007/1/4, Hans Hagen <pragma(a)wxs.nl>: > > While playing around with the new shell escaping in pdftex-1.40, I got > > the following crash: > > I can't reproduce this. > > Best > Martin > _______________________________________________ > ntg-pdftex mailing list > ntg-pdftex(a)ntg.nl > http://www.ntg.nl/mailman/listinfo/ntg-pdftex

4 5

Re: [NTG-pdftex] [Fwd: [NTG-context] crash with pdftex-1.40]
by Akira Kakuto 05 Jan '07

05 Jan '07

Hi Taco, >>> But this should be enough to reproduce it: >>> >>> \starttext >>> \input "\string|echo -n bla" >>> \stoptext > > I can reproduce the problem, but have no idea how to fix it. > > pdftex crashes when a pclose is called in aclose(), file > > openclose.c. But I don't know anything about pipes in pdftex > > I'll check it out. On windows (pdftex 1.40.0, ConTeXt 2007.01.02 13:44 MK II), \starttext \input "\string|echo This is a test." \stoptext works fine. The result is one page pdf with the sentence "This is a test." Thanks, Akira

1 0

[Fwd: [NTG-context] crash with pdftex-1.40]
by Hans Hagen 04 Jan '07

04 Jan '07

2 1

[ pdftex-Bugs-712 ] problems with \letterspacefont
by noreply＠sarovar.org 04 Jan '07

04 Jan '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: 2007-01-04 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-712 ] problems with \letterspacefont
by noreply＠sarovar.org 04 Jan '07

04 Jan '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: 2007-01-04 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

Re: [NTG-pdftex] [tex-live] \pdffontattr in pdftex-1.40.0-rc4
by Akira Kakuto 03 Jan '07

03 Jan '07

Hi Reinhard, > > Here is the try of a patch, which also shows what happens with the > > above file. Reinhard, does it remove the warnings in your case? > > Yes, after I applied your patch, the warnings don't appear any more. In this respect, the released version seems to be the same as rc4: I see the warning with cmap.sty. Thanks, Akira

6 8

bug
by Hans Hagen 03 Jan '07

03 Jan '07

Hi, ! pdfTeX error (\pdfglyphtounicode): not allowed in DVI mode (\pdfoutput <= 0). l.1 \pdfglyphtounicode this is wrong i.e. it should be possible to generate a format which can produce both dvi and pdf and in that case one should be able to preload the unicode mappings test by taco also indicate that those settings don't survive a dump/undump Hans ----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | fax: 038 477 53 74 | www.pragma-ade.com | www.pragma-pod.nl -----------------------------------------------------------------

2 2

[ pdftex-Bugs-719 ] few buglets in build.sh
by noreply＠sarovar.org 02 Jan '07

02 Jan '07

Bugs item #719, was opened at 2007-01-02 04:39 You can respond by visiting: http://sarovar.org/tracker/?func=detail&atid=493&aid=719&group_id=106 >Category: Build process Group: v1.40.0 Status: Open >Resolution: Accepted Priority: 5 Submitted By: Sanjoy Mahajan (sanjoy) >Assigned to: Martin Schröder (oneiros) Summary: few buglets in build.sh Initial Comment: Just compiled 1.40.0 using build.sh. I had to make a few changes to get it to run on my system (Ubuntu 6.10, i386). The purpose of the chunks in the attached diff: 1. The JOBS variable wasn't used, so I changed its name be descriptive for parallel builds, and used it (and a related load variable) in the line that appends to the MAKE variable. Also, without the necessary double quotes around that line, the rest of the line after $MAKE is interpreted as a command rather than as stuff to append to MAKE (run build.sh with "--parallel" to see the problem). I used the short option names for make, to keep the line within 80 chars. 2. "-a build" is an (incomplete) AND statement, rather than a test for the "build" directory existing. The test should be "-e build" to check for existence. 3. Brace expansion is suboptimal in sh and not very portable to different reimplementations of sh. According to the bash man page: "Brace expansion introduces a slight incompatibility with historical versions of sh. sh does not treat opening or closing braces specially when they appear as part of a word, and preserves them in the output. Bash removes braces from words as a consequence of brace expansion. For example, a word entered to sh as file{1,2} appears identically in the output." I don't know the full story, but the fake sh on my system (actually 'dash') doesn't do brace expansion at all. And if the bash man page is right, even if it did the expansion, it wouldn't help since build.sh uses braces in the middle of a word. So I rewrote the two uses of brace expansion to avoid it. The resulting "build.sh --parallel" worked fine in my limited tests. It still complained that it couldn't find the texmf tree, but I figure that's okay since pdftex'll eventually (I hope) use kpathsea and texmf.cnf. -Sanjoy ---------------------------------------------------------------------- You can respond by visiting: http://sarovar.org/tracker/?func=detail&atid=493&aid=719&group_id=106

1 0