[iText-questions] Encoding used by DefaultFontMapper
Hi iText'ers, I came across the following problem: printing Unicode with (Graphics2D)g.drawString() works fine on screen for all unicodes, but not using PdfGraphics2D class. I figured that this is due to the fact that the DefaultFontMapper used in PdfGraphics2D uses the font encoding CP1252. No means are provided to set the font encoding explicitly. I tried to write my own DefaultFontManager (which had the encoding for the basefonts set to IDENTITY_H or IDENTITY_V [what's the difference or reason for the two identities, anyhow]) but this leads (run the attached example) to the following exception, (encoding in DefaultFontMapper is set to BaseFont.IDENTITY_H) ExceptionConverter: com.lowagie.text.DocumentException: Identity-H at com.lowagie.text.pdf.Type1Font.init(Type1Font.java:252) at com.lowagie.text.pdf.BaseFont.createFont(BaseFont.java:360) at com.lowagie.text.pdf.BaseFont.createFont(BaseFont.java:313) at com.lowagie.text.pdf.DefaultFontMapper.awtToPdf(DefaultFontMapper.java:125) at com.lowagie.text.pdf.PdfGraphics2D.getCachedBaseFont(PdfGraphics2D.java:514) at com.lowagie.text.pdf.PdfGraphics2D.setFont(PdfGraphics2D.java:508) at com.lowagie.text.pdf.PdfGraphics2D.init(PdfGraphics2D.java:140) at com.lowagie.text.pdf.PdfContentByte.createGraphics(PdfContentByte.java:2168) at com.lowagie.text.pdf.PdfContentByte.createGraphics(PdfContentByte.java:2157) at DerivedScreen.init(DerivedScreen.java:50) at DerivedScreen.main(DerivedScreen.java:72) This stack trace raises several questions: Why is it trying to create a Type1Font? I'm only using arial.ttf, don't I? What would be the proposed way to give the user a chance for changing the encoding? The best (most convenient for the user) would be, if PdfGraphics2D.drawString() would switch encoding (and associated BaseFont) if a Unicode character outside the current encoding is encountered. Are there any means for that? Thanks for any help and cheers Erwin DerivedScreen.java Description: DerivedScreen.java
[iText-questions] Images in running text
Hi *, I'm trying to put little icons into running text, they sometimes are higher than the font, and they sometimes hang lower then the lowest character. Can somebody verify that such images, wrapped in chunks (new Chunk(img, 0,0) and added to one long lasting Paragraph, clash with the preceding and following lines of text? Maybe I have broken something, but I cannot find the place where the maximal height of a line is computed (and the 'leading of the current line' is adjusted). Put the other way, how can I ensure that such an 'Image in a Chunk' doesn't clash with other lines of text? Thanks for any hint, Cheers Erwin --- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf ___ iText-questions mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/itext-questions
RE: [iText-questions] Current CVS
Hi, i proposed this change when implementing the table-in-colspanned cells feature. This method is so 'internal' that no one outside the iText package should need to call it. It should be ensured that iText calls it when necessary internally. If the writers beeing in sub-packages cannot access it, then i missunderstood the term 'package protected' and it is indeed a bug then, however my pdfwriter accesses it correctly... and besides, i don't even know wether this particular change is due to my proposals :) Cheers Erwin -Original Message- From: Mark Hall [mailto:[EMAIL PROTECTED]] Sent: Saturday, June 22, 2002 9:18 AM To: Bruno Cc: itext-questions Subject: [iText-questions] Current CVS -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Hi I've got a problem with the current CVS version. The table object used to have a public void complete() method. In the current version this method is specified as void complete(). Unfortunately all Writers depend on this method being public, because they call it before processing Tables. I'm not sure whether there's been a design change here or if this is just a bug. Greetings, Mark - -- The cause of the problem is: Webmasters kidnapped by evil cult. My GPG key is available at: http://www.edu.uni-klu.ac.at/~mhall/files/MarkHall.asc -BEGIN PGP SIGNATURE- Version: GnuPG v1.0.7 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iQEVAwUBPRQkmkks8txBCICQAQIsWwgAjbsBki1LZ25MWEYPQmKNTVpH0uQdh5OB zedK2Ptu9xNFZSPc9HARN0xMDpNwm4Y+KEF225iToaOoaTf1XBQ6km8299zQrHl4 LzTrKf1TkslnhNvlYR3NaifxuI5qwiDI4EavLoBieQibjSBbrR8ETFXi1woJp1Kb DzNqdqvPtV4oNlAcmTYPfY6jcTeYAt/XWBGOuY35LMcz1Q+Aseqmr7ZVWCyihNvR pCwSo9sYWdoSFb5lDy6WWXzNspoqLs51G+qPoM4idLV2QqffR5EqIm20rXLrxU2r HhG7jYjhxDK5MSLw64+gzj26bBZT+KJOFPX7K7N1pJu99WU1tHSJDA== =X4Ag -END PGP SIGNATURE- --- Sponsored by: ThinkGeek at http://www.ThinkGeek.com/ ___ iText-questions mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/itext-questions --- Sponsored by: ThinkGeek at http://www.ThinkGeek.com/ ___ iText-questions mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/itext-questions
