CentraDoc 6.3.n - Summary of Changes since 6.3

Introduction

This document contains a short description of changes on the 6.3 branch since the initial 6.3.1 release, grouped by version and ticket number.

6.3.3 tagged in [8545]

#722 API for replacing PostScript operators at run time:

void psgetoperator(PS_CONFIG *psc, const char *name, PS_OBJECT *o);
void pssetoperator(PS_CONFIG *psc, const char *name, OpProc proc);
void psinvoke(PS_CONFIG * psc, PS_OBJECT * o);

#720 PostScript fix issues conspiring against CID 0.

#719 Preview application didn't run under Vista, plus other Preview build fixes.

#717 Windows writing stdout and stderr from GUI app would produce occasional spurious errors when these files aren't open. Windows FFsStd now detects this case instead of relying on configuration at build time.

#715 Duplicate of RIP clipping fix from 6.3.1 - lib-sys.com website had wrong 6.3.2 RIP code.

#714 PostScript fixed bomb if image file object is the same as the base file.

#712 PostScript enhancements:

#711 Update open source libraries changes.txt files with attribution.

#710 PDF Read not providing font scale matrix to PDF Output driver.

#708 PostScript flushfile on read means seek to end of file. This is a little slippery to have as part of the FFS API, so be careful.

#707 PDF Handle case of embedded TrueType fonts without cmap tables.

#705 PDF don't use external fonts for GID_CHARSET. For now, the driver is the only place that can handle GID_CHARSET.

#701 PostScript multichar == 0 and driver failing DvText on-the-fly restarted text emulation of the string at the beginning.

#697 JasPer tweak for potential temp file deletion weirdness.

#690 PDF file contains "Symbol" instead of "SymbolEncoding" encoding.

#686 PDF Fixes for CJK-V text placement when emulated, improves #241.

#684 Postscript PreviewSDK EnablePSProcessGray=Y to force following DeviceGray ProcessColorModel.

#678 RIP switch to nearest neighbor if image width or height below arbitrary threshold. Fixes some ugly results from other image filters.

#677 More shader fixes related to #651.

#675 PDF annotation write fixes - The PDF/A spec is wrong about flags.

#673 Fixes for 64 bit build compatibility.

#672 PDF Acrobat accepts XRef streams in any version of PDF file.

#669 PDF Fixes for transparency group clipping.

#658 PDF TrueType fonts look at /Encoding first when picking 'cmap'.

#654 PDF attempts to block font stroking if stroke and file with thin lines and the same color. Work around for mixing ugly stroke against good fill.

#653 PDF Write Type 1 font creation support.

#652 RIP draft mode fixes.

#651 Many changes to enhance DvShade interface for driver native shading.

#646 PDF access External CJK TrueType fonts in RIP.

#645 PostScript several subtle / special case fixes:

* enable run operator for file written to temp directory
* fix executive reentrance problems with proc files nesting
* Problems with SubFile decode positioning on raw input file

#643 PDF allow file with version 0.0.

#642 PDF japanese characters should emulate, passing the wrong information to driver. Looks like a bug in #240 with the wrong fix in [6489].

#537 PDF handle UserUnit scaling on input.

Unticketed:

6.3.2 tagged in [8195]

#624 Preps based PostScript file using clippath to make decisions about placement and/or save restore the clippath (instead of clipsave/cliprestore). Quick fix handles this correctly if the clips are rectangles. clippath is still incorrect if the clips aren't rectangles.

#626 Build issues:

#629 riptest.c no longer hard-codes the temporary file name (temp.$$$) and also deletes it correctly on completion.

#632 New cap option gxc->no_textadjust, turns off gxc->multichar if DvText has adjust values, placing characters one at a time, so the driver doesn't have to.

#634 Color Spaces: PdfBlockICC=Y ini option to ignore embedded ICC Profiles in PDF files. This setting also maps CalGray to DeviceGray and CalRGB to DeviceRGB. An alternative Lab implementation that runs without ICC is also available; contact Liberty Systems for more information.

#637 PDF pattern color space with no explict fill color recursed unpleasantly, blowing the stack.

#639 Acrobat allows some ridiculous numeric input (e.g. ~= 1.7e308), as long as all the characters are numeric, the range isn't checked.

6.3.1 tagged in [8057]

References