0.2 (95-dec-12) --------------- First public release. 0.3 (96-jan-13) --------------- LZW patent workaround. Implemented inline images. Fixed (mostly) disjoint polygon fills. Added remote server stuff. Added page number on command line. Fixed problem with font encodings which caused character misalignment. Fixed inverted CCITT decoding and inverted image mask drawing. Now compiles under gcc 2.7.x (ignore those stupid 'unused parameter' warnings). Many minor bug fixes and optimizations. 0.4 (96-apr-24) --------------- Implemented DCT filter. Implemented PostScript output; wrote pdftops program. Implemented links. Implemented font rotation -- I was wrong: X11R6 servers *do* support font rotation (by specifying a matrix in place of a size). Added bindings for Home/End, Page Up/Down, arrow keys. Added initialZoom resource and -z option. Added geometry resource and -g option. Fixed image size off-by-one bug. Fixed bug where page content is reference to an array of streams. Cleaned up uninitialized variables which were causing random problems on various platforms. Manually skip whitespace before calling atoi() for startxref. Replaced calls to XrmCombineFileDatabase() with calls to XrmGetFileDatabase() and XrmMergeDatabases() so it will work under older versions of X. Fixed problem with reading multiple xref tables in updated PDF files. Check for encryption and print appropriate error message. Rudimentary dithering of images. Fixed bug in CCITTFax filter (pass mode followed by horizontal mode). Optimized drawImage() and drawImageMask(). Changed several things to ease porting: - changed '__inline' to 'inline' (to adhere to the ANSI standard) - surrounded interface/implementation pragmas with #ifdef _GNUC__ - got rid of empty array initializer lists - moved Operator type definition from Gfx.cc to Gfx.h - renamed String, uint, etc. - ability to uncompress to file (NO_POPEN and USE_GZIP flags) - added definitions of XK_Page_Up/Down and XPointer for old versions of X For VMS port: - use correct Xdefaults name for VMS, get rid of ltkGetHomeDir() - added '#include ' before all X includes - renamed files with multiple periods in their names Fixed window resizing infinite oscillation bug. Fixed problem with string-type (as opposed to stream-type) indexed color space lookup tables (which are used in inline images). If an X font is not found, try smaller and then larger sizes (this is useful for old or broken X servers which can't scale bitmap fonts). Added -rgb (color cube size) option. Various minor bug fixes. 0.5 (96-may-23) --------------- Fixed bug in LTKWindow which broke the remote server mode. Fixed PostScript output: - doesn't seg fault if file is unwritable. - incorrect DSC comment - need colon in '%%Page:'. - use 'imagemask' command for masks. - output filters in the correct order. - Stream::isBinary() checks the next-to-bottom, not top, stream. - if page width > height, rotate it 90 degrees. - if page is larger than paper size, scale it down. Set default MediaBox to 8.5" x 11" to deal with non-compliant PDF files which don't specify a MediaBox. Added DEBUG_MEM stuff in gmem.c and gmempp.cc. Fixed memory leaks: - LTKWindow didn't delete the LTKBox. - LinkAction needs a virtual destructor. Use $(RANLIB) variable in goo/Makefile and ltk/Makefile. Allocate image data after calling XCreateImage, using image->bytes_per_line -- works in 24-bit mode now. DCTStream rounds width of rowBuf lines up to the next multiple of mcuWidth, so last MCU doesn't run off end of buffer. Increase size of block (from 255 to 1024 bytes) read at end of file to search for 'startxref'. Skip past garbage at start of file, look for '%PDF'. Moved more compiler options out of Makefiles into Makefile.config. Top-level Makefile uses '$(MAKE)' instead of 'make' for making subdirectories. Space/PageDown/Next and Backspace/PageUp/Previous now moves to next/previous page if already scrolled to bottom/top of current page. 0.5a (96-jul-09) ---------------- [not a public release] For PDF 1.2 (a.k.a. Amber, a.k.a. Acrobat 3) support: - look for trailer after first xref instead of at end of file. Deal with font subsets by converting character names of the form 'Cnnnn' to the appropriate character from the standard encoding. Extract encoding from embedded Type 1 fonts. Kludge to fill one-pixel thick polygons. Changed X font encoding to use endash for hyphen (gets rid of too-long hyphens). Handle Resources key in Pages dictionaries (needed for pstoedit output). Fix comment handling in Parser (needed for pstoedit output). Move Bezier curve conversion from GfxState to XOutputDev; look at flatness parameter in GfxState. Change all of the path functions in XOutputDev (stroke, fill, clip) to use the same path transformation/conversion function. Rewrote PostScript output driver as a subclass of OutputDev; removed duplicated code (ps_ functions) from Gfx. Fixed bug in xref code with small (< 1024 bytes) PDF files. Implemented BX/EX operators. Added PDFDoc class. 0.6 (96-nov-12) --------------- Add support for PostScript output to stdout (-) and to a command (|lpr); added -ps option and psFile resource. Decryption is implemented but not included in the distribution due to legal restrictions: the decryption algorithm is a trade secret of RSA, Inc., and the U.S.A. still has bogus export controls on cryptography software. Added .xpdfrc config file: - Added fontmap parameter: user can map PDF font names to X fonts. - Added fontpath parameter: search for Type 1 font if encoding is not in PDF file. Incremental display: display is updated after every 200 commands. Added forward and backward by-10-page buttons. Links: - Implement links with "Launch" actions that point to PDF files. - Draw borders around links. - Handle links with named destinations. - GoToR links specify a page number instead of a page reference. Optimizations: - Rewrote Stream to use buffering, and added lookChar() functions; rewrote Lexer to take advantage of this. - Use hash tables for name->code mapping in font encodings. - Made XOutputDev::doCurve() iterative, changed /2 to *0.5, and changed the flatness test. Added file name to window title. Implemented RunLength filter. Implemented forms. Convert ObjType to an enum. Changed isDict("Pages") to isDict() (in Catalog.cc) to deal with incorrect PDF files. Changed color selection so that very pale colors don't map to white. Fixed bug in CCITTFax filter (multiple make-up codes). In GString::clear(): need to set length to 0 before calling resize(). Base initial window size on first displayed page, not page 1; deal correctly with rotated pages. Added ltkGetIntResource() and LTKApp::getIntResource(). PostScript output fixes: - Escape backslashes in strings. - When doing ASCII85 encoding, keep both chars of EOF marker ('~>') on same line. - Add extra line '%-EOD-' after image data streams; call wrapper functions for image and imagemask which look for this line -- this should fix the 'too much data in stream' bug. - Font tags can be reused for different fonts on different pages -- so use font object reference (number/generation) instead. Initialize character widths to zero (this caused crashes on OSF/1). Handle image masks which go outside of pixmap. Makefile.config changes: - Remove -o in C++ compile rule. - Add $(AR) variable. Code which read char widths from font dictionary read all but the last width. Add 'return 0;' to main() in xpdf and pdftops. Allow fonts to use StandardEncoding. Convert man pages to VMS help files. 0.7 (97-may-28) --------------- Implemented FlateDecode filter (for PDF 1.2). Basic xref table reconstruction for damaged files New pdftotext program converts PDF to plain text. Implemented menus in LTK; added a menu to xpdf. Added open and save functions; allow xpdf to start without any PDF file. Implemented text find. Implemented text select/copy. Change mouse cursor when it's over a link. Embed Type 1 fonts in PostScript output. Moved rotate functions to menu; added quit to menu. Fixed stroke color bug in PostScript output (was using fill color instead of stroke color; this sometimes caused lines to be missing (white) in PostScript output). Support Launch-type links -- pops up a dialog before executing anything. Expects the A (action) dictionary to contain a Unix dictionary with F (file) and P (paremeter) keys just like the Win dictionary. A moveto op all by itself should just be discarded, instead of generating a subpath with one point (this was causing seg faults). Balanced parentheses in strings don't need to be escaped. Tj operator in PostScript prolog didn't check for zero when dividing by length of string. Implemented selection in LTK; TextIn widgets support dragging/copy/ paste. Handle font subsets that use hex character codes. Added icon pixmap; added the XPMLIB and NO_XPM variables to Makefile.config. Fixed subtle bug in use of horizontal scaling parameter (it affects only the width of drawn characters, not positioning done in text space). Memory testing (with DEBUG_MEM): - gmalloc now fills memory blocks with garbage to catch unitialized fields. - gfree fills memory blocks with garbage to catch uses of freed blocks. Handle image masks which go off the pixmap on the top and/or left. Fixed inline functions which had two return statements (to make the HP, SCO, and other cfront-based compilers happy). Fixed bug which caused seg faults when following a link to a different file (info in LinkGoto object was used after link was deleted by loadFile). If page content is an array of streams, the streams are concatenated; objects and commands can span multiple streams. If file open fails, try lower-casing and upper-casing the file name. Commands should end when lexer sees a '/' character. GString::append(char *, int) was broken. Changed LTKScrollingCanvas redraw to be more efficient: copy as much as possible from window before copying from off-screen pixmap. Ignore gs (set extended graphics state) operator. Handle colorspaces (CalGray/RGB are treated as DeviceGray/RGB; the weird colorspaces are not yet implemented). Named destinations (for links) can be strings as well as names; deal with the names tree in the catalog. Clip to the page CropBox. Added '-q' to gzip options (to suppress warnings, in case user has -v in GZIP env var). Added 'include Makefile.config' to top-level Makefile. Added INSTALL variable to Makefile.config; used in top-level Makefile. Always initialize LinkDest left/bottom/top/right/zoom fields (bogus floating point values were causing crashes on Alpha). Added Makefile.config options for Digital Unix (DEC compilers), HP-UX (HP compilers), SCO Unix, and Evans & Sutherland ES/OS. Added flag to set stream mode in fopen call for VMS. Rewrote Link module. Pages with no contents shouldn't cause an error message. In PostScript output: pdfImM needs to set fill color before doing imagemask. If font doesn't specify character widths, use widths from built-in font, based on font flags. Fixed LTK scrollbar to delay before repeating and to control the period between repeats. Removed window/widget copy() methods (they were untested and unused). Unknown filter types produce a single error message instead of a stream of errors. Added a dummy target in top-level Makefile so making individual executables (e.g., 'make pdftops') should now work. Added optional xpdf-flip.ltk with buttons on right side instead of bottom of window. 