$Header: /home/cvsroot/dvipdfmx/ChangeLog,v 1.290 2011/03/11 03:15:31 chofchof Exp $ ChangeLog: Changes for DVIPDFMx =============================== 2011-03-11 Jin-Hwan Cho * dpxcrypt.c: Applied the patch for 64bit systems; suggested by Akira Kakuto. 2011-03-09 Jin-Hwan Cho * dpxfile.c, configure.in, man/extractbb.1: Applied the patches for TeX Live 2011 by Peter Breitenlohner. 2011-03-07 Matthias Franz * dvi.c, dvi.h, dvipdfmx.c, pdfencrypt.c, pdfencrypt.h, spc_pdfm.c: Added special pdf:encrypt to turn on document encryption from within the DVI file (suggested by Michael Joyner). Format: \special{pdf:encrypt ownerpw (abc) userpw (xyz) length 128 perm 252} If a parameter is omitted, the default value / empty password is used. The special must appear at the beginning of the file. 2011-03-05 Jin-Hwan Cho * dpxfile.c: Applied the secutiry patch written by Akira Kakuto (suggested by Karl Berry). 2010-10-11 Matthias Franz * dvipdfmx.cfg: Added quotes around command line parameters %i and %o (suggested by Karl Berry). 2010-05-29 Matthias Franz * dpxfile.c, dpxfile.h, pdfobj.c, pdfximage.c: Further changes to fix the "%hhu" problem on Solaris 9. 2010-05-26 Matthias Franz * pdfobj.c: Changed type of PDF minor number from char to short in check_for_pdf_version in order to fix a compilation problem with Solaris 9; reported by Ulrich Seidl. 2010-05-24 Matthias Franz * dvipdfmx.cfg: Generate PDF 1.5 as default (suggested by Karl Berry). 2010-03-28 Jin-Hwan Cho * dvi.c, dvi.h, dvipdfmx.c, spc_pdfm.c: Implemented a new special \special{pdf: minorversion num} which will override the value given either by the configuration file or by the command line option; requested by Karl Berry. 2010-03-21 Jin-Hwan Cho * cidtype2.c: Assume that the output encoding of CMap files is UTF-16, not UCS2 as the recent Adobe-Japan1-UCS2 shows; reported by Haruhiko Okumura and the patch were prepared by Akira Kakuto. 2010-02-22 Matthias Franz * pngimage.c: Made PNG code work with libpng14, as suggested by Karl Berry. 2010-02-07 Matthias Franz * type1.c: Fixed bug caused by fonts without data for private dictionary. (The private dictionary was omitted instead of inserting an empty dictionary.) This caused problems with Acrobat Reader, for example with the dsrom10 font. Bug reported by Vladimir Vlasov. 2010-02-07 Jin-Hwan Cho * dvipdfmx.c, pdfdoc.c, pdfobj.h, pdfobj.c, pdfencrypt.h, pdfenctypr.c: Now the trailer always contains ID array to work Screen annotations under Adobe Reader 9. 2009-12-14 Matthias Franz * dvipdfmx.c: Turn on quiet mode earlier if "-q" option is given. 2009-11-28 Matthias Franz * spc_dvips.c: Fixed bug in "ps: plotfile" special, found by Alexander Bostroem. 2009-11-18 Matthias Franz * man/extractbb.1, man/Makefile.am, Makefile.am, configure.in: Added man page for extractbb in new directory man. This addresses Debian bug #512357 and a request by Karl Berry. 2009-10-10 Jin-Hwan Cho * dvipdfmx.c, xbb.c: Change usage() as suggested by Matthias Franz and Karl Berry. 2009-10-09 Jin-Hwan Cho * dvipdfmx.c: Implemented the usual options --version and --help (suggested by Karl Berry.) 2009-09-19 Matthias Franz * epdf.c, pdfdoc.c, pdfdoc.h, xbb.c: Bug fix: The number of pages in a PDF document as reported by extractbb was always 0. (The code for reporting the number of pages of a PDF file was accidentally removed during a previous modification.) 2009-09-18 Matthias Franz * dpxutil.c, dpxutil.h and all calls to the changed functions: Polished code for hash tables. Now the function for freeing values is defined with ht_init_table. ht_append_table now produces an error if the key is already in use (but this could be changed easily). A new insert mode where a value does _not_ replace an existing value for the same key has been introduced, but there is no short name for it yet. 2009-08-30 Matthias Franz * dvipdfmx.cfg: Turned on optimization of PDF destinations by default. 2009-08-27 Matthias Franz * truetype.c, type1.c, type1c.c: Changed precision in font widths arrays from 1.0 to 0.1. (Applies only to Type 1, Type 1C and TrueType, not PK fonts.) 2009-08-23 Matthias Franz * epdf.c, pdfdoc.c, pdfdoc.h, pdfximage.c, pdfximage.h, xbb.c: Removed the feature of negative page numbers for PDF inclusion. It was never used and made the code more complicated. * dvipdfmx.c, pdfximage.c: Changed the code for PS and PDF page inclusion: The function mkstemp on NetBSD may return the same name twice if the file has been deleted after the first call. (Reported by Jukka Salmi.) This interfered with our file name caching. Now we don't delete the temporary file after PS->PDF conversion, but cache the temporary file name instead. An advantage is that we don't have to convert the same file multiple times if several pages are imported. 2009-08-17 Jin-Hwan Cho * tt_post.c: Fixed the crashing bug with Mona font found by Alexis Ballier. 2009-07-08 Jin-Hwan Cho * dvipdfmx.c: Add -E option that enables DVIPDFM emulation mode in DVIPDFMx. 2009-07-07 Jin-Hwan Cho * pdfximage.c, pdfximage.h, spc_pdfm.c: Add pdf_ximage_set_attr() and pdf_ximage_get_subtype() to resolve the conflict with bmpsize package in dvipdfm emulation mode; reported by Heiko Oberdiek. 2009-05-22 Jin-Hwan Cho * generic/pgf-doc-diff.version2cvs, generic/pgf-generic-diff.version2cvs: Removed the patches for the CVS version of pgf, because they are applied to the main stream. 2009-05-10 Matthias Franz * epdf.c: Don't try to convert included PDF files whose version number is higher than output file version. (If run as dvipdfm, conversion still occurs.) * Makefile.am, Makefile.in, dvipdfmx.c, dvipdfmx.h, jpegimage.c, pngimage.c, xbb.c, xbb.h: Added dvipdfm compatibility mode. This mode is triggered if dvipdfmx is called as "dvipdfm", and likewise extractbb as "ebb". (The latter mode was already present.) * dpxconf.c: Made predefined paper sizes match values given by libpaper. 