Hi, We were trying to convert a PS file that was captured from the SAP Spool to PDF, using GhostScript. But the 2 latin characters did not convert properly. The GhostScript Version we used is 8.14. Command used to Convert the PS file to PDF is: gswin32 -q -IC:/hp-dazel/os/lib/FONTS/Soft_Horizons -I../lib -dSAFER -dNOPAUSE - dBATCH -sOutputFile=file.pdf -sDEVICE=pdfwrite -c .setpdfwrite -f file.ps I am attaching the original PS file (file.ps), the converted PDF file (file.pdf) and the bmp file (ps2pdf_problem.bmp) which shows the difference between the 2 files. The arrows indicate, the characters which are not converted properly. Thanks Subrahmanyan.
Created attachment 662 [details] Original PS file Original PS file captured from the SAP spool.
Created attachment 663 [details] Converted PDF file Converted pdf file.
Created attachment 664 [details] Converted PDF file Converted pdf file.
Created attachment 665 [details] JPG file showing the differences JPG File showing the differences
Can't reproduce in 8.14 with the GPL fonts. Will try will Soft Horizon fonts.
Reproducable with CVS HEAD but only using Soft Horizon's fonts. Assigning to Igor for investigation.
I also noticed a problem with regular GS fonts in the text "ICO/DIC". After the distillation with GS the first "C" dissappears.
Created attachment 674 [details] A reduced test It only prints the text "I¨^O". Distilling with GS with regular GS fonts, the 2nd character is lost.
Both "Converted PDF file" are same.
Created attachment 675 [details] A PDF created with Adobe 6.0 Adobe Distiler 6.0 also has problems with the PS file. Particularly it skips Ccaron. Ghostscript doesn't skip Ccaron and displays its PDF correctly. But Adobe can't display Ccaron. I don't see anything wrong in the PDF created by GS. Likely we observe an Adobe problem.
Patch http://www.ghostscript.com/pipermail/gs-cvs/2004-May/004489.html makes URW fonts and SoftHorizon fonts to give same result. I still observe a problem with ccaron and Ccaron in Adobe Viewer on Windows. I believe it is Adobe problem. I intend to close this bug. Any objections ?
Same patch is applicable for to GS_8_1X. Here is exact instance : http://www.ghostscript.com/pipermail/gs-cvs/2004-May/004490.html
Hi, I tried viewing the converted PDF file with GSView and GS 8.14. I could see the problem existing. So it may not be the Adobe problem. Thanks & Regards, Subrahmanyan.
Dear Subrahmanyan, Did you apply the patch before testing GS ? The patch is attached to here. Igor.
Hi, Yes, I applied the Patch, but still it didn't work for me. Subrahmanyan.
Please give me a precise information : 1. Convert Original PS with unpatched GS and view it with unpatched GS. 2. Convert Original PS with patched GS and view it with patched GS. Do they give same raster ? I believe they should be different. Please attach both rasters.
Dear Igor, Sorry for the late response. Yes, I tried both the cases. And I could see the differences. When I converted the PS file with patched GS and view it with the Patched GS, I could see the document correctly. But with Adobe Acrobat, it is not showing properly. So should we need to contact Adobe for this problem? Thanks & Regards Subrahmanyan.
Yes, I think so. If you use Windows, the problwem may also happen in Adobe to Microsoft interfaces, I mean the emulation of PS fonts with Microsoft fonts. Here is my summary : The mentioned patch fixes the problem in pdfwrite. The patched revision creates a correct PDF. That PDF views correctly in Ghostscript with URW fonts. Meanwhile there still exist 2 problems, which to be reported to 3d parties : 1. The characters ccaron, Ccaron doesn't view in Adobe Reader on Windows. Please report to Adobe. 2. The characters ccaron, Ccaron doesn't view in GS with Soft Horizon fonts. Ghostscript prints "Substituting .notdef for Ccaron." because the glyph is missed in the font. Please report to Soft Horizon.
Hi, We have a similar problem. Here pdf file converted from PS using GS 8.13 or GS 8.14, is not displayed properly in Adobe Acrobat 6.0. It is observed that the output PDF file has one unknown character at the starting of each line. I am attaching the input PS file and output PDF file. Can you please go through this and find out, if it is a GS problem or Acrobat problem. Thanks Subrahmanyan.
Created attachment 692 [details] Input PS File This is original PS File.
Created attachment 693 [details] Output PDF File This is the Output PDF File, which has an unknown extra character at the begining of each line.
Can't reproduce. The attached "Output PDF file" views fine on Windows XP with Acrobat reader 5, Adobe Acrobat 6 Professional, and with Adobe reader 6. Converting the "Input PS file" with gs 7.14 and CVS HEAD got a document, which also viewes fine. Please explain more detailedly, what problem do you observe. Restoring the bug status to the previous state.
Oops, I've meant gs8.14.
Created attachment 695 [details] BMP file showing the differences. BMP File showing the differences in PS File and the PDF file. One unknown character is seen in the begining of each line, which is marked with read arrow in the BMP file. That character looks like small sqaure with 2 diagonal lines. I tried viewing the PDF file in Win XP also, with Acrobat 6.0. There also I could see the unknown characters.