0.7a (98-feb-22) ---------------- Moved find command from menu to toolbar button ('f' key still works). Support TrueColor visuals. Added a -cmap option and a installCmap resource to install a private colormap. Mouse button 2 pans the window. Selecting a URI link now executes a configurable command (from the urlCommand resource). Added a "link info" display which shows the URL or file for the link under the mouse. Don't draw (or convert to PostScript) text drawn in render modes 3 and 7 -- this is invisible text, used by Acrobat Capture; this text is still passed to the TextPage object so that selection works. Recognize (and quietly ignore) marked content operators (BMC, BDC, EMC, MP, DP). Recognize new color-setting operators (scn, SCN). Added A4_PAPER option. Embed external Type 1 font files (this currently only works with PFA files). Added "-level1" option (in xpdf and pdftops) to generate Level 1 PostScript. Setup autoconf -- replaced Makefile.config. Added SELECT_TAKES_INT flag, and use configure to autodetect (for HP-UX). Fixed appendToPath() to behave reasonably when appending ".." to root directory. Fixed array size in FlateStream::compHuffmanCodes() (was causing xpdf to crash under OSF/1). ASCII85Stream, ASCIIHexStream, and DCTStream didn't check for EOF and could run past the end of the stream in damaged files. Handle hex escapes (#xx) in names. Still allow the name /# for backward-compatibility. Check for NULL characters in encoding array in GfxFont.cc (was calling strcmp() with NULL which crashed under Solaris). PageAttrs::PageAttrs() didn't initialize crop box boundaries. Changed uses of lookup() to lookupNF() in XRef.cc. Fixed type checking of operators which take a variable number of args. Gfx::buildImageStream() doesn't need to check for parser (since I got rid of the bogus array-of-command thing). XOutputFont matches on font reference instead of font tag (similar to PSOutputDev fix). Fixed bug in position calculation for multi-char substitutions in XOutputDev. Cleaned up local variables which hid class variables. Optimized variable length decoding in CCITTFaxStream. Set link border width to zero if Border dictionary entry is missing. Throw away zero-length strings in TextOutputDev -- they don't have valid xMin/xMax values. Swapped order of XLIBS and XPMLIB in xpdf/Makefile. Deleted 'LTKApp::' in function declaration in LTKApp.h. Changed '(XKeyEvent *)&event' to '&event.xkey' in LTKApp.cc. Check that the link rectangle coordinates are in the correct order, and swap if necessary. TextOutputDev didn't set text to NULL, which caused pdftotext to segfault if it couldn't open it's output file. Fixed a hash table search bug in GfxFontEncoding::getCharCode(). Cleaned up colorspace code: rewrote GfxColorSpace and added GfxImageColorMap; cleaned up PSOutputDev::doImage. Handle named colorspaces in images. Correctly set the default color after a colorspace change. Old setcolor operators now set the colorspace. Fixed bug with uncompressed blocks in FlateStream. Fixed bug with fixed Huffman code table in FlateStream. Added hash table of X windows (for LTKWindow and LTKWidget) to LTKApp and replaced calls to XQueryTree with hash table searches -- this avoids a roundtrip to the server for each event and also fixes the problem where XQueryTree crashed if the window no longer existed (with leftover events from a destroyed window). (Thanks to Yair Lenga for the suggestion.) Create a new GC for selection -- xor black and white (instead of LTK foreground and background). Fixed crash with blank lines in .xpdfrc. Allow spaces in font descriptors in fontmap lines in .xpdfrc. Check for bogus object number in XRef::fetch(). Use MacRomanEncoding for TrueType fonts that don't specify an encoding. Certain PDF generators apparently don't include FontDescriptors for Arial, TimesNewRoman, and CourierNew -- set GfxFont flags appropriately. Fixed a bug in width guessing in GfxFont -- sans serif and serif were swapped. Rewrote XRef::readXRef() to avoid using a parser to read the xref entries. Added NO_TEXT_SELECT option. Ignore APPn/COM/etc. markers in DCT streams. Replaced select() with XMultiplexInput() in LTKApp.cc for VMS. Handle WM_DELETE_WINDOW protocol -- if you ask the window manager to delete the xpdf window, xpdf will exit cleanly; other windows/dialogs are simply closed. Optimized DCT decoder; switched to integer arithmetic. The "/Type /Annots" field in an annotation dictionary is optional. Check for null nameTree in Catalog::findDest(). In XOutputDev, search user font map before default font map. Added "normal" SETWIDTH parameter to all font descriptors in XOutputDev (some systems have a narrow-width Helvetica font). Added FOPEN_READ_BIN and FOPEN_WRITE_BIN to support Win32. Added a hack which allows better font substitution for some Type 3 fonts. Also allow character names of the form /nn and /nnn. Added and to LTKApp.cc (needed by AIX and IRIX for bzero() declaration for FD_ZERO). 0.80 (98-nov-27) ---------------- Support for some Japanese fonts (Type 0 fonts using the Adobe-Japan1-2 character collection, horizontal only). Added pdfinfo application. Added pdftopbm application. Added pdfimages application. Added -papercolor option and .paperColor resource. Fixed divide-by-zero problem in XOutputDev Type 3 font matrix kludge. Font subset char names can be 'Bxx' as well as 'Cxx' and 'Gxx'. Fixed bug in color space conversion in DCTStream filter (YCC->RGB was correct, YCCK->CMYK was broken). Added XRef::getDocInfo() and PDFDoc::getDocInfo() to support pdfinfo. Optimized GfxImageColorMap. Lexer::getStream(), getPos(), and setPos() check for null stream. Decryption code now does strings as well as streams. ASCII85 decoder rounds short tuples up instead of down. CropBox and MediaBox can be non-integers. PostScript output: - Use a rectangle operator. - Call setpagedevice with page size. - Insert %%PageOrientation comments. - Add paper size flags (-paperw and -paperh) to xpdf and pdftops. - If HAVE_POPEN is not defined, and user tries to print to '|...', the PSOutputDev destructor tried to write to the PS file. - Added support for forms (pdftops -form). Removed error messages for empty paths in stroke, fill, etc. operators. Don't allow flatnesses less than 1 in XOutputDev (this speeds up rendering a little bit when there are lots of tiny curves). Moved the font subset character name guessing from GfxFont to XOutputDev and TextOutputDev - now these files print correctly. Cast argument to XFree() to XPointer; add XPointer definition where necessary (portability fixes). Various minor VMS fixes. Changes to configure script and Makefiles: - Print a warning if X is missing. - Use C++ when checking select() argument type (HP-UX). - Use 0 instead of NULL when checking select(). - Default to gcc instead of c++. - Get rid of AC_C_INLINE -- this is meant for C, not C++. - Changed -USE_GZIP to -DUSE_GZIP. - Added ability to compile ouside of the source tree. - Added .cc.o rule to {goo,ltk,xpdf}/Makefile.in. - Added @LIBS@ to XLIBS in xpdf/Makefile.in. - In top-level Makefile.in: added '-' to clean commands; added distclean rule. - Create install directories. - Use INSTALL_DATA (instead of INSTALL) for man pages. - Changed xpdf-ltk.h rule to avoid leaving an empty file when ltkbuild fails. - Change things so that by default, ltkbuild is not built and xpdf-ltk.h is not rebuilt. - Use AM_PROG_CC_STDC to add compiler flags for ANSI C. - Modify autoconf's builtin macros to check for xlC. - Use Steve Robbins' smr_CHECK_LIB to test for pixmap library (Xpm) -- this should fix the problems on systems that have the library but not the include file. - Added better test for sys/select.h, sys/bsdtypes.h, strings.h, bstring.h. - New VMS make scripts from Martin P.J. Zinser. - Moved dependences into Makefile.in -- this gets rid of problems with 'cc -MM' (which is gcc-specific) and 'include Makefile.dep' (which isn't supported by all make implementations). Also changed all non-system include files to '#include "..."' (from '<...>'). Tweaked the TextOutputDev heuristics slightly. Modify Gfx to use a stack of resources -- this is necessary for Form XObjects, which can define their own local resources; also modified PSOutputDev to dump fonts used by forms. Look for excessively large MediaBox (compared to CropBox) and shrink it to CropBox. Minor fix to scrolling when dragging a selection. Various patches for pdftex and Win32 support. Deal with Separation colorspaces by using their alternate colorspace entry. Added PBMOutputDev for pdftopbm application. Added ImageOutputDev for pdfimages application. Separated XOutputDev into LTKOutputDev + XOutputDev. Added support for 1-D and mixed 1-D/2-D (Group 3) decoding to CCITTDecode filter. Added resetImage(), getImagePixel(), and skipImageLine() to Stream class; used these in XOutputDev, PSOutputDev, and ImageOutputDev. Implemented predictor for LZW and Flate filters. In pdfImM1 in PSOutputDev prolog: div should be idiv. Changed output from printUsage() function in parseargs to look nicer. 0.90 (99-aug-02) ---------------- Added Type 1/1C font rendering, using t1lib. Added "backward" and "forward" buttons. Added fit-page and fit-page-width zoom factors; replaced zoom-in and zoom-out buttons with a zoom popup menu. Type 1C fonts are converted to Type 1 and embedded in PostScript. Support vertical Japanese text. Added Japanese text support (EUC-JP) to pdftotext. Bumped PDF version to 1.3. Added stub functions for ri and sh operators. (But there are still some missing 1.3 features.) Added -raw option to pdftotext. Minor changes to allow compiling under MS Visual C++ 5.0. Top-level makefile: changed 'mkdir -p' to '-mkdir -p'. Configure script: added X_CFLAGS to smr_CHECK_LIB(Xpm). Added Xpm_CFLAGS to xpdf/Makefile.in (this is needed to get the -I for the xpm includes). Rewrote code that handles font encodings; added support for Type 1C fonts. In the setpagedevice dictionary in PostScript output - added a /Policies entry which tells the PS interpreter to scale the page to fit the available paper. Changed PageUp behavior slightly: move to bottom (instead of top) of previous page. TextPage used character's dx (width + char space) instead of just its width. Read base URI from document Catalog (for URI-type links). Minor change to configure script to avoid using 'unset'. Fixed bugs in CropBox inheritance. Fixed a bug in resource stack handling for form objects. Display forms even if they have a missing/incorrect FormType. Fixed a bug in stream predictors -- the predictor parameters (width, comps, bits) don't need to match the actual image parameters. Completely rearranged the predictor code. Fixed PostScript output to correctly handle stream predictors. Don't segfault on empty (zero-page) documents. Added the xpdf.viKeys feature. Added the ffi and ffl ligatures to XOutputDev and TextOutputDev. Pdftotext and pdfimages now check okToCopy(). Added a '-q' flag to all programs (except pdfinfo) to suppress messages and errors. Deal with DeviceN colorspaces by using their alternate colorspace entry. Change PostScript output so setpagedevice is only called once, at the very beginning of the document (to avoid problems on duplex printers). Changes to configure script and makefiles for DOS/DJGPP. FontEncoding::getCharCode() looked for (code>0) instead of (code>=0). Added keypad arrow keys, etc. to xpdf. Minor changes to gfile.{h,cpp} () to compile under VC++. Fixed CCITTFaxStream to correctly handle all parameters. Modifications to gfile.{h,cc} for Acorn. Some minor changes for OS/2. Added 'SHELL = /bin/sh' to Makefiles. Compare file version to pdfVersionNum+0.0001 to avoid floating point precision problems. Added LDFLAGS to Makefiles. Removed strip entirely from configure and Makefiles. Fixed a bug in choosing the correct DCTStream color transform. 0.91 (2000-aug-14) ------------------ Added TrueType font rendering, using FreeType. Support for Chinese fonts (Type 0 fonts using the Adobe-GB1-2 character collection). Decryption code is included with the main distribution (now that the US export regulations are a little bit less irrational). Added (very basic) support for generating PostScript with Japanese fonts -- only tested with ghostscript so far. Added support for generating EPS files (pdftops -eps). Much-improved image filtering in xpdf (for downsampling and for transforms other than 0/90/180/270-degree rotations). Implemented a basic full-screen (presentation) mode (xpdf -fullscreen). (There is currently no way to switch between window and full-screen modes on the fly -- this will be fixed in a later release.) Added "reload" menu item. Do a better job with anti-aliased Type 1 fonts on non-white backgrounds. Handle Lab color spaces. Handle non-null user passwords. Avoid security hole with tmpnam()/fopen() -- added openTempFile() in goo/gfile.cc. [Thanks to Joseph S. Myers for pointing this out.] Filter out quote marks (' and ") in URLs before running urlCommand to avoid a potential security hole. [Thanks to Frank Doepper for pointing this out.] Fixed TrueColor detection to look at the complete list of available visuals, not just the default visual. In gfile.h, changed NAMLEN(dirent) macro to NAMLEN(d). Removed copyright character from start-up banners. In the open and save dialogs, if the open/save button is pressed with no file name, the dialog is not canceled. Added Xpm_CFLAGS to ltk/Makefile. XOutputDev::updateLineAttrs was using dashLength before it was defined. In top-level Makefile.in, use INSTALL_PROGRAM instead of INSTALL. In man page, recommend -q instead of -err in .mailcap. Changes for GNOME / Bonobo support: - Separated Stream classes into BaseStream and FilterStream trees. - Got rid of all FileStream-specific stuff. - Added a PDFDoc constructor that takes a BaseStream* instead of a file name. Allow newlines inside strings (Photoshop does this). Don't require whitespace between tokens in consecutive content streams (for pages that specify an array of streams). Look at MissingWidth when constructing font character widths array. Fixed a bug that caused incorrect PostScript output for images that use 8-bit indexed color spaces with < 256 colors in the palette. Handle case where embedded font file is bad (this was seg faulting). Minor changes for Windows/pdftex. Work around a bug in PDF files from the IBM patent server. Fixed bugs in PostScript form generation: use pdfStartPage instead of pdfSetup; problem with inline images. Minor bug fix in