fig2vect 1.1.23

Change log

Change log

1.1.23
- Bugfix: Due to an #ifndef instead of an #ifdef the automatic checking of the LANG environment variable for ``.UTF-8'' was not disabled as intended. This is fixed now.
1.1.22
- Java output driver creates new functions:
  - setTextEnabled(String s, boolean f)
    to enable/disable one text.
  - setTextInLayerEnabled(int l, boolean f)
    to enable/disable all texts in a layer.
  - enableAllTexts()
    to enable all texts.
  - Graphics2D getTransformedGraphics(Graphics g, PageFormat pageFormat)
    to obtain a graphics context with the same transformations applied as for printing.
  - Rectangle2D.Double getLayerRectangle(int l)
    to retrieve the used rectangle in a specific layer.
- Bugfix in the Java output driver: Some images names (those without slashes/backslashes) were not printed correctly. This is fixed now.
1.1.21
- Changed default setting: On systems with LANG=xx_YY.UTF-8 jFig and XFifg 3.2.5 do not write UTF-8 encoded texts. So the default setting was changed from ``utf-8=auto'' (inspect LANG environment variable) to ``utf-8=no'' (not UTF-8 encoded texts in Fig file).
- Bugfix in Java output driver: Spaces were used to force numbers to a given string length. This is ok for normal number but not for object names consisting of a character and a number as the space divides the object name into an identifier and a number.
  Now the use leading zeores.
1.1.20 (new features)
- Java output driver,see FAQ: Java driver for details.
- Option "page size = w h [ llx lly urx ury ]" to specify the page size and optionally the printable range in PS points.
  For the PS output driver only.
- Option "page alignment = alignment" to choose the position of the graphics on the page. For the PS output driver only.
- Option "image alignment = alignment" to embedded images into the available space.
1.1.19 (Documentation)
- Developer documentation was added using doxygen.
1.1.18 (internal change)
- The dksf_must_rebuild() function is used to check whether a destination file is already up to date.
1.1.17 (new feature)
- We ensure not to use exponent notation when printing double values.
1.1.16 (bugfix, new features, new configuration mp.tex in fig2vect.cfg section 1)
- Bugfix: Due to a wrong index calculation the first column of embedded PNG and JPEG images was damaged when writing PDF output.
  This is fixed now.
- A configuration setting
```
color = boolean
```
  was added to allow to generatoe grayscaled or colored output.
  This option has no effect on included images, they are passed through as they are.
- The MetaPost output driver now interprets the
```
skip all texts = yes
```
  option. If this option is set no texts are issued at all.
  This allows to combine the MetaPost driver and TeX driver. In this combination the MetaPost driver produces all graphics elements except the texts which are produced by the TeX driver.
- The
```
fill bounding box = yes
```
  option was added for the MetaPost output driver.
  Unlike EPS or PDF there is no BoundingBox information in the *.mp file but the correct bounding box information and correct shifting is needed when combining MetaPost and TeX output driver. So this option fills the bounding box calcutlated by fig2vect using white color in the *.mp file, this results in a correct bounding box in the *.0 or *.mps file.
- A new configuration named ``mp.tex'' was added to section 1 of fig2vect.cfg. When uprading from a previous version make sure to transfer the entire contents of fig2vect.cfg section 1 from the source archive to your per-machine configuration file (typically /etc/fig2vect/fig2vect.cfg or /usr/local/etc/fig2vect/fig2vect.cfg).
1.1.15 (bugfix)
The configuration file fig2vect.cfg contained a typo "verbose outputt" in the SVG configuration in previous verions.
This is fixed now.
1.1.14 (new feature)
- The PDF driver now adds page attributes if the Fig file references a PNG image containing an alpha channel.
  The
```
	allow pdf page attributes =
	
```
  setting can be used to enable/disable the creation of pdf page attributes.
1.1.13 (new feature in the Windows version)
When switching standard output to binary mode before doing output the mode is set back to the original text/binary mode before the program exits.
1.1.12 (new feature)
- When showing the current configuration the strings are now passed to the new dkapp_stdout() function using codepages if necessary.
  This creates better output for non-english languages when used on Windows systems.
1.1.11 (new feature)
- Two new long options were added.
  Run
```
fig2vect --configuration-file
```
  to get the current configuration file listed.
- Run
```
fig2vect --write-configuration-file
```
  and provide a configuration files contents to be saved.
  The purpose of these two options is to allow other programmers to wrap a GUI around fig2vect allowing to choose a configuration or to manage configurations.
- The command line option -A was added to suppress messages about alpha channels transferred to PDF.
  These messages are of interest for new users of fig2vect but may be boring after a while.