[iText-questions] DefaultFontMapper.insertDirectory is (too) slow
Hi Bruno, we are facing the problem, that inserDirectory processes 250 font files, consuming more time than acceptable (8 seconds). I was browsing to code and observed that for each font file a basefont is created. This includes loading mapping tables (which is what we need in the first place) but this also include construction of kerning tables, and numerous other datastructure: TrueTypeFont.process() ... fontName = getBaseFont(); fullName = getNames(4); //full name familyName = getNames(1); //family name fillTables(); readGlyphWidths(); readCMaps(); readKerning(); ... I would propose that these font information are only loaded when the font is actually needed in the document, and thus speeding up the insertDirectory-call. A lazy Font construction, if you wish. Is this problem recognized, planned to be solved or am I completely mistaken here? Cheers Erwin ___ Don't miss the 2002 Sprint PCS Application Developer's Conference August 25-28 in Las Vegas - http://devcon.sprintpcs.com/adp/index.cfm?source=osdntextlink ___ iText-questions mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/itext-questions
RE: [iText-questions] DefaultFontMapper.insertDirectory is (too) slow
The problem with this is that the file ist completly read in to generate a new
basefont.
Because the FontFactory.createFont(...) calls
new TrueTypeFont(filename, ...) and similar for other file types, only to get the
names of the font
String names[][] = bf.getFullFontName();
and then bf=null; again.
I would hope that a FontMapper.getFullFontName(filename) could be implemented without
really building a new basefont, but only by peeking into the file.
How difficult would that be?
Cheers
Erwin
-Original Message-
From: Paulo Soares [mailto:[EMAIL PROTECTED]]
Sent: Friday, June 14, 2002 12:34 PM
To: Erwin Achermann; iText ML (E-Mail)
Subject: RE: [iText-questions] DefaultFontMapper.insertDirectory is
(too) slow
If you just need a couple of fonts use this class (untested):
import com.lowagie.text.pdf.*;
public class SpeedFontMapper extends DefaultFontMapper {
/** Inserts a single font into the map. The encoding
* will be CODEBaseFont.CP1252/CODE but can be
* changed later.
* @param file the font file name
* @return CODEtrue/CODE if the font was inserted
*/
public boolean insertSingleFont(String file) {
String name = file.toLowerCase();
BaseFont bf = null;
try {
if (name.endsWith(.ttf) || name.endsWith(.afm)) {
bf = BaseFont.createFont(file, BaseFont.CP1252, false,
BaseFont.NOT_CACHED, null, null);
String names[][] = bf.getFullFontName();
bf = null;
insertNames(names, file);
}
else if (name.endsWith(.ttc)) {
String ttcs[] = BaseFont.enumerateTTCNames(file);
for (int j = 0; j ttcs.length; ++j) {
String nt = file + , + (j + 1);
bf = BaseFont.createFont(nt,
BaseFont.CP1252, false,
BaseFont.NOT_CACHED, null, null);
String names[][] = bf.getFullFontName();
bf = null;
insertNames(names, nt);
}
}
}
catch (Exception e) {
return false;
}
return true;
}
}
Best Regards,
Paulo Soares
-Original Message-
From: Erwin Achermann [SMTP:[EMAIL PROTECTED]]
Sent: Friday, June 14, 2002 11:05
To: iText ML (E-Mail)
Subject:RE: [iText-questions]
DefaultFontMapper.insertDirectory is
(too) slow
Paulo,
it was (is) my first attempt to remedy for the problem by
making one only
instance of the DefaultFontMapper in our application. And
every thread
uses this one. But still it takes too long for the
DefaultFontMapper to
read in our font directory. We cannot afford the time on
app startup, and
even less on the first 'toPdf'-action. All I'd expect from
the FontMapper
is to build a table mapping font-name and -attributes to
the appropriate
font-file, so that the fontfactory later will know which
file to load
(avoiding a scan of the font directory in each
FontFactory.getFont()-call). But the DefaultFontMapper does
way more than
that, apparently it instantiates each possible font... and
then throws
most of each instantiated font away again.