FlateStream::loadFixedCodes(). Added %%DocumentMedia and %%PageMedia comments to the PostScript so that gsview (a Windows frontend for ghostscript) gets the right paper size. Draw AcroForm fields that have appearance annotations. Bounds check gray, CMYK, and RGB values (in GfxColor). Moved the link border drawing code into Page (from PDFDoc). Minor modifications for pdftohtml. PSOutputDev: use the Type 3 font scaling kludge from XOutputDev. Separation color spaces were handled incorrectly in images. Fixed a bug with form bounding boxes. Modified the t1lib support -- replace libt1x code with my own code. Type 1 and TrueType fonts are now handled similarly, and clipping works on Type 1 fonts. Don't print copyright banner (xpdf); add -v switch to get copyright and version info (all apps); get rid of -err switch (xpdf). Automatically reload the PDF file if it has been changed, i.e., if the modification time is different. Fixed a memory (malloc size) bug in CCITTFaxStream. Fixed two bugs in FontEncoding::hash() -- handle zero-length character names (which were found in a (buggy?) PDF file), and handle character names with high-bit-set characters (use unsigned ints). Added PDFDoc::isLinearized() and corresponding code in pdfinfo.cc. Handle files with an incorrect page count in the Pages dictionary (FOP, from the Apache project produces a page count of 0). Handle TrueType equivalents to the Base14 fonts (Arial, TimesNewRoman, CourierNew) -- Adobe's tools use these names without embedding the fonts. Tweaked the Type 3 font sizing kludge. Changed pdfimages (ImageOutputDev) so it doesn't output JPEG files for 4-component color spaces, since these seem to confuse most image viewers. Added support for generating OPI comments (pdftops -opi). In XOutputDev::drawImage() and drawImageMask(), check for images that are completely off-page. Use the provided alternate or a default (DeviceGray/RGB/CMYK) color space for ICCBased color spaces. Incorporated MacOS-specific code from Leonard Rosenthol. Configure script switches to C++ for the strings.h/bstring.h test. Gfx::opRestore() calls clearPath() to handle (apparently) buggy PDF files produced by FreeHand. The /Type field in most dictionaries is optional (PDF 1.3 change). Move printCommands variable definition into Gfx.cc. If page is smaller than paper, center the PostScript output. Fix a minor bug in the SELECT_TAKES_INT detection in the configure script. TextOutputDev filters out control characters. Changed enough occurrences of 'char *' to 'const char *' to keep gcc 2.95 from barfing. Support for Latin-2 and Latin-5 in pdftotext (however, this will only work if the PDF file contains correct font encodings, which seems to be rare). TextOutputDev converts "eightoldstyle" to "eight", etc. Don't use the return value from sprintf() -- most systems return the length, but some return the string. Minor fixes for SunOS 4. Configure script looks for both select() and fd_set in sys/select.h. Configure script checks for gethostbyname() in -lbsd (for LynxOS). Fix missing closepath bug in PostScript output. Change PostScript portrait/landscape mode selection so it only uses landscape if the page width is greater than the paper width. Tweaked the VMS code in makePathAbsolute(). 0.91a (2000-oct-11) ------------------- Implemented separable CMYK PostScript output (the -level1sep switch to pdftops). Implemented Pattern color spaces with tiling patterns (polygon fills only). Implemented Stamp annotations. Implemented Named link actions. Fixed a really dumb bug in the TrueColor code in SFont (which affects both Type 1 and TrueType font rendering on 16-bit displays). Rewrote the GfxColorSpace / GfxColor code. Switched from djgppcfg to dj_make.bat (from Michael Richmond). Bug in the Type 1 encoding parser -- couldn't handle lines of the form 'dup NNN/name put' (with no space between the code and the name). Fixed the mkstemp() test in configure.in -- switched from AC_TRY_COMPILE to AC_TRY_LINK and added . Added DESTDIR to top-level Makefile.in. Fixed an incorrect OPI comment in PSOutputDev. Minor tweak to the CCITTFax code to avoid writing past the end of an array on an invalid data stream. Xpdf crashed if the user selected 'reload' when no document was loaded. Look for character names of the form "xx" (two hex digits with no leading alphabetic char) and 'cNNN' (decimal digits with a leading alphabetic char that happens to be a hex digit). FlateStream didn't correctly handle zero-length streams. Xref reconstruction didn't handle the case where the opening "<<" immediately followed "trailer" with no intervening whitespace. Fix the %%DocumentSuppliedResources comment in EPS output. Scale annotations to fit their rectangles. Added Stream::close() to handle cases (e.g., patterns) where a Stream object is used multiple times before it is deleted. Added the topLevel arg to Gfx::go() so it doesn't call out->dump() for every pattern element (and form). Rearranged the GfxResources class. Clean up white space handling in Lexer. Make the dpi parameter to PDFDoc::displayPage etc. a double - this avoids margin gaps with fit-page and fit-width. Fix a rounding problem in xpdf.cc that was causing the window to sometimes be one pixel too small. Fixed a minor bug in dealing with Base-14 TrueType font names. Fixed Lab -> RGB color space conversion. Added support for opacity values (from PDF 1.4) to GfxState and OutputDev. [Thanks to Leonard Rosenthol.] Implemented type 2 functions; rearranged the Function class hierarchy. 0.91b (2000-oct-29) ------------------- Print a warning about Type 3 fonts (XOutputDev, PSOutputDev). Added the scroll lock behavior to 'n' and 'p' keys in xpdf. Change FileStream buffer size to a #define'd constant. Renamed Pattern to GfxPattern to avoid clashes with Windows and MacOS types. Added CNS (Big5) Chinese font support (CHINESE_CNS_SUPPORT); renamed CHINESE_SUPPORT to CHINESE_GB_SUPPORT. 0.91c (2000-nov-19) ------------------- Fix an endianness problem in the Type 1 font code which resulted in an incorrect display with "-t1lib plain" on big-endian systems. CCITTFax stream decoder will skip over extra zero bits at end of line, even if EncodedByteAlign flag wasn't set. Added Big5 support to pdftotext (with CHINESE_CNS_SUPPORT enabled). Fixed a typo in the CNS/Big5 encoding translation table. Change the form code in PSOutputDev to store images in arrays of strings. The xref reconstruction (for damaged files) now also looks for 'endstream' tags, and the parser uses this information when setting up stream objects. In pdfinfo, convert Unicode chars in the 00xx range into 8-bit chars; print a warning if there are any other Unicode chars. 0.92 (2000-dec-03) ------------------ Fixed %%BeginResource comment (for xpdf procset) in PostScript output. Added "-title" switch and "xpdf.title" resource to set the window title. Check for in addition to . Upgraded the configure script to smr_macros 0.2.4 - this should fix a bug where configure wasn't correctly finding t1lib. 0.92a (2000-dec-17) ------------------- Added 'extern "C" { ... }' in various places for ANSI C++ compliance. Tweaked the code that figures out DPI for fit-to-page and fit-to-width modes. Fixed the image transformation code in XOutputDev -- no more missing lines. Implemented color key image masking in XOutputDev. 0.92b (2001-jan-07) ------------------- Fixed a bug in the error-checking code in the Separation/DeviceN color space parsing functions. [Thanks to Lidia Mirkin.] Added wheel mouse support (mouse buttons 4 and 5). [Thanks to Thorsten Schreiner.] Added preliminary support for FreeType 2 (disabled by default). 0.92c (2001-jun-04) ------------------- Fixed a bug in the new image transformation code. Look for character names of the form "", instead of looking for names beginning with a few specific letters. T1FontFile::T1FontFile wasn't initializing vars, and ~T1FontFile wasn't checking before calling T1_DeleteFont -- this caused crashes if it tried to open a nonexistent font file. Catalog::Catalog didn't set baseURI to NULL early enough. Tweak the check for strings.h in the configure script. Yet another fix for the image rotation code in XOutputDev -- off-by-one problem when upsampling. Handle Type 1/1C encodings when using FreeType 2. Allow FreeType2 to render user-supplied Type 1 base fonts. Opening a new file from full-screen mode tried to scroll. Fixed a bug in GfxFont constructor (missing check for NULL base font name). Don't crash if a Type 1 font's FontBBox is non-integer. Pdfinfo prints page size. Tweak for the alpha hack in T1Font/TTFont: sample the middle pixel instead of the top-left pixel. Automatically activate the text input widget in the find window. Changed a Japanese char code mapping in XOutputDev and TextOutputDev: period was being incorrectly mapped to small circle (end-of-sentence character). Add the 0/+/-/z/w key bindings to control the zoom setting. Fixed ImageOutputDev (pdfimages) to correctly handle inline image masks. Extract ascent/descent info from font descriptor. 0.92d (2001-jun-26) ------------------- Embed TrueType fonts in PostScript output. (Added a "-noembtt" flag to pdftops.) Extract encoding from TrueType fonts. Moved Function classes to a separate file (Function.h/cc). Implemented multi-dimensional sampled Functions. Implemented Type 4 (PostScript calculator) Functions. For Type 0 fonts, FontDescriptor is in descendant font, not parent. [Thanks to Lidia Mirkin.] Added the "-htmlmeta" option to pdftotext. In TextOutputDev, when computing the number of blank lines to insert, do a sanity check on the result. If both FlateDecode and some other filter (e.g., DCTDecode) were applied to an image stream, getPSFilter() crashed instead of just returning NULL. Handle the /Identity function. 0.92e (2001-aug-23) ------------------- Widths in font dict should override built-in font widths. Changed "rotate left/right" menu items to "rotate clockwise/counterclockwise". The link parsing code choked if the Border array was incorrect (too short). Modified PSOutputDev to output CMYK for fill/stroke colors. 0.93 (2001-oct-25) ------------------ Implement PDF 1.4 (128-bit) decryption. Bump supported PDF version number to 1.4. Text output for Simplified Chinese. [Thanks to Cheung Siu Fai.] Read an app-defaults file for Xpdf. Read a system-wide config file (/etc/xpdfrc) if ~/.xpdfrc doesn't exist. Accept and verify owner password; if correct, allow all actions. Added a "-level2sep" option to pdftops to generate Level 2 separable PostScript. The PostScript separation convention operators are used to handle custom (spot) colors. [Thanks to Thomas Freitag for help on this.] Add support for FreeType 2 to the configure script. Warning: this requires FT 2.0.5 or newer. Fixed the bounding rectangle overlap test in the disconnected subpath fill hack in XOutputDev. Stupid typo in font name table in PSOutputDev. Changing the zoom setting with a keyboard shortcut didn't update the displayed setting. Modified the mouse wheel support and added the second wheel (mouse buttons 6 and 7). [Thanks to Michal Pasternak.] Character and word spacing is affected by horizontal scaling (display and PS output). [Thanks to Eddy Ng.] Rotation specified by the text matrix, character spacing, and horizontal scaling interacted incorrectly (display and PS output). Some broken Type 1/1C fonts have a zero BBox -- kludge around this by assuming a largeish BBox. Handle PDF files with an incorrect (too small) xref table size. Allow "-?" and "--help" as aliases for "-h" (all apps). Correctly handle unescaped parens in strings in Lexer. Fixed a bug in LTK where a menu got posted multiple times if you right clicked while a page was being rendered. Removed a comma inside a string in configure.in. Kludge around broken PDF files that use char 32 but encode it as .notdef instead of space. Clean up various compiler warnings: use constructor args like "fooA" if there is a field named "foo". Everything now compiles cleanly under gcc 2.91.66, 2.95.2, and 3.0.1. Page objects now read all of the page rectangles (MediaBox, CropBox, BleedBox, TrimBox, ArtBox), as requested by the pdfTeX folks. Added a new PDFRectangle struct to hold these. Use XOutputDev's Type 3 font size hack in TextOutputDev too, so it does a little better job of extracting text in Type 3 fonts. Modify pdfimages to write one-bit images as PBM files. Work around a bug in cygwin's implementation of fseek. 0.93a (2001-nov-21) ------------------- Implemented the sh (shaded fill) operator for the axial shading type. Minor fixes to avoid compiler warnings. Cleaned up global variables -- moved many into instance vars and function args. Minor fixes for OS/2. Fix the system config file path for VMS. Fix an uninitialized var in XOutputDev that caused crashes on Alphas. Don't incrementally update the display in full-screen mode. For Type 1/1C fonts, use the FontBBox from the PDF FontDescriptor (instead of the one in the font file) if present -- this avoids problems with fonts that have non-standard FontMatrixes. Add the Euro character to WinAnsiEncoding. Track the bounding box of the clip region to make rendering patterns more efficient. Fix openTempFile() for Win32. 