2009-05-06 Jin-Hwan Cho * dvipdfmx.c, pdflimits.h, pdfobj.c: Improved PDF version handling. * cmap_read.c: Fixed a silly bug; reported by Fumihito YOSHIDA. 2009-05-06 Matthias Franz * cid.c, dvipdfmx.c, pdflimits.h, pdfobj.c: Fixed minor issues with PDF version handling, following a question by Akira Kakuto. The minimal and maximal PDF versions allowed are now defined in pdflimits.h. 2009-05-03 Matthias Franz * dvipdfmx.c, epdf.c, epdf.h, pdfdoc.c, pdfdoc.h, pdfobj.c, pdfobj.h, pdfximage.c, pdfximage.h, xbb.c: Improved PDF version handling. A "Version" entry in the document catalog is now honored. Moreover, dvipdfmx accepts to write PDF 1.7, and extractbb doesn't complain about PDF versions any more. Encryption keys too long for PDF 1.3 now lead to an error instead of a silent change of PDF version. 2009-05-02 Jin-Hwan Cho * tt_table.c: Fixed a bug of reading numOfExSideBearings; reported by Tomas Janousek. 2009-04-30 Matthias Franz * dvipdfmx.cfg: Use ps2pdf for PS-to-PDF conversion on all platforms. 2009-04-29 Jin-Hwan Cho * spc_color.c, spc_pdfm.c, spc_util.c, spc_util.h: Allow changing strokecolor or fillcolor only; for example, "pdf:scolor Red []" changes fillcolor only and "pdf:scolor [] Red" changes strokecolor only, but "pdf:scolor Red" changes both fillcolor and strokecolor. * pdfparse.c: Parse stream objects in a strict way as DVIPDFM because DVIPDFMx already has another specials for stream objects. 2009-04-26 Matthias Franz * pdfobj.c: Don't print unnecessary newlines in PDF dictionaries. * pdfobj.c, pdfobj.h, xbb.c: Added "%%HiResBoundingBox" and "%%PDFVersion" lines to extractbb output. * dpxutil.c, dpxutil.h, dvipdfmx.c, pdfdoc.c,pdfdoc.h, pdfnames.c, pdfnames.h, pdfobj.c, pdfobj.h, spc_pdfm.c, specials.c, dvipdfmx.cfg: Rename PDF destinations to shorter names (in order to save space). Together with the removal of unused PDF destinations implemented previously, this is more or less equivalent to Heiko Oberdiek's LaTeX package "hypdestopt". 2009-04-20 Jin-Hwan Cho * pdfdoc.c: Bug fix: \special{pdf:bann} ... \special{pdf:eann} (e.g., \hyperlink) cause a problem when it was surrounded by the pair \special{pdf:bcontent} ... \special{\special{pdf:econtent}. 2009-04-19 Jin-Hwan Cho * generic/pgf-doc-diff.texlive2008, generic/pgf-generic-diff.texlive2008: Patches and a new dvipdfmx driver for PGF, Version 2.00 included in texlive 2008. * generic/pgfsys-dvipdfmx.def.version2, generic/pgfsys-dvipdfmx-patch.tex.version2: Removed contributed files because of new patch files pgf-doc-diff.texlive2008 and pgf-generic-diff.texlive2008. 2009-04-18 Jin-Hwan Cho * generic/pgf-doc-diff.version2cvs, generic/pgf-generic-diff.version2cvs: Patches and a new dvipdfmx driver for PGF, Version 2.00-cvs 20090418. * contrib/context/spec-dpx.tex, contrib/latex/animate.sty.sample, color.sty.sample, dvipdfmx.def.sample, graphics.sty.sample, movie15.sty.sample, sample-imagemask.tex: Removed contributed files already applied to texlive 2008. 2009-04-08 Jin-Hwan Cho * t1_load.c: Ignores the second /Subrs dicts and /CharStrings not followed by dict; Adobe's OPO_____.PFB and OPBO____.PFB had those problems. 2009-04-07 Matthias Franz * xbb.c: Removed binary mode option "-b". In ebb, just ignore it. 2009-03-30 Jin-Hwan Cho * pdfdoc.c: Fixed the bug: bbox->llx is not extracted for PDF images. 2009-03-29 Matthias Franz * xbb.c: Changed output-to-stdout option to "-O". 2009-03-24 Jin-Hwan Cho * cff.c, dpxutil.c, pdfencoding.c, pdfencrypt.c, mem.c: Revive system.h to work with TeXLive. 2009-03-23 Matthias Franz * dvipdfmx.c, jpegimage.c, jpegimage.h, pngimage.c, pngimage.h, xbb.c: Several changes related to xbb/extractbb: Combined code with jpegimage.c and pngimage.c, added number of pages to xbb output for PDF files, introduced quiet mode (option "-q") and output to stdout (option "-s"), improved test of filename check in dvipdfmx.c (now uses >libgen.h>), new version numbering for xbb (same as dvipdfmx version) 2009-03-15 Jin-Hwan Cho * data/Makefile.am, data/Makefile.in, data/EUC-UCS2: Revive the cmap EUC-UCS2; suggested by Nobuyuki Tsuchimura. 2009-03-16 Matthias Franz * pdfdev.c, pdfdraw.c, spc_html.c, spc_pdfm.c, spc_tpic.c: Added comments to show the PDF operators inserted by each call of pdf_doc_add_page_content. This makes it easier to find out which part of the source is responsible for a PDF operator in the output. Example: pdf_doc_add_page_content(work_buffer, len); /* op: q cm * 2009-03-12 Matthias Franz * dvipdfmx.c: Don't print messages in quiet mode. * dpxfile.c, dpxfile.h, pdfximage.c, data/dvipdfmx.cfg: Added conversion specifier "%v" (PDF version) to PS-to-PDF conversion template. Also changed default template for GhostScript to use the script ps2pdf instead of calling gs directly. (This works for all platforms.) * pdfobj.c: Fixed a bug concerning empty PDF strings: Empty PDF strings were reduced to NULL, which could lead to errors when calling string functions. 2009-01-15 Jin-Hwan Cho * specials.c: Apply Jonathan Kew's fix reported on August 7, 2008. (Sorry that it was too late!) 2008-12-11 Matthias Franz * xbb.c: Fixed an off-by-one and an off-by-two error in the generation of output file names for extractbb. (Debian bug #506329, reported by G\'abor Braun.) * mpost.c, pdfcolor.c, pdfcolor.h, pdfdev.c, pdfdev.h, pdfdoc.c, pdfdraw.c, pdfdraw.h, spc_color.c, spc_pdfm.c: Polished the code for color support. PDF color operators will not be generated any more if old and new color coincide. 