- ntg-pdftex - NTG Mailing Lists

bug in XObject
by Vit Zyka 03 Jun '05

03 Jun '05

Hallo pdftexdevelopers, there is a bug in \pdfxform. See the next example: ----------------------- \pdfcompresslevel=0 \footline={\hfill}% \setbox0\vbox{\hsize8.75cm\noindent Ahoj} \pdfxform attr{/Matrix [1.2 0.25 0.3 1 0 0]} 0 \pdfrefxform\pdflastxform \end ------------------------- The result XObject looks like: ------------------------ 1 0 obj << /Type /XObject /Subtype /Form /Matrix [1.2 0.25 0.3 1 0 0] /BBox [0 0 248.031 8.856] /FormType 1 /Matrix [1 0 0 1 0 0] /Resources 6 0 R /Length 69 >> stream 1 0 0 1 0 -833.034 cm BT /F1 9.963 Tf 0 834.971 Td[(Aho)-55(j)]TJ ET endstream endobj ------------------------ See the double /Matrix entry. That results in ignoring transformation by Acrobat (R) v.6 and 7. (v.4 and v.5 is OK). I believe that it is (exceptionally) bug in pdftex not in Acrobat ;) vit

5 5

[ pdftex-Bugs-362 ] pdfobj produces extra /Filter for object streams with own filter
by noreply＠sarovar.org 02 Jun '05

02 Jun '05

Bugs item #362, was opened at 2005-06-02 12:55 You can respond by visiting: http://sarovar.org/tracker/?func=detail&atid=493&aid=362&group_id=106 Category: None Group: v1.21a Status: Open Resolution: None Priority: 5 Submitted By: Jens Elstner (daguy) Assigned to: Nobody (None) Summary: pdfobj produces extra /Filter for object streams with own filter Initial Comment: When I use the following code with pdftex and activated compression \pdfcompresslevel=9 \pdfobj stream% attr{/Filter/ASCIIHexDecode/FunctionType 0/Encode[0 255]/BitsPerSample 8/Domain[0 1]/Size[2]/Range[0 1 0 1 0 1 0 1]/Decode[0 1 0 1 0 1 0 1]}% {00000000FFFFFFFF}% output tint function The generated pdf file contains the object as shown below 3 0 obj << /Filter/ASCIIHexDecode/FunctionType 0/Encode[0 255]/BitsPerSample 8/Domain[0 1]/Size[2]/Range[0 1 0 1 0 1 0 1]/Decode[0 1 0 1 0 1 0 1] /Length 17 /Filter /FlateDecode >> stream x┌30Ç'7'c endobj The compression added an extra /Filter line and produces an error or does not work the way it was intended anymore. Proposed fix for this would be to check the <attr- spec> for /Filter and omit any compression if a user filter function is present already. Yours sincerely, Jens Elstner ---------------------------------------------------------------------- Comment By: Heiko Oberdiek (oberdiek) Date: 2005-06-02 14:52 Message: Logged In: YES user_id=1276 Hello, not a fix, just a workaround: \begingroup \pdfcompresslevel=0 % then pdfTeX does not % add /Filter itself \immediate\pdfobj stream attr{/Filter/...}{...}\relax \endgroup Yours sincerely Heiko <oberdiek(a)uni-freiburg.de> ---------------------------------------------------------------------- You can respond by visiting: http://sarovar.org/tracker/?func=detail&atid=493&aid=362&group_id=106

1 0

[ pdftex-Bugs-362 ] pdfobj produces extra /Filter for object streams with own filter
by noreply＠sarovar.org 02 Jun '05