0.93b (2001-dec-11) ------------------- Added a duplex option to PSOutputDev and a -duplex switch to pdftops. Added XRef::PDFgetDocInfoNF() for pdftex project. Updated the VMS build script. 0.93c (2001-dec-12) ------------------- Completely rewrote the code that handles font encodings: - everything is Unicode-based - 16-bit fonts are handled much more cleanly - text output encoding can be set more flexibly New .xpdfrc config files. 1.00 (2002-feb-01) ------------------ More work on the font encoding rewrite: - use the ToUnicode font dict entry - pdfinfo and pdftotext (with '-htmlmeta') convert info strings to the selected text encoding Added key bindings for forward ('v') and backward ('b'). Added the pdffonts program which lists the fonts used in a PDF file. Fixed several problems in the TrueType font embedding code (for PostScript output). Accept named destination on command line. Added several new items to pdfinfo: file size, PDF version, tagged (yes or no), XML metadata (with the -meta option). Pdftops didn't get the portrait/landscape setting correct for PDF files with rotated pages. The TrueTypeFontFile class (including the Type 42 converter) now understands cmap format 6. Improved the "about" window -- mention the GPL, add a list of key bindings. Added Zcaron and zcaron characters to WinAnsiEncoding. The '0' keyboard shortcut didn't update the zoom popup menu. Handle the complete list of alternate names for the Base14 fonts. Fixed substitute font scaling in XOutputDev - scale only the width, not the height. Implemented stitching (type 3) functions. Handle the case of moveto/closepath/clip, which defines an empty clipping region. Move dependences into separate Makefile.dep files; get rid of the distdepend target. Move all of the configure-script-generated -D options out of the Makefiles and into a top-level .h file (aconf.h). Cleaned up the FreeType 1/2 detection code in the configure script. Pdfinfo prints dates in a more readable format. Fixed a bug in the Paeth image predictor. Handle annotations with multiple states. Another workaround for buggy X servers: clip points that are way out of bounds. Added libpaper support (for Debian). Generate PostScript DSC resource comments for PS (not just EPS) files. The save and restore (q/Q) operators shouldn't save/restore the path. Performance optimization: disable pattern drawing in TextOutputDev. 1.00a (2002-feb-25) ------------------- Added an optimized special case for one-bit images in XOutputDev. Implemented CID TrueType font embedding; added a psEmbedCIDTrueType option. The initialZoom X resource was broken. The reverse MacRoman encoding should return 32 for "space" (not 202, which is an alternate encoding). Tweaks to the FreeType 2 support: only disable hinting if the bytecode interpreter is disabled (i.e., disable autohinting but not bytecode hinting); add some padding to the glyph cache for CJK fonts. Added level3 and level3Sep options for the psLevel setting and corresponding -level3 and -level3Sep options to pdftops. Added a -level2 option to pdftops for consistency. Avoid a divide by zero in pdftotext. [Thanks to William Bader.] Added a Greek language support package. [Thanks to Alexandros Diamantidis and Maria Adaloglou.] Don't bother trying to extract a "builtin" encoding from a TrueType font. Accept either a page number or a page reference in a link destination. Update the fontFixedWidth flag in GfxFont after reading the char widths (used by the Acorn RiscOS port). Removed yet another (illegal but not caught by gcc) class specified from a .h file. Avoid using snprintf - it's not available everywhere. Improved the CMYK->RGB transform. Use mkstemps where available. 1.01 (2002-may-20) ------------------ Implemented Type 3 fonts. Implemented PostScript CID font embedding; added a psEmbedCIDPostScriptFonts option. Implemented PostScript 16-bit font substitution; added psNamedFont16 and psFont16 options. Moved the initialZoom setting from X resources to the xpdfrc file. Implemented the radial shading type in the sh (shaded fill) operator. [Thanks to Mike Sweet.] Added an 'include' command to the xpdfrc format. Added the displayNamedCIDFontX option so different fonts can be used within one character collection. Added a simple reverse video mode (-rv switch, xpdf.reverseVideo resource). Implemented stroked text in XOutputDev (with t1lib and FreeType2). [Thanks to Leonard Rosenthol.] Implemented stroked text in PSOutputDev. Added a built-in Unicode map for UCS-2. New key binding in xpdf: 'g' activates the page number text field. PSOutputDev will now embed external TrueType fonts in addition to external Type 1 fonts. The psEmbedType1Fonts and psEmbedTrueTypeFonts options were missing the "Fonts" suffix. Documentation in xpdf.1 for -freetype option was wrong. Added the Big5ascii Unicode map to the Chinese-traditional support package (maps 7-bit ASCII straight through). [Thanks to Lawrence Lai.] Modified the EUC-CN and EUC-JP encodings to pass 7-bit ASCII straight through. [Thanks to Lawrence Lai.] Avoid a divide by zero in XOutputDev. [Thanks to Simon Burge.] Remove old code in openTempFile that removed an extension from the name returned by tmpnam. Tweak the scrolling behavior when switching pages. [Thanks to Case Jones.] In the code that guesses character names (for font subsets), also handle names of the form 'ABnnn'. [Thanks to Colin Granville.] Fix the transform code for annotations. Improved the CMap file parser to handle more general PostScript lexical conventions. Added '-enc' option to pdfinfo. Added the small caps and oldstyle numbers from Adobe's Unicode corporate use area to the Latin1 and ASCII7 Unicode maps. The code in TextOutputDev that guesses Type 3 font size could generate a zero size, which resulted in div-by-zero errors. Various tools (including Adobe's) occasionally embed Type 1 fonts but label them Type 1C - so check for a '%!' at the start. Some tools embed Type 1C fonts with an extra whitespace char at the beginning - just skip over it. Fixed a typo in the Simplified Chinese add-to-xpdfrc file. Updates to dj_make.bat and the djgpp build instructions. Added a Turkish language support package. Updated VMS build scripts. [Thanks to Martin Zinser.] Modify the incremental display update code to redraw less often if most of the commands are vector graphics, as opposed to text and images. Tweak the Type 1 font bbox code to look at the bboxes in both the PDF font object and the embedded font file. Fixed the ETenms-B5-H CMap file (for traditional Chinese) to map the Latin characters to their proportional versions. Added an optional displayCIDFontX entry for one of the Arphic TrueType fonts in the traditional Chinese 'add-to-xpdfrc' file. Remove leading '-' on include statements in Makefiles. Added psASCIIHex parameter. Added the GBK Unicode map to the simplified Chinese language pack. Pdftotext now opens the text file in binary mode to avoid Microsoft's annoying automatic end-of-line translation stuff. Added an executeCommand function in goo/gfile.cc. [Thanks to Mikhail Kruk.] The %ALDImagePosition OPI comment was wrong if the page was scaled to a different paper size. The OPI code was saving the default transform matrix before calling setpagedevice, which can change the matrix. Fixed a crash when an inline image dictionary contains garbage. Upgraded to autoconf 2.53. Use unsigned int file offsets, to allow access to PDF files in the 2-4 GB size range; use fseek64/ftell64 if available. Fixed two floating point exception cases that came up with broken PDF files. Avoid a crash when printing an error message regarding an unnamed font. Default link border width should be 1. [Thanks to Michael Pfeiffer.] Minor tweak to build with FreeType 2.1.0. Handle "weird" characters in PostScript font names. PSOutputDev now handles PostScript XObjects. Added several more page attributes for the pdftex project. Transferred the copyright to Glyph & Cog, LLC. 2.00 (2002-nov-04) ------------------ Switched to the Motif toolkit. Support multiple open documents (in separate windows). Added document outlines to the viewer. Modified the text extraction (placement) algorithm. Implemented the JBIG2 decoder. Added a Latin2 language support package. Added support for movie annotations. Switched back to native LZW decompression code. Text extraction from Type 3 fonts was (partly) broken. The owner password checking code was missing a step in the case of 128-bit encryption. Added the 'printCommands' option to the xpdfrc file. Added key binding for '?' to bring up the about/help dialog. In TextOutputDev, ignore any text that's outside the page bounding box. Text extraction throws away "tiny" characters after the first 20000 per page, to avoid really slow runtimes with PDF files that use special fonts to do shading or cross-hatching; added the 'textKeepTinyChars' option to disable this behavior. Text extraction discards duplicated text (fake boldface, shadow effects). Added ctrl-F as a key binding for find. Added a "find next" function, bound to ctrl-G. Added ctrl-P as a key binding for print. Modified the DCT decoder to handle progressive and non-interleaved JPEG streams. Added key bindings for ctrl-Home and ctrl-End. Allow the initialZoom setting to be made in either the xpdfrc file or as an X resource. Added a Hebrew language support package. [Thanks to Roy Arav.] The "make distclean" target now creates (empty) Makefile.dep files in the three subdirectories. Initialize XRef::ownerPasswordOk. Correctly handle stroking of Type 3 fonts in PSOutputDev. Generate correct PostScript for fonts with "weird" character names (e.g., "("). Generate correct PostScript for images using Indexed color spaces with DeviceN base color spaces. Added lowercase Roman numerals to ISO-2022-CN.unicodeMap (simplified Chinese support package). Tweak the image scaling code to better handle flipped (top-bottom and/or left-right) images. Generate correct PostScript code for inline images and images in Type 3 fonts which are too large for a single PS string. Correctly handle indexed color spaces whose base color spaces have component ranges other than [0,1]. Optimized the DCT decoder. Fixed mistakes in the list of key bindings in the about/help dialog. Optimized the Flate decoder. Add literal names for punctuation and digits to the Unicode name table. Cygwin's popen wants mode "r", not "rb". Fixed a bug in the Type 4 function parser (the "if" operator wasn't parsed correctly). Fix a bug in PS output for TrueType fonts with no PDF encoding. Make the bbox size in FTFont more liberal (to avoid problems with fonts that have incorrect bboxes). Reverse the colors in PBM files generated by pdfimages, so the common case (an image mask filled with black) comes out correct. Add fseeko/ftello support which is basically identical to fseek64/ftell64. [Thanks to Nassib Nassar.] Modified column assignment in text extractor to account for characters that convert to multiple characters in the output encoding. Fix TrueType fonts which have an incorrect cmap table length. Work around a pragma bug in the version of gcc that ships with MacOS X 10.2. [Thanks to Frank Siegert and Andrew Stone.] Fix a problem that was causing an infinite loop when a damaged content stream contains an 'ID' command inside a dictionary. Handle the case where systempapername() returns NULL (libpaper support). Handle fonts which are defined directly in the font resource dictionary rather than as separate objects. Track process colors in Level 1 separable PostScript. Pdfinfo now checks the return value from mktime to avoid seg faults in flakey strftime implementations. If duplex is not enabled in PostScript output, leave the duplex setting alone, allowing the spooler to insert its own setting. Added three missing fclose calls. Change the default encoding for TrueType fonts (used when the PDF file doesn't specify an encoding) from MacRomanEncoding to WinAnsiEncoding. Move X_CFLAGS to the end of the list in CXXFLAGS (in Makefile.in) to avoid some of the FreeType2 include path problems. Fixed an obscure bug in the LZW decoder. [Thanks to Martin Schroeder.] Fixed a bug in decryption when using the newer (PDF 1.4) algorithm with shorter-than-128-bit keys. Minor optimization for image data streams: the ImageStream class can return an entire buffered line. 2.01 (2002-dec-05) ------------------ Redesigned the text extraction process: - process the text into "reading order" - added a "-layout" flag to pdftotext to switch back to the old style, where physical layout is maintained - use of the "-raw" flag is no longer recommended Added the -reload option for xpdf (in remote mode). Added support for external CID fonts; added the displayCIDFontT1 and displayNamedCIDFontT1 commands to the xpdfrc file. Handle the case of moveto/newpath/clip, which defines an empty clipping region (just like moveto/closepath/clip). Accept XYZ link destinations with missing array elements. Fix some problems with state save/restore triggered by Type 3 fonts that reference other fonts. Accept bogus font names based on "Symbol": Symbol,{Bold,Italic, BoldItalic}. Fixed color and font resource names in the xpdf man page. Was using delete instead of gfree in OutlineItem::~OutlineItem. Set the busy cursor in the find dialog while searching. Map variants of the copyright, trademark, and registered trademark symbols to the proper Unicode codes, not to Adobe's corporate use area codes. Fixed a floating point exception bug in TextOutputDev (check for a too-small denominator). Fixed a typo in TextOutputDev, in the code that generating blank lines to add vertical whitespace. Config files whose last line didn't end with a LF (or CR+LF) weren't being handled correctly. The code that handled CIDToGIDMaps in Type 2 CIDFonts was broken. Check the per-glyph bounding box in Type 3 fonts, and don't try to cache glyphs with bogus bboxes. Allow ToUnicode CMaps to use fewer than four hex digits in the Unicode char indexes. Added multithreading protection to the GlobalParams class. Fixed a bug in end-of-stream detection with the TIFF predictor. Added some characters to MacRomanEncoding to match up with Apple's definition. 