Cheers
Erwin
-Original Message-
From: Paulo Soares [mailto:[EMAIL PROTECTED]]
Sent: Friday, June 14, 2002 11:28 AM
To: Erwin Achermann; iText ML (E-Mail)
Subject: RE: [iText-questions]
DefaultFontMapper.insertDirectory is
(too) slow
There's no problem here. First of all the class
DefaultFontMapper is a
generic class that may not be suitable for all the cases. You
can sub-class
it or create a new one if special capabilities are needed.
You only have to create an instance of it once and then use
it everywhere in
other threads; I don't think the load time is that important
in this case.
Lazy evaluation would require scanning the directory and
reading all the
files to find the font, saving no time in the process.
Best Regards,
Paulo Soares
-Original Message-
From: Erwin Achermann [SMTP:[EMAIL PROTECTED]]
Sent: Friday, June 14, 2002 10:08
To: iText ML (E-Mail)
Subject:[iText-questions]
DefaultFontMapper.insertDirectory is (too)
slow
Hi Bruno,
we are facing the problem, that inserDirectory processes
250 font files,
consuming more time than acceptable (8 seconds). I was
browsing to code
and observed that for each font file a basefont is created.
This includes
loading mapping tables (which is what we need in the first
place) but this
also include construction of kerning tables, and numerous other
datastructure:
TrueTypeFont.process()
...
fontName = getBaseFont();
fullName = getNames(4); //full name
RE: [iText-questions] doc.close() closing OutputStream
The rule is actually very simple: if the user opened the stream and passed it to the library, then its the _users_ responsibility to close that stream, and analogously, if the library opened the stream i would expect it to close it as well. Any unbalance open-close-stream calls disturb my feelings ;) Cheers Erwin -Original Message- From: Bruno Lowagie [mailto:[EMAIL PROTECTED]] Sent: Tuesday, May 28, 2002 4:20 PM To: [EMAIL PROTECTED] Subject: RE: [iText-questions] doc.close() closing OutputStream Quoting Roman Novichenok [EMAIL PROTECTED]: Thought it would be useful to have the option of not closing the stream within iText library. In some cases, it would be useful. You are right about that. You were also right that changing the original method is not a good solution: it would break the functionality for all other iText users and applications. Your question to add a new method was justified, but since there is a simple way to solve the problem outside the library, adding a new method is no priority. For the moment I am working on other projects, so I leave iText development to Paulo and some other developers for a while. I am saving all kind of mails and remarks though, and I will take them into account once I decide to take the iText work up again (but I really need to do other projects now in order to learn about new tools, technologies, programming standards,...). Bruno ___ Don't miss the 2002 Sprint PCS Application Developer's Conference August 25-28 in Las Vegas -- http://devcon.sprintpcs.com/adp/index.cfm ___ iText-questions mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/itext-questions ___ Don't miss the 2002 Sprint PCS Application Developer's Conference August 25-28 in Las Vegas -- http://devcon.sprintpcs.com/adp/index.cfm ___ iText-questions mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/itext-questions
[iText-questions] FontMapper and FontFactory
Dear iTexters, I was using the Fontmapper in conjunction with PdfGraphics2D. In an initialization I registered all system fonts to the Fontmapper and henceforth could use all my locally installed fonts. Worked like a charm! DefaultFontMapper fm = new DefaultFontMapper(); fm.insertDirectory(sun.awt.font.NativeFontWrapper.getFontPath(true)); java.awt.Graphics2D g = pdfWriter.getDirectContent().createGraphics2D(width, height, fm); Nice job. But now I come to the situation where I want to translate an XML file into Pdf. The XML contains a tag like font family=Arial bold=true size=13 italic=false ... this should boil down to adding a chunk(,attributes) which loads the arialbf.ttf font. Is there a way to use font mapper for easing this, too? Thanks and Cheers Erwin ___ Don't miss the 2002 Sprint PCS Application Developer's Conference August 25-28 in Las Vegas -- http://devcon.sprintpcs.com/adp/index.cfm ___ iText-questions mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/itext-questions
[iText-questions] PdfGraphics2D: how to get the text width
Hi all,
I am using PdfGraphics2D in place of an old PrinterGraphics2D. In my code I have the
following snippet:
width = g2d.getFontMetrics().getStringBounds(txt, flushIndex, flushIndex + flushLen,
g2d).getWidth();
to compute the horizontal extension of some string. This seems to be the culprit in my
case. It seems as if this sequence used on PdfGraphics2D is not yielding the correct
width. Can you affirm this observation? Is there a more elegant way to find out the
exact width of some rendered text? As a workaround I extended my own PdfGraphics2D by
the method:
public double getWidthPoint(java.lang.String s) {
return (double)baseFont.getWidthPoint(s,fontSize);
}
and my code snipped reads, now:
// width = g2d.getFontMetrics().getStringBounds(txt, flushIndex, flushIndex +
flushLen, g2d).getWidth();
java.lang.String tmp = new java.lang.String(txt,flushIndex, flushLen);
PdfGraphics2D pg = (PdfGraphics2D)g2d;
width = pg.getWidthPoint(tmp);
In my cases this fixes the problem. But I'm not totally calm, because the idea of
PdfGraphics2D was to have be able to use it as any other Graphics2D Object given
this situation, that goal is not attained :( Also my PdfGraphics2D.getWidthPoint(s) is
not taking into consideration that the font itself may carry an AffineTransform
different from identity, which will of course affect the resulting width. Any better
solutions?