02 Jun '05

Bugs item #362, was opened at 2005-06-02 12:55 You can respond by visiting: http://sarovar.org/tracker/?func=detail&atid=493&aid=362&group_id=106 Category: None Group: v1.21a Status: Open Resolution: None Priority: 5 Submitted By: Jens Elstner (daguy) Assigned to: Nobody (None) Summary: pdfobj produces extra /Filter for object streams with own filter Initial Comment: When I use the following code with pdftex and activated compression \pdfcompresslevel=9 \pdfobj stream% attr{/Filter/ASCIIHexDecode/FunctionType 0/Encode[0 255]/BitsPerSample 8/Domain[0 1]/Size[2]/Range[0 1 0 1 0 1 0 1]/Decode[0 1 0 1 0 1 0 1]}% {00000000FFFFFFFF}% output tint function The generated pdf file contains the object as shown below 3 0 obj << /Filter/ASCIIHexDecode/FunctionType 0/Encode[0 255]/BitsPerSample 8/Domain[0 1]/Size[2]/Range[0 1 0 1 0 1 0 1]/Decode[0 1 0 1 0 1 0 1] /Length 17 /Filter /FlateDecode >> stream x┌30Ç'7'c endobj The compression added an extra /Filter line and produces an error or does not work the way it was intended anymore. Proposed fix for this would be to check the <attr- spec> for /Filter and omit any compression if a user filter function is present already. Yours sincerely, Jens Elstner ---------------------------------------------------------------------- You can respond by visiting: http://sarovar.org/tracker/?func=detail&atid=493&aid=362&group_id=106

1 0

unwanted font replacement during inclussion
by Vit Zyka 02 Jun '05

02 Jun '05

Hi, I noticed the next strange behaviour. Let us create very simple file test.tex: ------------------ \font\bf=cmbx12 Normal text. \bf Bold text. \end ------------------ Now include this file by plain pdftex: ------------------ \pdfximage{test.pdf} \pdfrefximage\pdflastximage \end ------------------ Everything is correct. No let us try to insert the file by context: ------------------ \starttext %\font\bf=cmbx12 %Normal text. \bf Bold text. \pdfximage{test.pdf} \pdfrefximage\pdflastximage \stoptext ------------------ And that is!!! Text 'bold text' is typeset with totaly different font from cmbx12, so it is smashed by wrong metric! See the attached file. If I uncomment two lines, so the font cmbx12 is not only present in the included file, the result is correct. Where is the problem? Something mixing cm and lm? I though that included file comes with all font resources. Hmmmm, now I feel it is more pdftex-related then context-related question. vit (pdfetex v1.30) -- Vít Zýka --- http://typokvitek.com Automatic document typesetting Automaticka sazba dokumentu -- --

4 3

Re: [NTG-pdftex] inclusion of PDF with hz-alg expandeded font
by Vit Zyka 02 Jun '05

02 Jun '05

The Thanh Han wrote: > remember when I fixed it. Anyway, can you please try it with > the latest snapshot (djgpp is also available). If you need Jupiii! Yes, ~1.30 really fixes the problem. That is very good news. Thank you, Thanh, very much. (Also good news for me is, that pdftex preserves form objects with geometrical transformations during PDF inclusion; If someday it preserves also annotations it would be really amazing!) > help with getting the djgpp binary work, please feel free to Thanks to Akira Kakuto who is maintaining the fresh food in the http://www.fsci.fuk.kindai.ac.jp/kakuto/win32-ptex/web2c75.html I do not need translating. Vit

2 1

unwanted font replacement during inclussion
by Vit Zyka 01 Jun '05

01 Jun '05

Hi, I noticed the next strange behaviour. Let us create very simple file test.tex: ------------------ \font\bf=cmbx12 Normal text. \bf Bold text. \end ------------------ Now include this file by plain pdftex: ------------------ \pdfximage{test.pdf} \pdfrefximage\pdflastximage \end ------------------ Everything is correct. No let us try to insert the file by context: ------------------ \starttext %\font\bf=cmbx12 %Normal text. \bf Bold text. \pdfximage{test.pdf} \pdfrefximage\pdflastximage \stoptext ------------------ And that is!!! Text 'bold text' is typeset with totaly different font from cmbx12, so it is smashed by wrong metric! See the attached file. If I uncomment two lines, so the font cmbx12 is not only present in the included file, the result is correct. Where is the problem? Something mixing cm and lm? I though that included file comes with all font resources. Hmmmm, now I feel it is more pdftex-related then context-related question. vit (pdfetex v1.30) -- Vít Zýka --- http://typokvitek.com Automatic document typesetting Automaticka sazba dokumentu -- --