2.02 (2003-mar-24) ------------------ Rewrote the text extractor code that assembles words into lines to better handle vertically overlapping lines. Add the "match" option for paper size (in PostScript output). Added support for external 16-bit TrueType fonts; added the displayCIDFontTT and displayNamedCIDFontTT commands to the xpdfrc file. Added an Arabic language support package. Added the Windows-1255 encoding to the Hebrew language package. A missing NULL check was causing a crash when closing the file in a single window (which clears out the window, but leaves it open). Deal with TrueType fonts whose glyph data is out of order - this affected both FreeType rasterization and PostScript generation. Munge font names in PSOutputDev to avoid names that are problematic for ghostscript because they start with an out-of-limits number (e.g., 1e999foo). Modify the TrueType font encoding deciphering algorithm in yet another attempt to match up with Acrobat's behavior. Bounds check the indexHigh value in indexed color spaces. The text extractor no longer bothers trying to get an average character width for Type 3 fonts, since it generally doesn't work very well (because Type 3 metrics are unreliable). Don't crash if the user hits ctrl-G ("find again") before doing a find. Set the button pixmap foreground color correctly. Handle text drawn backward on 180 degree rotated pages. Added a magic call to XtUngrabButton after calling XmCreatePopupMenu which appears to prevent some very odd problems (idea taken from the DDD source code). Fix the MacOS X fix (needed to include ). Fixed a bunch of Motif 1.x / X11R5 incompatibilities. [Thanks to William Bader and Albert Chin-A-Young.] Fixed various bugs in previously untested code in the JBIG2 decoder. Modify the XPDFCore destructor to avoid a bogus warning message from OpenMotif 2.2. Modified the Type 1C font parser to do proper bounds checking. Fixed the bounds checking in the TrueType font parser. Text extractor shouldn't do block merging in physical layout mode. Fixed a problem in PSOutputDev in level2sep mode with images in a Separation color space and with a non-default Decode array. Text extraction with "-raw" was concatenating lines from the bottom of one column and the top of the next. Handle Type 1C subroutines in the font converters. Correctly handle progressive JPEG images whose scans are slightly different sizes (e.g., the Y scan rounds up to a multiple of 8 pixels and the Cb/Cr scans round up to 16 pixels). Avoid a potential divide-by-zero problem in TextOutputDev. Modified the T1Font and FTFont modules to correctly handle glyphs that are larger than the font's claimed bounding box. Tweak dupMaxDeltaX parameter in TextOutputDev to avoid triggering on double characters. Improved detection in pdfinfo for ISO paper sizes. [Thanks to Hartmut Henkel.] Xpdf wasn't responding to the TARGETS atom, which prevented pasting the selection into various applications. [Thanks to Phillip Ezolt.] Handle XObjects with recursive references in their Resources dictionaries (in PSOutputDev). Change PSOutputDev to deal with invalid PDF files that use non-embedded TrueType fonts with no encoding. Check for undersized Widths arrays in fonts. Add bounds checking code to Array class. Updated VMS build scripts. [Thanks to Martin Zinser.] Tweak the TrueType font handling code (again): - char codes in symbolic fonts may or may not be offset by 0xf000 - discard empty tables because they sometimes confuse FreeType Fixed bounds checking in the Flate decoder. Removed a bogus error message for exponential functions without explicit C0/C1 values. [Thanks to Hartmut Henkel.] Handle the other Unicode cmap type (platform=0) in TrueType fonts. Added support for the SGI Motif horizontal paned window widget. [Thanks to Felix Ritter.] Ignore extra elements in link destination arrays. Accept external Type 1 font files with a suffix of ".ps" or no suffix at all. Add a bounds check in the DCT decoder. Added instructions for building xpdf.exe under cygwin/XFree86. Tweaked the word separation parameter for raw-mode text extraction. 2.03 (2003-oct-10) ------------------ Rewrote the text extractor to: - do a better job with rotated text; - handle right-to-left scripts; - be faster. Changed the zoom setting to use a percentage (relative to 72 dpi) instead of a zoom "factor". If the PDF file has an outline, open the outline pane initially. Added -f and -l options to pdfinfo; print multiple page sizes. The HAVE_XTAPPSETEXITFLAG test in XPDFApp.cc was backwards. The BitsPerComponent entry is optional in image mask objects. Render any annotation with an appearance stream, instead of just Widget and Stamp annotations. Fix a bug in the TrueType font checker: the test for an unsorted 'loca' table was wrong. Modify the TrueType cmap selection algorithm yet again to try to match Adobe's behavior. Changed sqrt(2) to sqrt(2.0) in pdfinfo.cc to make various compilers happy. Fixed a deadlock problem (when MULTITHREADING is set); cleaned up some other problems with the locking code. Fixed a bug in the interpolation code for type 0 (sampled) functions. Implemented type 1 (function-based) shaded fills. Fixed some stupid bugs in the JBIG2 decoder (introduced with the previous optimization work). Fixed a typo in the code that parses vertical font metrics for CID fonts that was causing a seg fault. Fixed a couple of bugs that were causing seg faults with badly damaged PDF files. Limit the number of nested Forms to avoid infinite recursion (in buggy PDF files). Add a special case for rectangular clip regions - make sure these don't drop pixels on the right and bottom edges. Tell FreeType not to use glyph bitmaps when in anti-aliased mode. Read all of the border style info for links. All of the shaded fill types now do at least one bisection to avoid problems when the colors at the endpoints of the domain are the same. If the Length2 parameter for an embedded Type 1 font was incorrect (too small), pdftops was losing font data. Deal with (broken) DCT streams that use the same component ID number for different components. The MediaBox page attribute was not being inherited correctly. Fixed a bug in the Type 1C font converter related to local subroutines. The Type 1C -> Type 1 font converter was allocating the font dictionary one slot too small. Added a missing private dictionary entry to Type 1 fonts generated by the Type 1C converter. [Thanks to Michael Shell.] Fixed bugs in the tiling pattern fill code. Try the TrueType 0xf000 char code offset hack for the MacRoman encoding too (in addition to MS Symbol). Update the font metrics info for the Base 14 fonts to include the Euro character. SECURITY HOLE: Escape various characters in URLs before running a web browser (or movie viewer). [Fixed in 2.02p11] SECURITY HOLE: In the dialog used to verify "launch" links, provide a scrolling view if the command to be run is excessively long. [Fixed in 2.02p11] Added an option to disable insertion of page breaks (form feed characters) in extracted text (pdftotext -nopgbrk; xpdfrc "textPageBreaks" option). Check for 8-bit fonts that specify an out-of-range FirstChar or LastChar. Correctly handle an obsolete Type 2 charstring op (in the Type 1C-to-Type 1 font converter). [Thanks to Helge Blischke.] Use the font encoding info to fill in holes in the ToUnicode map. Added character names for Bulgarian (in the Cyrillic support pacakage) and Greek. Handle clipping to text in xpdf and pdftops. Fix color space detection in DCT decoder. [Thanks to Dwight Kelly.] Added the "unicodeToUnicode" xpdfrc option, intended (initially) for Arabic support. Handle the case in PSOutputDev where two font objects refer to the same embedded TrueType font, but with different encodings. [Thanks to Frank Siegert.] Kill any pre-existing path before drawing a form (or annotation). Save state before rendering page content; restore state afterward. Fix Stream::reset/close to work correctly with encoder streams; fix PSOutputDev to use Stream::close consistently. Fix a seg fault when hitting the 'back' button after closing a file. GfxState::getStrokeGray was returning the fill gray value (this only affected Level 1 PS output). Change PSOutputDev to reuse dictionaries in Level 1 mode (since Level 1 PS interpreters don't do garbage collection). [Thanks to Frank Siegert.] PSOutputDev was generating incorrect translations for landscape-mode pages. Implemented shading pattern color spaces. PSOutputDev wasn't correctly handling Type 3 fonts which used image resources (as opposed to inline images). [Thanks to Frank Siegert.] The fix from 1.00 which clipped out-of-bounds points was a bit too aggressive. Do proper Floyd-Steinberg dithering in XOutputDev. Don't automatically check for a null owner password (to match Adobe's behavior). Allow the FlateDecode filter in Level 3 PostScript output. Fixed small bugs in the Type 1C -> Type 1 converter and Type 1C -> Type 0 converter. [Thanks to Tom Kacvinsky.] Work around another weird Motif problem with the right button menu (which was sometimes causing the menu to not be displayed). Make the code that handles fonts defined directly in the resource dict more robust. Add a brief description of the outline pane to the xpdf man page. Ignore extra operands to content stream operators. Fixed a bug in the CCITTFax decoder. Allow the Count entry in a Pages dictionary to be a real number (because some PDF generators actually do this). Shading pattern fills weren't being clipped correctly. Incorrect shallow copies in GfxRadialShading and StitchingFunction. The StitchingFunction destructor wasn't checking for funcs being NULL. Change the TrueType code-to-GID mapping code so it looks at the TrueType 'post' table. Set the print command in the print dialog once at startup, don't change it each time a file is (re)loaded. Generate the %%BoundingBox comment in regular PostScript files (not just EPS files). Fixed a bug in the Unicode CMap parser. 