2008-11-30 Matthias Franz * spc_tpic.c: Polished code. * pdfdev.c, spc_html.c: Concatenated transformation matrices for images. * data/dvipdfmx.cfg, dvipdfmx.c, pdfdoc.c: Fixed two problems with removal of PDF destinations: Non-link annotations produced a warning, and broken links were not detected if no valid destination existed. Removing unused destinations is now default behaviour and can be prevented by using the flag "-C 0x10". Since the feature is still new, this flag is set in dvipdfmx.cfg. * spc_color.c, spc_color.h, spc_dvips.c, spc_dvips.h, spc_misc.c, spc_misc.h, spc_pdfm.c, spc_pdfm.h, specials.c: Removed unnecessary (no-op) handlers for specials. * cid.c, cid.h, dpxconf.c, dpxconf.h, dvi.c, dvi.h, fontmap.c, fontmap.h, numbers.c, numbers.h, pdfcolor.c, pdfcolor.h, pdfdev.c, pdfdev.h, pdfdoc.c, pdfdoc.h, pdfdraw.c, pdfdraw.h, pdffont.c, pdffont.h, pdfobj.c, pdfobj.h, pdfresource.c, pdfresource.h, pst_obj.c, pst_obj.h, tfm.c, tfm.h, tt_table.c, tt_table.h, type0.c, type0.h: Commented out unused functions. 2008-11-08 Matthias Franz * pdfcolor.c, pdfcolor.h, spc_tpic.c: Made tpic specials respect color settings. Before, they always used black for drawing commands. (Debian bug #485567, reported by Daniel Skorka.) 2008-11-03 Matthias Franz * dpxutil.c, dvipdfmx.c, mem.c, pdfdoc.c, pdfdoc.h, pdfnames.c, pdfnames.h: Remove unused PDF destinations. (Packages like hyperref create destinations regardless of whether they are used or not.) This option is currently turned off by default; it can be enabled with "-C 0x10". * dvipdfmx.c: Improved parsing of "-v" command line option; contributed by Akira Kakuto. Note that there are still problems if an argument to an option (or the name of the DVI file) starts with a minus sign * pkfont.c: Corrected bounding box computation for PK fonts. Before, it was always [0, 0, 0, 0]. Also removed a redundant call to pdf_encoding_add_usedchars. 2008-10-13 Matthias Franz * cff.c, cff.h, cff_dict.c, cidtype0.c, pdfdev.c, t1_load.c, type1.c, type1c.c: Various speed improvements: cff.c: cff_add_string now has an option to avoid checking for duplicate strings. pdfdev.c: p_dtoa does not call library functions any more. t1_load.c: t1_decrypt is not called for each byte seperately any more. (Well, this does not make a big difference.) 2008-09-29 Matthias Franz * epdf.c, pdfdoc.c, pdfdoc.h, pdfobj.c, pdfobj.h, xbb.c: Made PDF parsing more robust. In particular, we can now detect (illegal) loops in a PDF document hierarchy. Also, much of the PDF parsing code in epdf.c and xbb.c has been combined to the function pdf_doc_get_page in pdfdoc.c. Loops in PDF page trees are detected as well. Moreover, the various bounding boxes of a page are now handled correctly: MediaBox and CropBox are inherited, but not the others. 2008-08-31 Matthias Franz * type1.c: Fixed a bug when processing "seac" operators in Type 1 fonts. If the base or accent char were not used otherwise, an "Invalid SID" error could occur. 2008-08-05 Matthias Franz * type1c.c: Fixed a buffer overflow when using 256 glyphs of a CFF font (plus ".notdef"). 2008-08-04 Matthias Franz * type1c.c: Fixed a segfault when processing CFF fonts. 2008-07-16 Matthias Franz * pdfnames.c, pdfobj.c, pdfobj.h: Replaced pdf_copy_object by the (almost) equivalent, but simpler function pdf_transfer_label. Simplified pdf_compare_reference. * pdfnames.c: Reintroduced warning about duplicate object keys. 2008-06-20 Jin-Hwan Cho * spc_pdfm.c: Fixed a bug on beads with nonzero depth; reported by Alexander Grahn. 2008-06-13 Jin-Hwan Cho * spc_pdfm.c: Fixed a bug on annotations with nonzero depth; reported by Alexander Grahn. 2008-06-07 Jin-Hwan Cho * contrib/latex/movie15.sty.sample: Support movie15.sty, version 2008/01/16. * pdfobj.c, pdfobj.h, pdfparse.c, spc_pdfm.c, pdfnames.c: Allow PDF stream of length zero and suppress some warning messages. 2008-06-06 Jin-Hwan Cho * contrib/latex/animate.sty.sample: Support animate.sty, version 2008/06/03. 2008-06-05 Jin-Hwan Cho * mpost.c, pdfdev.c, pdfdev.h, pdfdoc.c, pdfdoc.h, spc_pdfm.c: DVIPDFMx supports PGF version 2.00. New specials "pdf:bcontent", "pdf:econtent", and "pdf:code". Moreover, "pdf:exobj << additional_dict >>" merges additional dictionaries into the XObject dictionary, not resources. * pgfsys-dvipdfmx.def.version2, pgfsys-dvipdfmx-patch.tex.version2: A sample DVIPDFMx driver for PGF version 2.00. Note that pgfsys-dvipdfmx-patch.tex must be loaded after loading the basiclayers of PGF. Otherwise image mask and transparency group do not work. 2008-06-01 Jin-Hwan Cho * pdfdev.h, spc_pdfm.c, spc_util.c, dvipdfmx.def.sample: Changed the option "mask" to "hide" in "pdf:image" special. New "pdf:stream" special for a string given externally. * pgf-generic.diff, pgfsys-dvipdfmx.def, xcolor.sty.sample: Renamed and moved pgf patches for version 1 to contrib/generic. Removed "xcolor.sty.sample". 2008-05-29 Jin-Hwan Cho * spc_dvips.c, pdfdoc.c, spc_html.c, spc_misc.c, pdfximage.c, pdfximage.h, spc_pdfm.c, spc_util.c, pdfdev.h: Support PGF image masking; support additional attribute dictionary for included images as pdfTeX. * xcolor.sty.sample, pgf-generic.diff, pgfsys-dvipdfmx.def, sample-imagemask.tex, dvipdfmx.def.sample: Sample files for image masking and xcolor, pgf patches. 2008-05-28 Jin-Hwan Cho * spc_pdfm.c: Remove an warning message which occurs with the PGF transparent model. * spc_util.c: Support dvips color names defined in dvipsname.def with "pdf:bcolor" special, e.g. "pdf: bcolor JungleGreen". 2008-05-23 Jin-Hwan Cho * data/Makefile.am, data/Makefile.in: Removed the obsolete CMap files in the data directory. * data/Makefile.in, src/Makefile.am, src/Makefile.in: Apply Karl Berry's symlink patch. 2008-05-22 Matthias Franz * cff_dict.c, dpxconf.c, dpxfile.c, dvipdfmx.c, epdf.c, fontmap.c, pdfdoc.c, pdfdraw.h, pdfencoding.c, pdfobj.c, pst_obj.c, sfnt.c, spc_pdfm.c, xbb.c, truetype.c, pdfencoding.c, pdfdraw.c: Fixed some compiler warnings issued by "gcc -Wall". 2008-05-21 Jin-Hwan Cho * pdfdoc.c: Fix the timezone bug in asn_data() based on the dvipdfm-0.13.2d code; this patch was suggested by Nobuyuki Tsuchimura. 2008-05-20 Matthias Franz * dvipdfmx.c, pdfdoc.c, pdfdoc.h, pdfobj.c, pdfobj.h: Polished object stream code; removed command line option flag 0x10 (no object streams). 