1.1.10 (new feature)
- If PDF images containing an alpha channel are included in documents processed by pdfLaTeX, the document preamble should contain a line
```
\ifpdf\pdfpageattr{/Group <</S /Transparency  /I true /CS DeviceRGB>>}\fi
```
  to prevent PDF viewers from using CMYK as intermediate colorspace as this would eventually result in color shifting.
  The LaTeX output driver now checks the PNG images referenced in the *.fig file for alpha channels. If at least one alpha channel is found, the line above is included automatically into the *.tex output file just before the
```
\begin{document}
```
  line (when producing a full LaTeX source file).
  The PDF output driver also does the check and issues a warning to remember you to include this text line into your LaTeX source code. The TeX output driver also prints this warning when not producing a full LaTeX source file.
1.1.9 (bugfix)
- In the LaTeX output driver non-special text is UTF-8 decoded and re-encoded for LaTeX if necessary. In previous versions characters were skipped if no LaTeX encoding was found in the encoding tables.
  This is fixed now: The driver re-encodes such a character to UTF-8 and writes it to output. We hope, the inputenc package will handle it correctly.
1.1.8 (new feature)
- An option ``-a'' was added to generate a file name for the output file automatically if no output file name was specified.
  The name is based on the input file name and the output driver.
1.1.7 (1 new Feature and 2 bugfixes)
- Fig2vect can now process Fig files containing UTF-8 encoded text.
  See the FAQ page for details.
- Font encoding for PDF output.
  Previous versions of fig2vect did not specified an encoding when using fonts in PDF output. The default encoding does not allow to use german umlauts...
  Version 1.1.7 explicitly specifies /WinAnsiEncoding.
- Bugfix in SVG output, now writing correct font names.
  Previous versions of fig2vect contained a bug, when using a font configuration file, a suffix ``.svg'' was added to the file name (this was necessary for file names in the built-in list, but not for file names from configuration files). So instead of ``filename.svg'' the output was ``filename.svg.svg''.
  This is fixed now.
1.1.6 (Bugfix)
- The mmp output driver sometimes did not always start a new figure using ``\beginfig(...)'' when detecting a gap in the layers of graphics elements.
  This is fixed now.
1.1.5 (Bugfix)
- Texts using the default color (-1, black) were converted to texts using random colors in SVG output. The reason was an incomplete initialization of the color cell structure.
  This is fixed now.
1.1.4
The new configuration entry
font configuration file = filename
allows to use a font configuration file for flexible mapping of PostScript fonts to
- either system fonts
- or SVG variants of the GhostScript fonts.
Example font configuration files are shipped with fig2vect:
- webfont.cfg
  shows how to reference downloadable SVG variants of the GhostScript fonts.
- winfont.cfg
  shows how to reference system fonts.
- combfont.cfg
  shows how to reference both system fonts and SVG variants of the GhostScript fonts.
All three files need customizing (correct locations, make sure to reference only fonts which are installed on your system...)
Version 1.1.3
In previous versions I didn't use the term EPS and PS correctly. Now we have a PS/EPS driver which can produce either EPS or PS files. The showpage and setpagedevice operator can be used only when creating PS files.
Version 1.1.2
- Bugfix:
  If the setpagedevice operator is used, we can print %!PS-Adobe-x.0 only, not %!PS-Adobe-x.0 EPSF-x.0
Version 1.1.1
- Bugfix:
  In previous versions, if only one end of an arc had an arrowhead attached the wrong end of the arc was cut.
  This is fixed now.
Version 1.1.0
- Configuration changes:
  - Default settings are changed to produce best quality instead of small files.
  - Configuration entry changed, instead of the former
```
large pattern cell = no
```
    use
```
tiled patterns = yes
```
    now (for smaller files on the cost of less quality).
  - Configuration entry
```
ps setpagedevice = yes
```
    or
```
ps setpagedevice = no
```
    was added. If this entry is turned on, the page size (paper size) is set to match the bounding box. This entry should be turned on for images to be processed by GhostScript. It should be turned off for images to be included into LaTeX documents. A configuration "[eps.standalone]" was added, to produce standalone images.
    [Updated 2007/05/13]: Instead of writing setpagedevice into the ps file it is better to provide the -dEPSCrop argument to GhostScript.
- Bugfix: There were compilation errors when compiling fig2vect on systems where either libbz2 or zlib was missing.
  Some messages used "c->msg" instead of "c->msg1". This is fixed now.
Version 1.0.3
- Bugfix: The PDF output driver aborted conversion silently on problems with PNG files. This is fixed now, on problems with PNG files the driver writes an error message before the conversion is aborted.
- Bugfix: The scripts/comment.txt file was changed, all occurances of "--" were replaced by " - " (except in the comment delimiters).