I do not observe the boxes you show in the "BMP file showing the differences". The PDF file contains the charater 0x20 in that positions, and does not embed the related font, which is Courier with no Encoding. The viewer must paint space character. In your case likely it paints .notdef . This definitely is a problem of the viewer. I guess this is a problem of a Windows NT localization, which could be contributed either by Microsoft for a specific locale or a specific localized version of Windows, or by Adobe with a localized Acrobat Reader, or by a 3d party IME software or by a 3d party font software.
Created attachment 707 [details] The "Input PS File" converted by Adobe Distiller 6 I've converted you "Input PS File" with Adobe Distiller 6 on my system and unpacked streams in there with the pdfinflt.ps tool. Do you observe boxes with it ?
No, I don't see any extra characters in the file which you created using Adobe Distiller 6. This file is perfectly fine.
Created attachment 709 [details] The last PDF with no Encoding I've edited the Adobe's PDF with removing Encoding entries. This makes it closer to the Ghostscript output. Please check whether your viewer shows boxes with it.
As I know, you could reproduce the problem with Windows XP. Please give me more information about that system : - the version number; - A service pack version number; - The PDF viewer version; - What locales are installed ? - Which installed 3d party products could contrubute/modify fonts ?
When you see boxes in Adobe Reader, please open File/DocumentProperties/Fonts and copy the table to here.
I don't see the boxes, with the attached PDF file, which you edited by removing Encoding entries.
This is what I saw in the "File->DocumentProperties->Fonts" Dialog, when I see boxes in Adobe Reader. ------------------------------------- Courier Type: Type 1 Encoding: Standard Actual Font: CourierStd Actual Font Type: Type 1 Times-Italic Type: Type 1 Encoding: Standard Actual Font: TimesNewRomanPS-ItalicMT Actual Font Type: True Type Helvetica-BoldOblique Type Type 1 Encoding: Standard Actual Font: Arial-BoldItaicMT Actual Font Type: TrueType Courier-Bold Type: Type 1 Encoding: Standard Actual Font: CourierStd-Bold Actual Font Type: Type 1 Times-Bold Type: Type 1 Encoding: Standard Actual Font: TimesNewRomanPS-BoldMT Actual Font Type: True Type Courier-Oblique Type: Type 1 Encoding: Standard Actual Font: CourierStd-Oblique Actual Font Type: Type 1 Times-BoldItalic Type: Type 1 Encoding: Standard Actual Font: TimesNewRomanPS-BoldItalicMT Actual Font Type: TrueType Courier-BoldOblique Type: Type 1 Encoding: Standard Actual Font: CourierStd-BoldOblique Actual Font Type: Type 1 Helvetica Type: Type 1 Encoding: Standard Actual Font: ArialMT Actual Font Type: TrueType Helvetica-Bold Type: Type 1 Encoding: Standard Actual Font: Arial-BoldMT Actual Font Type: TrueType Times-Roman Type: Type 1 Encoding: Standard Actual Font: TimesNewRomanPSMT Actual Font Type: TrueType Helvetica-Oblique Type: Type 1 Encoding: Custom Actual Font: Arial-ItalicMT Actual Font Type: TrueType -------------------------------------
Regarding to the Comment #32 : On my system all fonts are substituted with Type 1 fonts. This is a difference with the user's system, which uses some True Types. Please attach your file "*:\Program Files\Adobe\Acrobat 6.0\Resource\Font\AdobeFnt*.lst" or its analogue. Thank you.
Created attachment 714 [details] AdobeFnt07.lst file This is the file, which you asked me to attach, in comment #33.
The problem described in the Comment #19 is moved to the bug 687510. Please write all further related issues to there, because this bug was initially opened for another problem.
Created attachment 1638 [details] cust.ps Attaching a test file receieved from the customer. It almost same as receieved before, but we need to keep the file name for references. When cust.ps is converted with Adobe Distiller and then the generated PDF is viewed with Adobe Reader, the 2 characters disappear. Therefore I conclude that the problem lies in Adobe Reader. Likely they missed the character Ccaron from their font Helvetica-Bold. Ghostscript 8.00 can't convert that text due to another minor problem, and therefore it replaces the font Helvetica-Bold with a bitmap font, which is synthesized while the conversion. For doing that, Ghostscript uses an URW font NymbusSanL-Bold, which does include Ccaron. Therefore you can see it in cast-good.pdf . Thus what you need is to use the URW font instead the Adobe font either in the conversion, or while viewing the result. I guess you prefer the fist. As a workaround I suggest to apply distiller parameters : << /NeverEmbed [false /Helvetica-Bold] >> setdistillerparams I've tried it and got a good result - see attachment. As a next step, maybe you'll need it to be done automatically. I mean that you may want Ghostscript to recognize missed glyphs and apply /NeverEmbed [false ...] automatically. Unfortunately I have no idea how to make Ghostscript to know which glyphs Adobe missed. Therefore I can't suggest an automatic method.
Created attachment 1639 [details] cur.pdf A PDF generated by Ghostscript from cust.ps using << /NeverEmbed [false /Helvetica-Bold] >> setdistillerparams .