2008-05-20 Jin-Hwan Cho * contrib/latex/dvipdfmx.def.sample: Fix ithe problem when including a PDF image with bounding box having nonzero llx or nonzero lly. 2008-05-18 Jin-Hwan Cho * pdffont.c, pdffont.h, type1.c, type1c.c, truetype.c: Fix the problem with MusiXTeX's Type1 fonts. Some characters of these fonts have different widths from the corresponding TFM files; reported by Yasuhiro Okazaki and Jean-Pierre Coulon. DVIPDFMx now reads character width from TFM if found. Otherwise, the information comes from the font file. However, Type0 fonts (not simple fonts) do not refer to TFM. 2008-05-18 Matthias Franz * epdf.c, pdfobj.c, pdfobj.h, xbb.c: Added PDF 1.5 parsing. Note that many PDF 1.5 files have compressed xref streams with DecodeParms entry. Since we don't support this, we cannot read these files. * epdf.c: Allow page import from tagged PDF files. Tags are ignored. 2008-05-18 Jin-Hwan Cho * dvipdfmx.c, dpxfile.c, dpxfile.h, otl_conf.c, spc_pdfm.c, system.h: Arranged MiKTeX related codes according to MiKTeX SVN repository. * fontmap.c, fontmap.h, pdfdev.c, pdffont.c: Ignore obsolete remap option -r in DVIPDFM fontmaps. 2008-05-17 Jin-Hwan Cho * dvipdfmx.c, xbb.c, xbb.h, src/Makefile.am, src/Makefile.in, contrib/latex/dvipdfmx.def.sample: DVIPDFMx now extracts bounding box information. The binary dvipdfmx can be linked as extractbb, ebb, xbb, and it works as same as the linked name. (extractbb == xbb) This idea was suggested by Karl Berry. 2008-05-17 Jin-Hwan Cho * Makefile.in, aclocal.m4: Remove acinclude.m4. Give a new tag DVIPDFMX-20080517. * cid.c, cidtype0.c, cid_p.h, truetype.c, cidtype2.c, fontmap.c, fontmap.h, tt_aux.c, tt_aux.h, type1c.c: New fontmap option '-v' to give StemV value explicitly; especially for Chinese truetype fonts, e.g., Simsun; this feature was suggested by Yue Wang. * dvipdfmx.def: Support automatic generation of .xbb if -shell-escape is given. Don't forget to replace $TEXMF/tex/latex/dvipdfmx/dvipdfmx.def to contrib/latex/dvipdfmx.def.sample. * mpost.c: Support MPS files generated by MetaFun. * pdfximage.c: Try distiller specified in the configuration file if interpreting MPS file fails. 2008-05-16 Matthias Franz * pdfobj.c: Fixed incorrect size entry in xref stream dictionary. * pdfparse.c: Fixed a bug in hex string parsing. 2008-05-16 Jin-Hwan Cho * dvipdfmx.c, spc_pdfm.c, dpxfile.c, otl_conf.c: Restore the application name from "dvipdfm" to "dvidpfmx". Notice that "dvipdfmx.cfg" must be located in "$TEXMF/dvipdfmx". * dvipdfmx.c, fontmap.c, fontmap.h, spc_pdfm.c: Support font mapping files with DVIPS/pdfTeX format. 2008-05-13 Matthias Franz * cid.c, dvipdfmx.c, epdf.c, mpost.c, pdfencoding.c, pdfobj.c, pdfobj.h, pdfparse.c, pdfparse.h, pdfximage.c, pdfximage.h, spc_pdfm.c, xbb.c: Changed code for PDF import so that objects are not duplicated if they are imported through several pages of the same file. Some PDF parsing functions now take an additional argument of type pdf_file. If this is NULL, them indirect references are not accepted. 2008-05-10 Jin-Hwan Cho * dpxfile.c, pdffont.c: Support Mac OS X Resource Fork Font (/rsrc) with TrueType outlines. 2008-05-09 Jin-Hwan Cho * truetype.c, cidtype0.c, cidtype2.c, pdffont.c, pdffont.h, sfnt.c, sfnt.h, tt_glyf.c, tt_cmap.c, tt_table.c, tt_aux.c, dpxfile.c, dpxfile.h: Support Mac OS X Data Fork Font (.dfont) with TrueType outlines. 2008-05-08 Jin-Hwan Cho * cidtype0.c, cidtype2.c, tt_aux.c, tt_table.c: Support Mac TrueType fonts without OS/2 table. 2008-04-28 Matthias Franz * src/pdfencoding.c, src/pdfobj.c: No declarations after statements allowed in C89. (contributed by Vladimir Volovich via Karl Berry) * src/Makefile.am: xbb needs zlib too. (contributed by Karl Berry ) * m4/maw-kpse.m4, maw-libs.m4 (removed): Renamed ... * m4/chi-kpse.m4, cho-libs.m4 (new): ... into these. * Makefile.am: Added `ACLOCAL_AMFLAGS -I m4'. * configure.in: Use new names; respect library dependencies; moved macro definitions ... * m4/maw-time.m4 (new): ... to here. * acinclude.m4 (removed): Moved macro definitions ... * m4/maw-kpse.m4, maw-libs.m4 (both new): ... to here. (all contributed by Peter Breitenlohner via Karl Berry) 2008-03-01 Matthias Franz * dvipdfmx.def: Write bounding box information only if necessary. * dvipdfmx.def, spc_util.c: New 'clip' parameter for 'pdf:epdf' and 'pdf:image' specials. 'clip 1' enables clipping (default), 'clip 0' disables it. 2008-02-13 Matthias Franz * dvipdfmx.def, epdf.c, pdfdoc.c, pdfximage.c, pdfximage.h, spc_dvips.c, spc_html.c, spc_misc.c, spc_pdfm.c, spc_util.c, spc_util.h: Added new 'page' parameter to PDF and PS inclusion. This can be used with the graphicx package and the (experimental) graphics driver dvipdfmx.def (see contrib/latex). Example: \includegraphics[page=4]{a.pdf} Page numbers can also be negative; the last page is -1. Remarks: 1) Currently PDF objects may be duplicated in the output if several pages are included from the same file. -- done (2008-05-13) 2) The graphicx package currently uses the same bounding box information (that of the first page) for all pages. 2008-02-13 Jin-Hwan Cho * acinclude.m4, aclocal.m4, configure.in, configure: Karl Berry reports configure fixes for TeXLive. * pdfencoding.c, pdffont.c: Fix an encoding bug by the MinionPro package; reported by Juho Lee. 2008-02-12 Matthias Franz * pdfencrypt.c: Encrypted PDF files with key lengths strictly between 40 and 128 bits could not be opened. Changed code to meet Adobe Reader's behaviour for revision 3 (PDF 1.4) encryption. 2008-02-08 Matthias Franz * epdf.c, pdfobj.c: Improved PDF document parsing: pages without contents, undefined indirect references and media boxes with indirect references are now allowed. (Undefined indirect references are needed for documents with several cross-reference tables if the trailer contains a reference to an object outside the part of the xref table seen so far.) 2008-01-11 Matthias Franz * dvipdfmx.c: Fixed a bug in processing DVI comments. * cmap_read.c, pst.c, pst_obj.c , t1_load.c: Fixed a bug in postscript number parsing; reported by Ulrike Fischer. 