1 0

manual update
by Pawel Jackowski 01 Jun '05

01 Jun '05

Hi there, A couple of tiny corrections commited to pdfTeX manual. I'm going to add samplepdf file with resources to CVS, somewhere near the manual sources. Objections? Regards, -- Pawe/l Jackowski P.Jackowski(a)gust.org.pl

2 1

Re: [NTG-pdftex] inclusion of PDF with hz-alg expandeded font
by Vit Zyka 01 Jun '05

01 Jun '05

The Thanh Han wrote: > ah, sorry, I misunderstood the problem. It seems to be a bug > of pdftex with pdf inclusion (handling fonts in embedded > pdf's). I am not sure when this was fixed, but when I run > your test file, the output seems ok (see attachment). So I > suggest to upgrade your pdftex binaries (sorry that I don't > know any better > solution). Strange, I read in your PDF: PDF Producer: pdfetex 1.21b That is (hopefully) my version of pdfetex. So let us return to the minimal example. I send you .tex, .log, and .pdf files for comparision. In addition I notice another strange behaviour with inclusion. Type 1 font cmb12 (embeded subset) was after inclusion replaced by Times New Roman Bold(?). The rest cm* font are OK. It needs more investigation from me, so, please, can you focus on the hz problem? Thank you very much Vit

1 0

[ pdftex-Bugs-359 ] pdftex generates errors when it does not have to - eg \pdfinfo
by noreply＠sarovar.org 30 May '05

30 May '05

Bugs item #359, was opened at 2005-05-30 19:51 You can respond by visiting: http://sarovar.org/tracker/?func=detail&atid=493&aid=359&group_id=106 Category: None Group: None Status: Open Resolution: None Priority: 3 Submitted By: Taco Hoekwater (taco) Assigned to: Nobody (None) Summary: pdftex generates errors when it does not have to - eg \pdfinfo Initial Comment: The following submitted on behalf of Jonathan Fine: pdftex generates errors when it does not have to. This fault prevents the build of Python docs on some platforms. The fix may be to remove the error-generating code· For more details, see my posting to comp.text.tex. Google link: <http://groups-beta.google.com/group/comp.text.tex/msg/4b4842202ce88cda?hl=en> Copy and paste ============== There is, in my opinion, a problem with pdftex. (This is not to say that the macros used could not be improved.) pdftex exits on \pdfinfo error. TeX does not exit on error. == $ pdftex This is pdfTeX, Version 3.14159-1.10b (Web2C 7.4.5) **\pdfoutput 0 {/usr/share/texmf/pdftex/config/pdftex.cfg} *\pdfinfo pdfTeX error (ext1): \pdfinfo used while \pdfoutput is not set. <*> \pdfinfo No pages of output. Transcript written on texput.log. == The present semantics of \pdfinfo are confused. For example, I _can_ use \pdfinfo and generate dvi! I just have to be a bit sneaky. == $ pdftex This is pdfTeX, Version 3.14159-1.10b (Web2C 7.4.5) **\pdfoutput 1 {/usr/share/texmf/pdftex/config/pdftex.cfg} *\pdfinfo {} *\pdfoutput 0 *\shipout\hbox{} [1] *\end Output written on texput.dvi (1 page, 128 bytes). == Here, if pdf is not being generated, pdfinfo is ignored. Suppose the semantics of \pdfinfo were something like: \pdfinfo{ ... } Stores information that is written to the PDF file, if pdftex generates a PDF file. Otherwise ignored. Then the error that prompted this thread would not have occurred. -- ---------------------------------------------------------------------- You can respond by visiting: http://sarovar.org/tracker/?func=detail&atid=493&aid=359&group_id=106

1 0

another snapshot
by The Thanh Han 28 May '05

28 May '05

Hi, I have uploaded another snapshot with the patch removing the spurious space after some keywords by Hartmut. Thanh

1 0