Thanks and Cheers
Erwin
___
iText-questions mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/itext-questions
RE: [iText-questions] PdfGraphics2D: unbalanced save/Restore Exception
Thanks Paulo, for pointing me gently enough to my own error. I fixed it and it works nicely now. ;) Cheers Erwin -Original Message- From: Paulo Soares [mailto:[EMAIL PROTECTED]] Subject: RE: [iText-questions] PdfGraphics2D: unbalanced save/Restore Exception Last g to be created is not used or disposed. Best Regards, Paulo Soares ___ iText-questions mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/itext-questions
[iText-questions] PdfGraphics2D: unbalanced save/Restore Exception
Hi iTexters,
during the last week I was doing exactly a PdfGraphics class as you seem to have done.
I was only implementing the features which we currently are using. Especially I was
cheating with the FontMetrics. Now I realized this morning that you have done a more
complete implementation in the PdfGraphics2D, reason enough for me to abandom my
implementation and switch to the far more complete one of iText. But the problem I'm
observing is show in the following code:
/**
* codecreatePdf/code prints a set of pages into the PDF
* file.
*/
public void createPdf() {
// make an instance of a PdfGraphics Object for the document
DefaultFontMapper fm = new DefaultFontMapper();
java.awt.Graphics2D g = null;
// don't do anything if pageable is null
if (pageable != null) {
// read all fonts recognized be iText in system font path
fm.insertDirectory(sun.awt.font.NativeFontWrapper.getFontPath(true));
System.out.println(PdfJob.createPdf: started);
int pages = pageable.getNumberOfPages();
System.out.println(PdfJob.createPdf: pages to print:
+pages+\n);
// set pageSize of first page before the document opens!!
float width =
(float)pageable.getPageFormat(0).getPaper().getWidth();
float height =
(float)pageable.getPageFormat(0).getPaper().getHeight();
theDocument.setPageSize(new Rectangle(width, height));
// g.setPageSize(pageable.getPageFormat(0));
theDocument.open();
g = pdfWriter.getDirectContent().createGraphics(width, height,
fm);
for (int p = 0; p pages ; ++p) {
Printable thePage = pageable.getPrintable(p);
try {
// print!
thePage.print(g, pageable.getPageFormat(p), p);
System.out.println(PdfJob.createPdf: page
+ p + printed);
if (p pages-1) {
// next page, please! on its proper
PageSize
width =
(float)pageable.getPageFormat(p+1).getPaper().getWidth();
height =
(float)pageable.getPageFormat(p+1).getPaper().getHeight();
theDocument.setPageSize(new
Rectangle(width, height));
} // end of if (p pages-1)
theDocument.newPage();
// iText-Error!?
// g.dispose(); // use of dispose leads to a
// unbalanced save/restore exception not
using
// g.dispose() is producing erroneous PDF:
// Acrobat Reader complains about unmatched
// save/restore commands if the document
consist
// of Portrait and Landscape format. Acrobat
Reader
// show one empty additional page if the
document
// is only one page format
/
g =
pdfWriter.getDirectContent().createGraphics(width, height, fm);
} catch ( PrinterException e) {
e.printStackTrace();
}
catch(DocumentException de) {
de.printStackTrace();
}
} // end of for (int p = 0; p pages; p++)
// write the document and clear all references
end(); // = theDocument.close()
} // end of if (pageable)
} // end createPdf
___
iText-questions mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/itext-questions