2008-01-06 Matthias Franz * type1.c, type1c.c: Fixed a bug in the computation of glyph width arrays for Type1 and Type1C font dictionaries; reported by Ulrike Fischer. 2007-12-05 Jin-Hwan Cho * spc_pdfm.c: Fixed a memory deallocation problem in the ConTeXt sample of Francesco Pretto. 2007-12-03 Jin-Hwan Cho * cmap.c: Code mappings in CMap may overlap, but succeeding maps superceded preceding maps; reported and patched by Luo Jie. 2007-11-27 Jin-Hwan Cho * pdfdev.c: Restored the DVIPDFM code for calculation of kerns; reported by Keiichirou Shikano. 2007-11-22 Jin-Hwan Cho * pdfdev.c: Fixed a bug related to calculate kern; reported by Keiichirou Shikano. * spc_util.c: Fixed a memory deallocation bug. 2007-11-18 Matthias Franz * vf.c: Fixed small memory leak. 2007-11-17 Matthias Franz * pdfobj.c: Suppress unnecessary white-space in PDF arrays. * dpxfile.c, dvipdfmx.c, mem.c, mem.h, tfm.c, xbb.c: Fixed some small memory leaks. 2007-11-15 Jin-Hwan Cho * cid_basefont.h: Add a predefined simplified Chinese font, AdobeHeitiStd-Regular.otf. * pdffont.c: Fixed a bug causing problems with Type0 fonts; reported by Akira Kakuto. 2007-11-13 Jin-Hwan Cho * dpxutil.c, dpxutil.h, dvipdfmx.c, mem.c, mem.h, fontmap.c, pdfdoc.c, pdfdoc.h, pdfobj.c, pdfobj.h, subfont.c, subfont.h, vf.c, dpxfile.c, xbb.c: Applied the memory leak patch by Matthias Franz. * pdfencoding.c, pdfencoding.h, pdffont.c, type1.c, type1c.c, truetype.c, pkfont.c: Applied the encoding patch by Matthias Franz. * agl.c, agl.h: Introduced new predefined glyphlist.txt; contributed by Matthias Franz. 2007-11-03 Jin-Hwan Cho * mpost.c: Fixed a bug in skip_prolog(), which is unable to skip prolog. Now work with "prologues:=3" mode in metapost 1.000. 2007-07-03 Jin-Hwan Cho * cff_dict.c: Fixed a bug in pack_real() using work_buffer twice; found by Takanori Uchiyama and reported by Akira Kakuto. 2007-06-30 Jin-Hwan Cho * cidtype0.c, cff_dict.h: Fixed a bug checking cid >= CIDCount, which causes a problem with some CJK otf fonts without having the CIDCount value. 2007-06-28 Jin-Hwan Cho * cff_dict.c: Modified pack_real() to support musix20.pfb in Mac OS X 10.4 Quartz. It seems that Mac OS X 10.4 Quartz has a bug on reading real numbers in CFF font format. Since Mac OS X 10.4.10 the problem was solved. 2007-05-18 Jin-Hwan Cho * jpegimage.c, pngimage.c, pdfximage.c, pdfximage.h, dvipdfmx.c, xbb.c: Updated "xbb" to generate ".xbb" not ".bb". There is a big difference between "xbb" and "ebb" in handling bitmap images. In the case of bitmap images not having density information, there is no way to know the density (or DPI) of the image. "ebb" always assume that the image has 100 DPI so that 100 pixels = 1 in = 72 bp. However, in this case, screen capture image does not look clear. Thus, "xbb" follows pdfTeX, which uses density information if exists, otherwise 72 DPI, i.e., 72 pixels = 1 in = 72 bp, is used. With the modified "dvipdfmx.def" driver for the LaTeX graphics package, you can get the same result as pdfTeX. 2007-05-04 Jin-Hwan Cho * xbb.c: New program "xbb" extending "ebb". 2007-05-03 Jin-Hwan Cho * pdfximage.c, pdfximage.h, dvipdfmx.c: Support bbox for "pdf:image" special (experimental). The program "ebb" extracting the bounding box of image files always multiply 72/100 for bitmap images (png and jpg). It causes some trouble with LaTeX graphicx package which read the bounding box from ".bb" file usually generated by "ebb". New experimantal option "-C 0x0020" is introduced. With this option, dvipdfmx no longer multiply 72/100 for bitmap images so that users can get the same result as pdftex. 2007-04-25 Jin-Hwan Cho * spc_pdfm.c: Extended pdf:outline special to support open/close bookmark. The previous version 20070409 had problem because toclevel of part is -1 in some class files. 2007-04-24 Jin-Hwan Cho * spc_util.c: Predefined colors, Black, White, and Gray caused some problem in Adobe Reader 7 and 8; but not in other PDF viewers, e.g., Previewer.app. It was solved by changing cmyk value to gray value. 2007-04-17 Jin-Hwan Cho * dvipdfmx.c, pdfobj.c: The default of the PDF minor version was changed from 3 to 4. * type1.c: A bug was reported from Werner Lemberg (run nfssfont.tex for the font ptmr8y). It happened when any character is used more than one in one encoding file (e.g., texnansi.enc). To fix it, Encoding resource was revived (but Hirata noted that there was some problem with Ghostscript, not confirmed yet) and Widths resource was changed to look up correct glyphs according to the given encoding. * vf.c: Suppress an annoying warning message if no verbose option is given. 2007-04-13 Jin-Hwan Cho * tt_cmap.c: Support OpenType with PID=0, EID=3 2007-04-09 Jin-Hwan Cho * dvipdfmx.c: Updated the usage as suggested by Matthias Franz 2007-04-03 Jin-Hwan Cho * pdfobj.c, pdfdoc.h, pdfencrypt.c, pdfencrypt.h, pdfobj.c, pdfobj.h, pdfparse.c, dvipdfmx.c: Support object stream for PDF v1.5 or higher; contributed by Matthias Franz. Previewer.app in Mac OS X did not work with encryption under v1.5, but no problem with Adobe Reader 8 2007-03-26 Jin-Hwan Cho * pdfobj.c: Support xref stream for PDF v1.5 or higher; contributed by Matthias Franz 2007-03-17 Jin-Hwan Cho * dpxfile.c: Fixed a bug in reading .enc file; reported by Matthias Franz 2007-02-28 Jin-Hwan Cho * epdf.c: Fixed a problem in handling Cropbox 2007-02-14 Jin-Hwan Cho * cidtype0.c tt_glyf.c tt_table.h, tt_table.c: Fixed a bug in reading htmx and vtmx tables; reported by Jjgod Jiang based on the solution by Jonathan Kew 2007-01-21 Jin-Hwan Cho * fontmap.c: Fixed typos found by Jonathan Kew 2007-01-20 Jin-Hwan Cho * specials.h: Fixed warning message in the IRIX compilation, reported by Karl Berry and found by Olaf Weber 2007-01-19 Jin-Hwan Cho * fontmap.c: Fixed a bug reported by Shigeru Miyata on Dec 28, 2005 * spc_html.c: Fixed a bug reported by Nobuyuki Tsuchimura on Jan 19, 2007 2006-12-20 Jin-Hwan Cho * configure: Remake using autoconf-2.59 * configure.in config.h.in src/pdfdoc.c: Apply tm_gmtoff related modifications by Karl Berry 2006-12-15 Jin-Hwan Cho * configure aclocal.m4 Makefile.in src/Makefile.in data/Makefile.in: Remake using aclocal-1.6, automake-1.6, and autoconf-2.59 * configure.in src/Makefile.am data/Makefile.am: Apply configure related modifications by Karl Berry 2006-12-11 Jin-Hwan Cho * mpost.c, pdfdev.c, pdfdev.h, spc_pdfm.c, pdfcolor.c, pdfcolor.h, pdfdraw.c, pdfdraw.h: Temporary implementation for mps files generated by mptopdf. 