3.00 (2004-jan-22) ------------------ New PDF rasterizer ("Splash"). Added support for PDF 1.5: - JPX (JPEG 2000) decoder - XRef streams - object streams - DeviceN color spaces with up to 32 components - Added new CMaps to the CJK language support packages Replaced pdftopbm with pdftoppm (which can generate PBM, PGM, and PPM files). Reorganized the font file parser code into a new library ("Fofi"). Removed support for FreeType 1.x. Removed support for X server fonts - Xpdf (and pdftoppm) will now search for the URW fonts (from ghostscript). Changed the "-t1lib" and "-freetype" switches; replaced the "t1libControl" and "freetypeControl" config file options with "enableT1lib", "enableFreeType", and "antialias". Added the "-box" option to pdfinfo. Added imageable area support to PSOutputDev (for CUPS); added the "psImageableArea" config file option. Added the "-nocrop", "-expand", "-noshrink", and "-nocenter" switches to pdftops; added the "psCrop", "psExpandSmaller", "psShrinkLarger", and "psCenter" config file options. Dictionary size was in PostScript code generated for Type 3 fonts. The PS code generated for images in Type 3 characters was broken. Tweaked the text extractor. Accept xref entries that are one byte too short (matching Adobe's behavior). Change things so "xpdf -h" and "xpdf -v" work if DISPLAY isn't set. Fix a problem in the damaged file repair code that handles the trailer dictionary. Use the "Last" entries in "Outlines" objects - this avoids a problem with PDF files generated by buggy software that, e.g., sets the last item's Next pointer to point to itself. PSOutputDev was not handling DeviceN color spaces correctly in Level 2 images. Fixed a stupid little bug that broke PS output for JBIG2 images. Work around a Lesstif bug: set up an extra callback so hitting in the find dialog performs a search. [Thanks to Elliott Hughes.] Pdftops was crashing on zero page PDF files. Add an AC_PREREQ call to configure.in. Change the 'find' dialog so the text entry box resizes with the dialog. Skip extraneous zero bits at the start of a CCITTFax stream. The PostScript text clipping operator was missing a 'newpath'. [Thanks to Frank Siegert.] Fix a bug in tiling patterns with bboxes that don't start at (0,0). Fix a bug in Type 3 font handling with rotated text. The tiled pattern fill code was destroying the current path, which broke the fill+stroke operators when the fill color space was a tiled pattern. ICCBased color spaces don't always set their Ranges values correctly, so just use the values from the alternate color space. Modified GHash to accept int or void* - this avoids some conversion warnings. Check for missing Type 3 CharProcs - avoid a segfault. Pdffonts now marks all Type 3 fonts as embedded. Outline entries with no Title string weren't being handled correctly, resulting in segfaults. PSOutputDev now forces the text horizontal scale factor to be non-zero to avoid singular font matrices in the PS code. Tweaked the error recovery in the CCITTFax decoder. The LZW/Flate predictor should treat any Predictor value (in the stream dictionary) >= 10 identically. PSOutputDev and pdffonts check for NULL font objects (which can happen, e.g., because of missing CMap files). Swap the left and right mouse wheel button numbers. EPS output ("pdftops -eps") now uses the CropBox instead of the MediaBox as the EPS bounding box. 3.01 (2005-aug-17) ------------------ Added the continuous view mode, including the '-cont' switch and the 'continuousView' config file option. At high zoom levels, don't rasterize the entire page - this avoids problems running out of memory. Added "search backward" and "match case" options to the find dialog. Support explicitly masked images and soft masked images. Add support to DCTStream for 16-bit quant tables. Don't segfault if the user clicks on an outline entry with a broken destination. Changed the makefiles and configure script to skip building pdftoppm (in addition to xpdf) if X, Motif, or FreeType is not found; changed the error message in the configure script to match. Move an inline function in JArithmeticDecoder.cc to avoid compiler errors. Fixed a bug in the rasterizer that was sometimes causing infinite loops with round line caps on vertical lines. Various rasterizer optimizations. Look for intermediate resize events - try to avoid lagging when the user is doing an opaque resize. The FormType key in Form XObjects is optional. Handle external 16-bit TrueType fonts correctly, using the Unicode cmap. Add class declarations to TextOutputDev.h to work with stricter C++ compilers. Support FreeType's weird include file stuff (ft2build.h, etc.). Fixed a bug handling empty paths. Fixed a text positioning problem in PostScript output. Handle TrueType collections in FoFiTrueType.cc. FoFiTrueType constructor was reporting a failure if the post table was bad - this should be non-fatal. Type 1 font parser was missing a NULL test. Mask chars passed to isdigit in goo/parseargs.c to avoid problems with signed chars. Added more error checking to the CCITTFax decoder. Fixed a bug (computing the MCU size) in the DCT decoder. Change a test in the Splash stroke code to avoid x86 floating point weirdness. Reorganized the decryption code to allow security handler plugins; removed the NO_DECRYPTION #ifdefs. Added a plugin interface, initially just for security handlers. Support color key masked images and explicitly masked images in PS output (Level 2 only). When checking for aliases of the Base 14 fonts, ignore spaces in the specified font name. Handle encrypted PDF files that are missing the file ID string. Handle tiling patterns more efficiently in the PostScript output. Rewrote the code that handles color spaces in PostScript output. Fixed a bug in the Type 1C font parser - zero-length indexes (and zero-length names) weren't handled correctly. Handle shaded fills more efficiently in the PostScript output. Implement the remaining shading types (4-7). Rearranged the Splash color modes. Add the EarlyChange parameter to LZWStream when generating PostScript. Check for zero values in line dash arrays in PSOutputDev. Fixed an uninitialized variable in JArithmeticDecoder which was causing crashes. Treat unknown CMap names as identity mappings (to match Adobe's behavior). Fixed bugs in the XRef parser related to XRef streams in updated files. Added a missing call to FT_Done_Glyph which was causing a memory leak. [Thanks to Dave Formanek.] Fixed a bug in text copying that was causing the last word to be dropped on some pages. Tweaked the image width/height computation in Splash::drawImage and Splash::fillImageMask to make striped images work better. Ignore minus signs in the middle of numbers (to match Adobe's behavior). Missing '%s' in format strings for dates in pdftotext '-htmlmeta' mode. Change the TrueType code-to-GID mapping code so it looks at the standard name-to-Unicode mapping before the ToUnicode mapping defined in the font object. Added a matteColor setting (command line option and X resource). Tweaked the CMYK->RGB transform. Fix some problems in tracking the character position (to match up with Adobe's highlight file format). Handle moveto/closepath/stroke correctly. Check for singular text matrices and font size of zero in PSOutputDev. Clip PS output to the size of the page (avoiding any gibberish that lies outside the MediaBox, in the case where the MediaBox is smaller than the paper). If the line dash element in an annotation's Border array is of an invalid type (i.e., not an array), don't draw the link at all (this matches Adobe's behavior). Don't remap small caps and oldstyle glyphs in the name-to-Unicode table - it messes up TrueType font encodings. Pdftoppm wasn't setting the paper color correctly in mono and gray modes (this only showed up on big-endian machines). Missing NULL check was causing crashes when attempting to read non-PDF files that happened to contain the string '%PDF'. Fixed a problem in the text extractor that was breaking up words. Handle vertical text (CJK fonts) in PS output with TrueType fonts that are missing the vertical metrics tables. Handle the case where a font object and the corresponding embedded font are different types. Handle basic crypt filter functionality. Added more value checking in the XRef parser, to avoid potential security problems. Updated the CJK language support packages: replaced the displayCIDFontX references with displayCIDFontTT; added pointers to free TrueType fonts. Added a missing error message when SplashOutputDev can't parse an embedded TrueType font file. PDFCore and TextOutputDev now correctly handle searching for Unicode strings, including real Unicode case-folding. Throw away tiling pattern fills that are completely outside the clip region. The JPEG 2000 inverse reversible multiple component transform code was wrong. Fixed some bugs in shading pattern fills: clipping was wrong, and background color was not implemented. Added tool tips for the toolbar buttons. Decrease the max depth of recursive patch mesh filling if the pattern has a large number of patches. Highlight the find text whenever the find dialog is mapped. Handle page boundary boxes with reversed coordinates. Fixed a bug in the text extractor code that handles duplicated text. Optimization work on SampledFunction::transform(). Use the CropBox instead of the MediaBox as the display region. Dither for PseudoColor (8-bit) displays. Fix a bug in DCTStream that was causing an infinite loop with corrupted DCT image data. Fix a bug in the ToUnicode CMap parser. Fix a bug in the text extractor - negative font sizes weren't being handled correctly. Fix a bug in the text extractor - in certain cases, out-of-bounds text could cause crashes (generally only in damaged PDF files). Fix a read-past-end-of-array bug in the JBIG2 decoder. Fix a case where pdftops was generating lines longer than 255 chars. Optimize redraws - don't regenerate the XImage every time redrawRect is called. The ASCII85 decoder wasn't skipping whitespace properly. Optimize text extraction: skip (non-inline) image setup entirely. Added initial transparency support (stroke/fill alpha and blend mode). Added support for the overprint setting in PostScript output. Fixed various buffer overflow bugs. Handle negative font sizes and horizontal scaling correctly - this affected PSOutputDev for all text operators, as well as the TJ operator for all OutputDevs. Fixed a buffer overflow in the CCITTFax decoder. Fixed an out-of-order entry in the list of font name aliases. Fixed a backward loop in the PostScriptFunction code. Treat a zero-length base URI the same way as a nonexistent base URI. Add a divide-by-zero check in TextOutputDev (the problem was happening in cases of mixed horizontal and vertical text). PSOutputDev wasn't rounding the page bounding box coordinates correctly. Support the SOF1 marker in DCT (JPEG) image streams. Minor changes to GlobalParams.h and JPXStream.h because some compilers don't like anonymous structs inside anonymous unions. Xpdf now complains about a negative page number. Changed GString::cmp and GString::cmpN to correctly handle '\0' chars in the middle of strings. Fixed the radial shading code; corrected the handling of the 'extend' parameters. Added the gmallocn and greallocn functions. Fixed a bug in the TIFF image component predictor which shows up with components that are not 1 or 8 bits wide. Optimized FlateStream::loadFixedCodes(). For non-embedded Base-14 fonts, don't use the ascent/descent/bbox values from the FontDescriptor - various PDF generators get them wrong. Fixed a bug in the text extractor - words on the same line (especially in tables) were being split vertically onto multiple lines. Automatically select the correct radio button ("print with command" vs. "print to file") in the print dialog. Don't create the "open" and "save as" dialogs until needed - this avoids stat-ing every file in the directory at startup. Changed the Big5 and Big5ascii encodings (in the traditional Chinese language support package) to include characters from the Unicode database (which aren't mentioned in the Adobe character collection documentation). Added the '-pagecrop' switch to pdftops. Tweaked the RGB->gray and CMYK->gray conversion functions to match the PDF spec. The JPEG 2000 decoder wasn't correctly handling codeblocks split across multiple packets/layers. Fixed a typecast that caused compile errors on 64-bit systems. The CMap parser wasn't handling the 'cidchar' construct. Handle the case in PSOutputDev where two font objects refer to the same embedded 16-bit TrueType font, but with different CIDToGIDMaps. Changed the configure script to report more accurate warnings when it can't find X / Motif / FreeType. Encryption with revision=2 always uses a 40-bit key, regardless of the specified Length value. Yet another minor change to the TrueType font encoding deciphering algorithm. Don't completely invalidate the Catalog if one (or more) of the page objects are bogus -- just skip over those pages. Removed the workaround in pdftops for too-small Length2 values in Type 1 fonts -- it was causing problems on various PostScript printers. Started adding error checking to the JBIG2 decoder (this is nowhere near complete yet). Extended the "unicodeToUnicode" config option to also apply to CID fonts. Added the narrow Latin characters to the Adobe-Korea1.cidToUnicode file in the Korean language support package. Fixed the code that handles page rotation in PSOutputDev. When converting a Type 1C glyph to a Type 1 glyph, insert closepath operators as appropriate. Check for a sane 'loca' table in TrueType fonts (FoFiTrueType::parse). Fix PSOutputDev to correctly handle the case of an empty name in a font encoding. 