  Although the file was syntactically correct, Squiggle on Windows failed to use the fonts containing the comments from the previous version of the file.
Version 1.0.2
- Bugfix: There was a possible buffer overflow in the PDF output driver. This is fixed now.
Version 1.0.1
- Bugfix: The document level control comments were not processed by all drivers. This is fixed now.
- Bugfix: When processing a directory we now save the options before processing a file and restore the options after processing.
Version 1.0.0
- PDF driver finished. All intended features for the program are accomplished now, so we can switch the version numbering to 1.x.x.
- Bugfix in the MetaPost driver: For rotated colored text the driver wrote the rotation instead of the color. This is fixed now.
- Bugfix in the SVG driver: For rotated text the x-transformation function was used for both x and y co-ordinates. This is fixed now.
- New feature in the TeX driver: The driver can now create both code snippets for inclusion into LaTeX sources and full LaTeX sources to create standalone PDF-files.
- The number precision was unified between the drivers, the decimal digits after the decimal dot can be controled via the configuration file now.
- The PDF documentation was rewritten. The configuration entries are now sorted by driver and purpose.
Version 0.9.15 (bugfix and update)
- Short summary:
  - License change: This project uses a BSD-style license now.
  - Bugfix: System configuration directory was not defined.
  - New feature: TeX output driver.
- Details:
  - The sources contain a fallback definition
```
#ifndef SYSCONFDIR
#define SYSCONFDIR "/etc"
#endif
```
    Normally the system configuration directory is found automatically when the configure script is run. A definition is written to the Makefile. Programs searching for files in this directory need a compiler definition -DSYSCONFDIR=\"$(sysconfdir)\" in the Makefile which was not present. So "make install" installed files to /usr/local/etc but the program searched for these files in /etc. This is fixed now.
  - A TeX output driver was added. Using the configuration file entries
```
[eps.tex]
# to use exact fonts write
# normal text = handling:tex,font:fig,size:fig,mbox
normal text = handling:tex,font:similar,size:fig,mbox
skip all texts = yes
[tex]
normal text = handling:tex,font:similar,size:fig,mbox
```
    on can run
```
fig2vect -leps.tex f1.fig f1.eps
fig2vect -ltex f1.fig f1.tex
```
    and include the graphics using
```
\begin{figure}
{\centering
\input{f1.tex}
\caption{My image}
\label{fig:my-image}
}
\end{figure}
```
    The document LaTeX source imports the TeX source for the figure which uses includegraphics to import the EPS image and writes labels on top of the EPS image.
    The idea is the same as in transfig's pstex/pstex_t driver combination but fig2vect allows you to decide whether normal text is placed in the EPS file
```
[eps.tex]
# to use exact fonts write
# normal text = handling:tex,font:fig,size:fig,mbox
normal text = handling:none
skip all texts = yes
[tex]
normal text = handling:none
```
    or in the TeX file.
```
[eps.tex]
# to use exact fonts write
# normal text = handling:tex,font:fig,size:fig,mbox
normal text = handling:tex,font:similar,size:fig,mbox
skip all texts = yes
[tex]
normal text = handling:tex,font:similar,size:fig,mbox
```
    If handling is left up to the TeX file one can use the exact fonts as specified by the Fig file (font:fig) or use automatic font replacement (font:similar).
Version 0.9.14 (update)
- Documentation, string tables and other platform-independent read-only files are installed in ${prefix}/share now instead of ${prefix}/lib.
Version 0.9.13 (bugfix)
- Short summary:
  Improved handling of corrupt/invalid PNG-files.
  Previous versions of fig2vect crashed and dumped core when processing invalid/corrupt PNG files. This is fixed now.
  Previous versions did not list the used image libraries in the version text. Sorry for that, this is fixed now.
- Details:
  The PNG reference library uses the setjmp()/longjmp() functions if there are problems while reading a PNG file. In previous versions of fig2vect there was no setjmp(); instructions to set up the jump buffer.
Version 0.9.12
- Bugfix in the style comparison function in dkfigsvg.ctr.
- Configure script modified to print an error message and some help if the required dklibs library set is not found.
Version 0.9.11
- Documentation for backup rectangles.
Version 0.9.10
- Code to support backup rectangles.
Version 0.9.9
- SVG driver completed.
Version 0.9.8
- EPS driver completed.
Version 0.9.7
Bugfix release. Texts in MetaPost output were shifted as MetaPost automatically adds an offset to labels and margins to bounding boxes. In Fig files text is positioned by the user, we set
```
labeloffset:=0;bbmargin:=0;
```
to avoid additional offsets/margins.
Version 0.9.6 (no public release)
Starting development of EPS output driver.
Version 0.9.5
Makefile "makefile.msc" added for Visual C++.