2006-12-11 Jin-Hwan Cho * pdfcolor.c: Fixed the color bug reported by H. J. Lee and Pablo Rodriguez. 2006-12-07 Jin-Hwan Cho * pdfobj.c: Fixed a bug with xref, reported by Paul Guyot. 2006-12-06 Jin-Hwan Cho * t1_load.c: Modified to work with broken fonts, e.g. belleek/blsy.pfb. * dvipdfmx.c, dvi.c, dvi.h: Add xoffset and yoffset specials. 2005-12-18 Jin-Hwan Cho * dpxfile.c: Fixed a bug found by Karnes Kim. 2005-10-14 Jin-Hwan Cho * spc_tpic.c: Fix a bug on spc_handler_tpic_bk() and spc_handler_tpic_wh(); reported by Akira Kakuto. 2005-08-31 Jin-Hwan Cho * pdfdoc.c, pdfdraw.c: Stroke and fill color controls with gsave and grestore. 2005-08-30 Jin-Hwan Cho * pdfdraw.h, pdfdraw.c, mpost.c: Fix a bug on MetaPost manipulation. * pdfobj.c: Allow PDF images generated from Mac OS X. * truetype.c: Fix a bug on PK Type3 font. 2005-08-24 Jin-Hwan Cho * spc_pdfm.c: Fix a bug on finding XObject. 2005-08-23 Jin-Hwan Cho * pdfdraw.h, pdfdraw.c: Fix a bug on matrix concatenation and inverse matrix. * spc_pdfm.c: Fix a bug on giving object name in image. 2005-08-15 Jin-Hwan Cho * pdfdoc.c, spc_pdfm.c, specials.c: Fix a bug on article|bead|thread specials. 2005-08-14 Jin-Hwan Cho * spc_html.c, spc_pdfm.c, specials.c: Fix a bug causing memory leak, and correct a typo. 2005-08-12 Jin-Hwan Cho * pdfdev.c, pdfdev.h, spc_dvips.c, spc_pdfm.c, spc_util.c: Modified transform_info structure using pdf_tmatrix. 2005-08-11 Jin-Hwan Cho * pdfdoc.c, pdfdraw.c, pdfdraw.h, spc_pdfm.c: Optimization of changing transform matrices. The "pdf:literal" special works almost same as pdfTeX's "\pdfliteral" primitive. But "pdf:literal direct" is slightly different from pdfTeX's "\pdfliteral direct". It does not affected by "pdf:literal". 2005-08-08 Jin-Hwan Cho * spc_pdfm.c: Test implementation of \special{pdf:literal ...}. * spc_misc.c: Ignores \special{pos: ...} too. * spec-dpx.tex: Modified to fit the changes introduced in 2005-08-08. 2005-08-02 Jin-Hwan Cho * dvi.c, spc_pdfm.c: Fixed a few bugs introduced by changes in 2005-08-01. * dvipdfmx.def, spec-dpx.tex: Modified to fit the changes introduced in 2005-08-01. ConTeXt module still has a bug, e.g., \handleMPfshow. 2005-08-01 Shunsaku Hirata * specials.c, spc_html.c, spc_pdfm.c: Various minor changes to specials. Rewriting HTML special support. 2005-07-31 Shunsaku Hirata * specials.c: Fixed a bug introduced by changes in 2005-07-30. (@page1 not handled correctly) 2005-07-30 Shunsaku Hirata * colors.h: removed Color definitions placed into spc_util.c. * pdfdev.c, pdfdraw: Various changes. * pdfcolor, pdfdoc: set_bgcolor pdfcolor --> pdfdoc. * cid.c: Force fixedpitch option for printer/acro fonts. * spc_*, specials.c: * And many others.... 2005-07-28 Shunsaku Hirata * pdfdev.c: Simply use sprintf for printing numbers................. Fixed an ancient bug causing very frequent (unnecessary) font selection. 2005-07-21 Shunsaku Hirata * pdfencoding.c: Fixed bugs. * truetype.c: Update and various changes. Apply ligature GSUB for glyph name "a_b" and glyph names with AGL mapping to multiple Unicode character sequence. Suffixes "liga", "dlig", "frac" ... are always recognized as feature tag for glyph composition. But for "a_b.vert", dvipdfmx first compose glyph "a_b" (by some way) and then apply "vert" substitution to composed glyph. It may often behaves unexpectedly. Suffixes like "aalt1" is treated as alternate substitution; e.g., "a.salt1" is interpreted as first stylistic alternate of default glyph for "a". Those features are just for completeness. (esp. alternate GSUB) Reorganized various process of finding glyphs, but still complicated. TTC is supported (but index is always 0). 2005-07-20 Shunsaku Hirata * pdffont.c: Minor changes. * pkfont.c: Mostly rewritten. Allow .enc file to be specified in fontmap. But it is not for reencoding but to get glyph names for ToUnicode creation. * dpxfile.c: Check file type. * pdfcolor.c, vf.c: Don't use strcasecmp, strncmp. 2005-07-18 Shunsaku Hirata * pdfximage.c: Set image format conversion filter to NULL if empty string "" is given. Put various system dependent things into dpxfile.c. * dpxfile.c: Creation of temporary files, call other program via system() to apply filter to file. * dpxconf.c, dpxconf.h: new files Currently paper format related codes. * dvipdfmx.c: cleanup Papersize handling --> dpxconf.c Removed undocumented extension for page label. * dvipdfmx.c: 's' option for MPS-to-PDF conversion mode. Treat page numbers as suffix. For "dvipdfmx -M -s 1-3 foo", dvipdfmx first tries to open "foo". If "foo" exists, it then start conversion of "foo" (ignore 's'), if not, "foo.1", "foo.2", and "foo.3" is searched and multi-page PDF file is generated. 2005-07-17 Shunsaku Hirata * spc_dvips.c: Fixes in psfile special. 