3.02 (2007-feb-27) ------------------ Added anti-aliasing for vector graphics; added the vectorAntialias xpdfrc option; added the "-aaVector" switch to xpdf and pdftoppm. Implemented stroke adjustment (always enabled by default, ignoring the SA parameter, to match Adobe's behavior), and added the strokeAdjust xpdfrc command. Support PDF 1.6 and PDF 1.7. Added support for AES decryption. Added support for OpenType fonts (only tested with 8-bit CFF data so far). Added user-configurable key/mouse bindings - the bind/unbind xpdfrc commands. Cleaned up the full-screen mode code and added the ability to toggle it on the fly (the default key binding is alt-f). Pdfimages with the -j option now writes JPEG files for 1-component (grayscale) DCT images, in addition to 3-component (RGB) images. Fixed bugs in handling sampled (type 0) functions with 32-bit samples. Fixed some things to support DeviceN color spaces with up to 32 colorants. Pdftops now constructs the %%Creator and %%Title DSC comments from the relevant information in the PDF Info dictionary. Tweak the TrueType font encoding deciphering algorithm. Added the "mapUnkownCharNames" xpdfrc option. Fix a bug (that only showed up with certain window managers) in the intermediate resize event optimization. [Thanks to Michael Rogers.] Check for a broken/missing embedded font (this was causing xpdf to crash). Added support for transfer functions in PostScript output. Be a bit more tolerant of Link destinations that contain null values for positioning parameters. Use ordered dot dithering instead of clustered dot dithering at resolutions below 300 dpi (for monochrome output). Fixed security holes (bounds checking issues) in several places. Don't bother creating a SplashFont (allocating memory) for fonts that are only used for hidden text - this avoids problems with fonts of unreasonably large sizes. Clipping in TextOutputDev was off for characters on the left edge of the page. The scn and SCN operators weren't correctly handling colors with more than four components. FoFiType1::writeEncoded wasn't always correctly finding the end of the encoding. Use the ColorTransform parameter in the DCTDecode stream dictionary. Type 3 fonts are allowed to have a bbox of [0 0 0 0], which means "unspecified" -- don't issue error messages in that case. Perform the transform (to device space) in Splash instead of in SplashOutputDev -- this is needed to correctly handle round joins and caps on stroked paths. PSOutputDev now rasterizes any pages that use transparency. Limit the crop, bleed, trim, and art boxes to the edges of the media box (per the PDF spec). Change GString to increase the allocation increment by powers of two. Handle whitespace in hex strings in CMap files/streams. Use strings instead of names for separation colorant names in PSOutputDev. For explicitly masked images where the mask is higher resolution than the image, use the soft mask code. Avoid problems with very large x-steps in the PostScript output for tiling pattern fills. Avoid a divide-by-zero in stitching functions which have a subfunction with empty bounds. Honor the "Hidden", "NoView", and "Print" flags on annotations. Rewrote the pixel rendering code in Splash to use a single set of pixel pipeline functions. Added support for transparency groups and soft masks. Fixed the transparency blend functions to match the addendum published by Adobe. Changed Splash/SplashBitmap to store alpha in a separate plane. Setting the color space now selects the correct default color for that color space. Remove the mutex lock from GlobalParams::getErrQuiet() to avoid a deadlock when parseCIDToUnicode() or parseUnicodeToUnicode() calls it from inside a locked section. Added error checking (on the argument count) in the sc/SC/scn/SCN operators. Skip over notdef glyphs in TrueType fonts (which sometimes get drawn as little boxes), to match Adobe's behavior. Painting operations in a Separation color space with the "None" colorant or a DeviceN color space with all colorants set to "None" never mark the page. Fixed an obscure bug in the JPX decoder - it wasn't reading the extra stuffing byte in the case where the last byte of a packet header was 0xff. Change the TrueType font parser (FoFiTrueType) to change the glyph count rather than report an error if the 'loca' table is too small. Fixed a couple of bugs in the JBIG2 decoder. Added stochastic clustered dot dithering. Added the screenType, screenSize, screenDotRadius, screenGamma, screenBlackThreshold, and screenWhiteThreshold xpdfrc settings. PSOutputDev now correctly handles invalid Type 3 charprocs which don't start with a d0 or d1 operator FreeType 2.2.x support - get rid of the FT_INTERNAL_OBJECTS_H include, and add some 'const' declarations. Handle PDFDocEncoding in Info dictionary strings. Tweak the xref repair code - ignore whitespace at the start of lines when looking for objects. Added the "-exec" switch to xpdf. Removed the xpdf.viKeys X resource. Changed the color key / explicit masked image code in PSOutputDev to generate better PS code, including a Level 3 option. Tweaked the DEBUG_MEM code for performance. Move the JBIG2 global stream reading code into reset() instead of the constructor - this way, pdftotext doesn't end up reading the global stream. Added the "-preload" option to pdftops and the psPreload xpdfrc command. Added the "zoom to selection" command (on the popup menu). Fix a bug (in xpdf/pdftoppm/pdftops) with tiling patterns whose bbox size is different from their xStep/yStep. Implemented stroke with pattern color spaces. Following a link to a page whose CropBox was different from the MediaBox was resulting in an incorrect scroll position. Parse truncated date strings from the Info dictionary correctly. Change FoFiType1 to handle Type 1 fonts with two /Encoding keys. Extend the PSOutputDev shaded fill code to handle DeviceCMYK shaded fills in level2sep and level3sep modes. Detect infinite loops in the Page tree. Optimized the ASCII85Encoder code. Tweaked the text extractor to do a better job of lining up rows of text. Leave images compressed (or re-compress them with RLE) in PostScript output when setting up images for forms and Type 3 fonts (or with -preload). Extend FoFiType1 to handle Type 1 fonts with octal character codes in their encodings. Use a custom string formatter to avoid problems with locale-based decimal formatting (commas instead of periods) in PS output. Allow comments in PostScript-type functions. Change the TrueType font parser (FoFiTrueType) to delete glyf table entries that are too short. 3.03 (2011-aug-15) ------------------ Added the "fixed pitch" text extraction mode. Modified "pdftops -paper match" to handle PDF files with different-sized pages, i.e., it will now select the matching paper size on a page-by-page basis. Add ability for pdftoppm to write to stdout. Added the pdfdetach tool. Implemented 256-bit AES decryption. Commented out the t1lib section in the configure script -- t1lib has some potential security holes, and hasn't been updated in years. Redesigned the font configuration xpdfrc commands: removed the displayFontT1, displayFontTT, displayNamedCIDFontT1, displayCIDFontT1, displayNamedCIDFontTT, displayCIDFontTT, psFont, psNamedFont16, and psFont16 commands; added the fontFile, fontFileCC, psResidentFont, psResidentFont16, and psResidentFontCC commands. Switched from GPLv2 to dual v2/v3 licensing. Performance: cache tiling patterns. Implemented text fills with pattern color spaces. Rewrote the image and image mask rendering code to be more accurate and faster. Fixed a bug in PDFCore that could sometimes cause crashes at high zoom levels. Implemented embedded CMap streams. Added the 'setSelection' command. Added the 'rotateCCW' and 'rotateCW' commands. Added the 'psFontPassthrough' xpdfrc command. Added the 'launchCommand' xpdfrc command. Implemented alpha-type soft masks. Added "Form: AcroForm|XFA|none" to pdfinfo output. Added support for image masks filled with pattern color spaces. Text search wasn't finding all occurrences of a string on rotated pages (i.e., pages where the primary rotation, as displayed, was not horizontal). The text extractor now uses "ActualText" spans if they are present. Modified PSOutputDev so it always check for transparency; if Splash is not available, it now prints a warning. Handle the blending color space for soft masks. Added the disableFreeTypeHinting xpdfrc option. Added the psAlwaysRasterize xpdfrc option. Added support for transfer functions in the rasterizer. Optimized the JPEG 2000 decoder to use less memory. Do fill adjustment (similar to stroke adjustment) on simple rectangular fills. Added the antialiasPrinting xpdfrc setting. Added '%i', '%j', and '%k' (mouse pointer page and position) to the available options for the 'run' command. Links with the underlined border style were being drawn with the lines over, instead of under, the text. Add #include to XPDFTree.cc (to avoid problems with certain compilers). Change XRef::fetch() to avoid infinite loops caused by odd damage to the xref table (e.g., where a stream object's "Length" value is an indirect reference that points to another stream object). Minor fix in GString.cc to deal with an error in newer C++ compilers (pow() can take int or double args). Handle embedded fonts which are declared with the wrong font type, including 8-bit fonts declared as CID fonts and vice versa; this was causing various problems, including crashes and invalid PostScript output. In text extractor, don't drop horizontally overlapping words onto a separate line. The numbers in the operand to the TJ operator should be multiplied by the current horizontal scaling parameter. Fixed a bug in the Type 1C-to-Type 1 font converter -- need to escape strings in the font dictionary. The zero-font-size check in PSOutputDev.cc was broken. Fixes for the form field appearance regeneration code: handle Unicode strings (by downconverting to Latin1); check for "True" in addition to "Yes" as a button value. Modify XPDFTree to limit the widget height to 32767 pixels -- this avoids crashes with very large outlines. Modify FoFiType1 to handle PFB headers in Type 1 font files. Allow image mask decode arrays to be [1.0 0.0] in addition to [1 0]. Tweak the form field appearance regeneration code. PSOutputDev now sets up resources for DR dictionaries in forms (in case the form field appearances are regenerated). For TrueType fonts which are marked symbolic, the cmaps should be ignored. Change the handling of numeric characters in Unicode text output -- they are now treated as left-to-right, which isn't strictly correct, but does result in correct visual formatting. Modify FoFiTrueType to handle bogus loca table entries where the offset is past the end of the glyf table. Fixed shading pattern fills to do clipping and background fills correctly. Change the code that reads the page tree to be more flexible with regard to improperly constructed trees. Changed the PostScript header comment from "Produced by xpdf/pdftops x.yy" to "XpdfVersion: x.yy" to make it (mostly) DSC-compliant. Fixed PSOutputDev to handle page rotation correctly where the PDF page is rotated 90 or 270 degrees and the page width (before rotation) is greater than its height (and vice versa for 0 or 180 degrees). Unbalanced save/restores inside form XObjects (which can be happen in damaged or incorrectly constructed PDF files) could cause crashes. The CCITTFax decoder could go into an infinite loop on certain types of corrupt input. Added the "drawAnnotations" xpdfrc command. Added the "psUncompressPreloadedImages" xpdfrc command. Escape newlines and other non-printable characters in the custom color DSC comments in PostScript output. Added the tilingPatternFill and *ShadedFill functions to PreScanOutputDev to speed it up. Too many elements in an image's Decode array should not be a fatal error. Fixed a buffer overflow security hole in StreamPredictor. Empty pages (zero-length content streams and null contents) were causing crashes in continuous view mode. Handle line dash arrays that start with a zero element correctly. PreScanOutputDev was not correctly detecting all transparency - in some places, it was looking only at the blending mode, and not at the opacity setting. Force halftone screen size to be a power of 2. This allows optimizing the halftoning code. Fixed a bug in the JBIG2 MMR decoder (and also in the CCITTFax decoder) that was causing array references with negative indexes. Fixed a bug in the transparency code that was causing memory overruns. Fixed a 64-bit bug in the ASCII85 encoder. The focusToPageNum was crashing in full-screen mode - it should simply do nothing. Added '%p' (current page number) to the available options for the 'run' command. Tweak the behavior with PDF files that have pages of different widths in continuous mode: scroll horizontally to avoid blank space on the left side (when changing pages via entering a page number or clicking on a link). A closepath followed by a lineto should create a new subpath for the lineto. Fixed a buffer overflow in the CCITTFax decoder. Adobe Acrobat appears to ignore the flatness setting when rasterizing curves, so Xpdf now does the same. (Oddly, Acrobat passes the flatness setting through to PostScript output, which ends up making the PS file look different from the PDF file - so Xpdf continues to pass flatness through to PS files, too.) Pdfimages now ignores tiling pattern fills (for performance reasons). Fixed a bug in PDFCore that could sometimes cause crashes at high zoom levels. Use std::sort (with functors) in place