2005-07-17 Shunsaku Hirata * dpxutil.c, dpxutil.h: Added few functions to make fontmap.c independent from pdfparse.c. Various parse_xxx from pdfparse.c. mangle_name() removed. * pdfparse.c, pdfparse.h, dpxutil.c, dpxutil.h, ...: Several parsing routine moved to dpxutil.c. parse_c_ident() is not really for "C" ident. parse_c_string() is not really for "C" string. And there are too many unclear parse_xxx routines unrelated to PDF in "pdfparse". There are many uncertainty in parsing \special too. * spc_xxx: skip_white() wrongly used in many places. skip_white() skips from '%' to newline... * pdf:tounicode, pdf:mapfile, and pdf:mapline using very different syntax than other pdf: prefix specials. * dvi.c, pdffont.c: Minor changes in debug and error message. * psimage.c, psimage.h, t1crypt.c, t1crypt.h: Removed. * t1_load.c: Merge t1crypt * pdfximage.c, pdfximage.h: Merge psimage psimage.c is not really for PostScript support. It just call external program to do conversion form PS to PDF. There are no reason to restrict this to from-PS conversion. * fontmap.c: Do not use pdfparse.c. Added 'u' option to specify "ToUnicode". This is for testing. Don't use this. * pdffont.c: Read ToUnicode CMap specified in fontmap. * spc_util.c: HSB color... Converted to RGB. * cidtype0.c, cidtype2.c: Do not use mangle_name in dpxutil.c. * Various fixes related to pdflimits.h: This only contains PDF_NAME_LEN_MAX. 2005-07-16 Shunsaku Hirata * fontmap.c, fontmap.h Create fontmap entries for each subfont, containing SFD mapping information and "link" to fontmap entry "foo@SFD@". Removed pdf_lookup_fontmap_record2(). * dvi.c, mpost.c, pdfdev.c: Modification needed by changes in fontmap handling. * spc_pdfm.c: Minor changes in fontmap specials. Fixed leak in "append" fontmap. * subfont.c, subfont.h: Minor fixes and changes. * mpost.c: Bug fix. Fixed a bug (typo) in mp_setfont. 2005-07-14 Shunsaku Hirata * subfont.c, subfont.h: xxx_sfd_yyy -> sfd_xxx_yyy Added code for testing SFD file handling. Added various comments. * fontmap.c: Check if subfont ID is valid before loading SFD mapping table for subfonts. Changes related to the above changes in subfont.c. 2005-07-13 Shunsaku Hirata * dvipdfmx.c: minor fix agl_set_verbose() called in pdffont.c too. * fontmap.c: Workaround for problems in subfont fontmap lookup. Delete all conflicting explicit fontmap entries for subfonts in pdf_insert_fontmap_record() when adding fontmap entry with @SFD@ extension. * subfont.c, subfont.h: Scan SFD file for valid subfont IDs. (Required by changes made to fontmap.c mentioned above.) Removed unnecessary file static variables. Simplified file reading and parsing. * dvi.c: set_verbose() for subfont.c * fontmap.c: Correctly handle fontmap entry like "dmj@DNP@12". 2005-07-11 Shunsaku Hirata * agl.c: Fixed various small bugs 2005-07-09 Shunsaku Hirata * fontmap.c, fontmap.h, pdfdev.c: Fake-bold again 'b' option. This is experimental feature. * pdfdev.c: Graphics state is not properly managed yet. Fake-bold and set_rule changes linewidth without using pdf_dev_setlinewidth(). Enclose set_rule with q and Q. * agl.c: Removed unused variable 2005-07-08 Shunsaku Hirata * tt_cmap.c: Bug fix Format 12 cmap subtable support not working... * fontmap.c: Test for UCS-4 'p' option to specify UCS plane. This is just for testing purpose. Don't use that. 2005-07-08 Shunsaku Hirata * tt_cmap.c: Bug fix in ToUnicode creation ToUnicode CMap creation (inversion of TrueType cmap) was not working for format 12 cmap subtable. 2005-06-27 Shunsaku Hirata * cid_basefont.h: Add missing font definitions. * data/config/dvipdfmx.cfg: Forgot to mention that 'D "..."' line is commented. Now there are no default setting for image converter, so PS inclusion will always fail unless you modify dvipdfmx.cfg. 2005-06-26 Shunsaku Hirata * data/config/cid-x.map: empty Contained examples for fonts with unclear license. Not maintained for a long time. * data/CMap/Adobe-Identity-UCS2: removed Wrongly named file. Content of this file is actually identity mapping. * type0.c: Create dummy ToUnicode CMap "Adobe-Identity-UCS2". * specials.c: spc_push_object() always create label for objects. Objects created with pdf:obj should always be written to output. * ainclude.m4, configure.in: Define __TDS_VERSION__, require xbasename in kpathsea, removed unused macros, simplified. * dpxfile.c: Use __TDS_VERSION__ instead of HAVE_KPSE_ENC_FORMAT and others. * dvipdfmx.c: Alwyas use xbasename(). * fontmap.c, fontmap.h: Fake bold and OTL options removed. * otl_conf.c: Deactivated. * pdfdev.c: Fake bold option removed. (undocumented) * configure.in, acinclude.m4: Various changes. Error if kpathsea library doesn't have xbasename(). Define __TDS_VERSION__, removed HAVE_KPSE_XXX_FORMAT. * configure and others: recreated 2005-06-20 Shunsaku Hirata * tt_gsub.c, tt_gsub.h: Alternate substitution LookupType 3 (Alternate Substitution) GSUB lookup. Unused. 2005-06-13 Shunsaku Hirata * tt_gsub.c: Added support for LookupType 7 (Extension Substitution). Large CJK fonts may want them. Not tested well. 2005-06-09 Shunsaku Hirata * tt_glyf.c: Set top-side bearing to (sTypoAscender - bbox.ury) for fonts without vmtx table (vertOriginY = sTypoAscender). 2005-06-09 Shunsaku Hirata * dpxfile.c: Fixes for old kpathsea library. 2005-06-08 Shunsaku Hirata * dpxfile.c, dpxfile.h: Fallback to old path searching behaviour. Warn about files found in old location. * dvipdfmx.c: set_verbose() for dpxfile. 2005-06-08 Shunsaku Hirata * cid_basefont.h: Update for Adobe Reader 7 * cid.c, pdfobj.c: Accept version tag 1.6 2005-06-06 Shunsaku Hirata * acconfig.h: removed * data/CMap/glyphlist.txt: --> data/config/ No appropriate place yet. * cmap_read.c: Fixed a bug in reading CIDSystemInfo dict. * pdfencoding.c, pdfencoding.h: pdf_load_ToUnicode_stream(). Load ToUnicode CMap from file and then create ToUnicode stream object. See also changes in truetype, type1, type1c, and pdffont. * pdffont.c: First try to load ToUnicode CMap with file name 'tfm_name'. See also changes in truetype, type1, type1c, and pdfencoding. * truetype.c, type1.c, type1c.c Don't create ToUnicode if fontdict already have them. See also changes in pdffont and pdfencoding. * Makefile.am, data/Makefile.am, src/Makefile.am: Fixes, simplified, updated... There are problems in data/Makefile.am yet. Installation directory for .map and .cmap is not correctly set. * Makefile.in and others: recreated * configure.in: 2005-05-07 Shunsaku Hirata * spc_color.c: Color special without push/pop allowed? * spc_pdfm.c: pdf:literal broken for more than 2 years? Fixed a bug in creating annot. * data/config/dvipdfmx.cfg: Comments... 