of qsort (if available) - this can be significantly faster. Hitting "close" while Xpdf is in full-screen mode (with only one file open) was crashing. Tweak the TrueType font encoding deciphering algorithm. Rewrote the CCITTFax decoder inner loop - this fixes a security hole. Fixed two security holes (missing bounds checks) in the DCT decoder. Do the correct zooming for "Fit" and "FitR" destinations. Implement the rotation value in the form field appearance regeneration code. When PSOutputDev rasterizes a page (because it contains transparency), the image is broken into stripes if it is above a size threshold. Don't clip the other page boxes to the MediaBox at the intermediate (Pages) nodes; only do it at the leaf (Page) nodes - the other boxes can be specified before the MediaBox is specified. Split several special cases off from Splash::pipeRun, for performance. Add a sanity check for excessively large font sizes (which caused problems because Xpdf tried to allocate memory for a font cache). Fixed a bug in the GfxCIDFont constructor involving reading the vertical metrics. Rewrote the code that handles annotation transforms - it was not handling non-rectangular transforms correctly. Tweak the Type 3 bbox code to allow some slack on the left/bottom edges, as well as the top/right edges. Fixed a bug that was causing problems rendering Type 3 fonts to 1-bit monochrome output. Handle the case where AES padding is invalid. Changed XRef::getNumObjects to return the number of used entries, instead of the same value as returned by XRef::getSize(). Ignore bogus entries in the TrueType table directory instead of immediately giving up on the font. Fixed a bug in the radial shading code that could cause divide-by-zero errors. Tweaked the TrueType font fixup code to deal with an extra zero entry at the end of the loca table. Increased font cache sizes to improve performance. Tweaked the TrueType notdef kludge to skip the notdef glyph only if it was caused by a character for which a Unicode/MacRoman mapping could not be found. Added another font size sanity check, this time for Type 3 glyphs. Added initial support for optional content (layers) - no GUI yet. For CID fonts that specify the 'Adobe-Identity' or 'Adobe-UCS' collection, use an identity mapping for char code-to-Unicode conversion. Updated the error function to take a category argument, and to use GString::format instead of printf; added an error callback. The PDF spec claims that names are limited to 127 chars, but Distiller 8 will produce longer names, and Acrobat 8 will accept longer names -- Xpdf will now handle them, too. Change the Catalog code so it doesn't load the entire page tree at startup (this only helps the command line tools - the viewer scans all of the pages at startup). Clip opacity values to the range [0,1]. Handle glyph names of the form 'unixxxx' (similar to Ann, Axx, etc.). Resolution of rasterized pages in PostScript output was not being computed correctly (resulting in overly large images). Extend the mapUnknownCharNames config command to cover CID fonts. Zero-length segments in the middle of stroked paths were not being handled correctly -- they should be completely ignored, except for the special case of a zero-length path with round line caps. Various optimizations to the Splash rasterizer. Allow "Identity" and "Default" as the transfer function in soft masks. Tweaked the ToUnicode CMap parser to allow <00xx> char codes for 8-bit fonts. TextPage::clear() was not clearing the lists of underlines and links. Changed the CCITTFax decoder to correctly handle the interaction between the EndOfLine and EncodedByteAlign parameters. Fixed a bug where xpdf wouldn't go to a destination specified on the command line, if continuous mode was enabled (xpdf -cont file.pdf +foo). Fixed a bug in the FreeType interface code which was causing incorrect positioning/sizing of characters. Tweaked the FreeType interface code - use 'light' hinting with Type 1 fonts (because it generally looks better). Tweak the Windows font searching code to handle "Black" fonts. Fixed a bug in the PostScript-type function parser -- real numbers that start with a decimal point weren't being handled correctly. Changed the way filled-and-stroked text is handled -- use Splash to do both the fill and the stroke (rather than using the font ending to do the fill and Splash to do the stroke); also turn off stroke adjustment when drawing stroked text. Ignore generation numbers on compressed objects (to match Adobe's behavior). Changed the PostScript PageSize policy from 3 to 6 (choose next largest paper size, don't rescale). Check that the mask values for color key masked images are in range -- both Ghostscript and Distiller choke on the PostScript if invalid mask colors are specified. Fixed a bug in generic region decoding in the JBIG2 decoder. Fixed the lexer to handle large real numbers correctly. Pdftops wasn't correctly handling tiling patterns that use both fill and stroke operators. Report an error if the Size array in a sampled function includes any zeros. The PostScriptFunction.codeString field wasn't being initialized correctly. Invalid object streams were not being handled correctly. Check for loops in xref tables. Handle the case where a content stream ends after the 'ID' operator which starts inline image data. Check for invalid object streams (i.e., an object stream can't be inside another object stream. Add a recursion limit to the object parser - this avoids stack overflows with various sorts of damaged PDF files. Handle various parameter settings in the extended graphics state dictionary. An invalid document outline (missing fields) could cause a crash. Added an overprint preview mode (for CMYK output only). Correctly handle FitH/BH/V/BV link destinations that contain null values for positioning parameters. If the PDF file doesn't define a BaseURI, set one based on the location of the PDF file -- this allows relative links to be handled correctly. Use ResusableStreamDecode when generating Level 3 PostScript for an explicitly masked image. Tweak the Type 1 font parser to handle encodings with multiple characters on one line. Invert subtractive color components before passing them to the blending function. Fix an invalid array access in SplashOutputDev::setSoftMask() in CMYK mode. A PS interpreter may attempt to read past the end of a preloaded image, so we need to check for that. Fixed several overflow/uninit bugs in JBIG2Stream. Fix the CCITTFax decoder to correctly find end-of-file (RTC/EOFB) markers when the EndOfLine parameter is false. Don't limit the startxref offset to 10 digits - some PDF generators use extra leading zeros. Use the "DV" (default value) field in text annotations if the "V" (value) field is missing. Increase the number of digits printed for floating point numbers in PostScript output -- it was running into numerical accuracy problems on large pages. Fixed integer overflow bugs in Catalog.cc and GfxFont.cc. Zero-length tables in TrueType fonts should be treated as missing. ToUnicode CMaps map char codes to Unicode; .cidToUnicode files map CIDs to Unicode -- ToUnicodeCMaps were being handled incorrectly. Added the "psRasterResolution" and "psRasterMono" xpdfrc commands. Added code to FoFiTrueType to check for entries in the table directory with bogus tags -- this handles the case where the number of tables given in the header is too high. Negative shift values in the PostScript bitshift operator were being handled incorrectly. GfxICCBasedColorSpace was not correctly bounds-checking the number of components. Check SampledFunction input values for NaNs. Fix a divide-by-zero when the page width or height is zero. Fix a bug positioning text in PostScript output - if the last char in a string has an incorrect width in the PDF font object, that messes up the total string width, so we have to position individual chars. In PostScript output, if CID font substitution fails, drop all text in that font. Handle PDF files that set the stream/string decryption filters to Identity (i.e., no encryption). Avoid passing a zero font size to FreeType. Tweak raw mode in the text extractor to handle words on the same line drawn in the wrong order. Pdftops was generating a singular transform matrix for annotations whose bounding boxes had zero width or height. Handle embedded OpenType CFF fonts with CIDToGIDMaps. Remove the old kludge from PSOutputDev that was converting char 32 from ".notdef" to "space". Allow CCITTFax images to be more than 32k pixels wide. Modified the CMap parser to handle usecmap with Identity-H/V. Added some parameter checking in the JPX decoder. Added checks for infinite loops in PDF objects - for color spaces, functions, optional content, and outline items. Add support for mouse button bindings up to button 32 (old limit was 7). Fixed a bug in the decryption code for revision 3 with keyLength < 16 -- the owner password was not being handled correctly. [Thanks to Matthias Franz.] Optimize SampledFunction: pull index computation code out of the transform function; cache the last transform. Tweaked the font naming algorithm in PSOutputDev. Treat mirrored (as opposed to rotated) text the same as upright text in pdftotext. Fixed buffer overflows in Splash and SplashBitmap. Check for bogus character codes (e.g., ) in ToUnicode CMaps. The radial shading code (in both Gfx.cc and PSOutputDev.cc) was not computing the s bounds properly. Drop empty subpaths in clip and fill (but not stroke) operations - this can significantly speed up clip performance in the weird case where a PDF file does " re m h W n". Added code to FoFiTrueType to check for an invalid loca format field in the head table. The axial shading code (in Gfx.cc, but not PSOutputDev.cc) was not computing the incremental polygon vertices correctly. Set the character width in TextOutputDev to something sensible for vertical fonts. Added basic support to the text extractor for vertical writing mode. The non-interactive tools (pdftotext, pdftops) now free Page objects after using them, avoiding performance problems with pages that have huge resource dictionaries. Check for line dash arrays like [0], and draw nothing at all (to match Acrobat's behavior). Add a sanity check for the ascent/descent values in FontDescriptors. Single-point paths with a line dash were causing a crash. Correctly handle Level 3 PostScript output with masked images inside patterns. Tweaked the xref repair code so that it runs if the catalog is messed up (in addition to running if the xref table is damaged). If Indexed color space tables (streams or strings) are too small, reduce the max index instead of throwing away the color space. Change the CMap parser to allocate memory only when it sees a mapping, not when it sees a 'codespacerange' declaration -- this avoids allocating huge amounts of memory for CMaps with large, unused codespaceranges. In monochrome mode, treat lines with width <= 1 as hairlines (to match Acrobat's behavior). Cache the last transform for PostScript-type functions. Added the "-rawdates" option to pdfinfo. Optimized ImageStream::getLine(). Fixed the Hue, Saturation, Color, and Luminosity blend functions to match Adobe's spec. Fixed the non-isolated group compositing computation. Skip extraneous unused data at the end of JBIG2 segments. Change the algorithm that stroke/fill adjustment uses so that the edges of adjacent strokes/fills line up. Do stroke adjustment on end caps when cap style is butt or projecting. Fixed a security hole: Gfx.parser was not being initialized to NULL. Fixed a security hole: integer bounds check in the Type 1 encoding parser in FoFiType1.cc. If an embedded font object is invalid or non-existent, do font substitution (same as if there were no embedded font). TextOutputDev was reusing an old font in the case where the font changed but the font size and character positioning stayed the same. When starting a transparency group, copy the fill/stroke colors from the graphics state. Tweaked the fixed-point code. When a TrueType font is declared resident (with a psFont command), don't munge the encoding. Look for URIs starting with "www." and treat them as absolute "http:" URIs, not as relative URIs (to match Adobe's behavior). Added code to FoFiTrueType to check for a zero-entry cmap table. Tweaked the font substitution code to do a better job of scaling the substituted font. Require at least two splits in the axial shading color bisection. Optimized JBIG2Stream::readGenericBitmap(). JPXStream wasn't correctly handling row padding for images with fewer than 8 bits per component. Optimized the ToUnicode CMap parser. Added a "whole words only" option to text searches. Check for valid component parameters in the DCT decoder. Implement embedding of external 16-bit fonts (without subsetting) in PostScript output. Added the minLineWidth xpdfrc command. Added warning messages for font substitutions.