2005-05-04 Shunsaku Hirata * README: update * pdfparse.c: There were several problems in handling broken PDF files. * pdfobj.c: There were several problems in handling broken PDF files. Stream dict is freed twice when parsing stream object failed. * tfm.c: PK font creation was not working due to incorrect design size value returned by tfm_get_desing_size(). * index.html: Update and fixes for links. Many sentences doesn't make sense when the document is printed. Stylesheet default to "plain" style. Encoding set to "utf-8". * style.css: renamed to style_dflt.css * style_plain.css: Plain style stylesheet. 2005-03-07 Shunsaku Hirata * pdfcolor.c: Color not correctly reset at page break? Maybe more bugs... * pdfdoc.c: Encapsulating docxxx within single struct pdf_doc. 2005-03-05 Shunsaku Hirata * mpost.c: Text handling was not correct for subfonts. * pdfencoding.c: Increased verbose level for ToUnicode warning. 2005-02-03 Jin-Hwan Cho * otl_conf.c, spc_pdfm.c: Fixed remaining bugs related to MiKTeX file searching. * cvs tag: DVIPDFMX-20050201 (revised) 2005-02-01 Jin-Hwan Cho * cidtype2.c: Fixed a critical bug on the DW entry for CIDType2 fonts. * cvs tag: DVIPDFMX-20050201 2005-01-06 Jin-Hwan Cho * dvi.c: Fixed a bug on converting subfont character numbers. * snapshot: two snapshots DVIPDFMX-20040824 (release): before full update DVIPDFMX-20040912 (current): with the bug fix on 2005-01-06 2004-09-12 Shunsaku Hirata * specials: rewriting... All \special related files renamed to spc_xxxx and specials. Use function pointer table rather than doing everything with if-else or switch-case. * pdfparse.c: Few functions are expelled from pdfparse.c since they are not related to PDF parse at all. pdfparse still contains dvipdfmx only extension and they might be enabled even when parsing ordinary PDF files. * spc_pdfm, pdfparse: tounicode special moved to spc_pdfm. 2004-09-08 Shunsaku Hirata * ttf, tt_*: ttf.[ch] -> truetype.[ch], renamed few files, added tt_post. * unicode, pdfencoding: ToUnicode support migrated to pdfencoding. 2004-09-05 Shunsaku Hirata * pdfencoding: encodings -> pdfencoding * bug fixes. * pdflimits: This was not pdflimits nor implementation limit of PDF viewers at all. 2004-09-02 Shunsaku Hirata * pngimage, jpegimage: Improvement on embedded ICC profile support. Profiles with same MD5 checksum are shared among images. * pdfcolor: Added something for managing color space resources. * pkfont.c: Depends on tfm.c again. 2004-08-30 Shunsaku Hirata * pdfdoc, pdfdev: Mostly independent from dvi interpreter. pdfparse still depend on dvi.c through pdfspecial.c due to pdf: special @xpos and @ypos. * pdfdraw: Primitive graphics drawing functions and graphics state management: pdf_dev_moveto, pdf_dev_gsave, pdf_dev_lineto, pdf_dev_flushpath, and so on (migrated from mpost.c). PostScript, tpic, and all other graphics specials should not directly access page content stream but must use those functions instead. Text state and colors are still managed completely differently. * Added many comments in source and headers... * Rearranged initialization process again. pdf_init_fontmap: Initialize fontmapping database. pdf_open_document: Open output file, initialize various document resource handler, and set default media size... pdf_init_device: Unit conversion, precision and others. pdf_doc_begin_page: Initialize graphics state, etc. 2004-08-27 Shunsaku Hirata * cidtype0.c: Fixed several bugs in Type1 and CFF -> CID CFF converison. There might be more bugs... * doc/dpxunicode.pdf: Update. 2004-08-24 Shunsaku Hirata * cid_basefont.h: Bug fix Added missing '['. * doc/dpxunicode.pdf: Update * tt_cmap.c: 2004-08-22 Shunsaku Hirata * otl_conf.[ch]: New file OTL feature is managed by configuration file. * fontmap.c: -S and -L option removed. Added option -l. * doc/dpxunicode.pdf: Update. 2004-08-16 Shunsaku Hirata * Unicode support for CID OpenType. 2004-08-15 Shunsaku Hirata * fontmap.c: -f option changed to -S. Added -L for ligature substitution. * tt_gsub.c and others: New fontmap option -L Map ligatures not in Unicode to PUA. ToUnicode support handle substituted glyphs. See, doc/dpxunicode.pdf for details. 2004-08-14 Shunsaku Hirata * otl_opt.[ch]: New file Handle fontmap options for OTL features. 2004-08-11 Shunsaku Hirata * type1.c: Bug fix. The seac operator support was broken since 2004-8-02. * fontmap.c: New fontmap option "-f tag" and "-b number". Specify OpenType layout feature with -f option for "unicode" endoding font. Glyph substituion specified by -f option is not reflected in ToUnicode CMap yet. Only single substitution GSUB subtable is supported. The "-b" option specifies "boldness" parameter for fake-bold (stroke line width of fill-then-stroke text). * ...: Changes for new features mentioned above. 2004-08-08 Shunsaku Hirata * tt_cmap.c, cidtype0.c: Bug fixes 2004-08-08 Shunsaku Hirata * Makefile.in, Makefile.am: cmap_parse --> cmap_read. * cmap_parse.[ch]: Remaned to cmap_read.[ch] * cmap_write.[ch]: New file Create CMap PDF stream object. Output is not efficient yet. * dvi.c: Bug fix Multiply pdf:pagesize special page height and page width by DVI magnification factor. * agl.[ch], unicode.[ch], ....: Use cmap_write. * dpxfile.[ch]: New file ..... 2004-08-07 Shunsaku Hirata * Type1 font accept "unicode" encoding. Converted to CIDFontType 0 CIDFont. 2004-08-02 Shunsaku Hirata * New files: t1_load.[ch], t1_char.[ch] type1.c --> type1.c, t1_load.c, t1_char.c. t1_load: load Type1 font (PFB) into cff_font structure. t1_char: type 1 charstring interpreter and type 1 to type 2 charstring converter (type 2 encoder). There are many duplicated code in cs_type2 and t1_char. 2004-07-31 Shunsaku Hirata * New files: pdfdraw.[ch], pdfcolor.[ch] * mpost.c, tpic.c: Use pdfdraw and pdfcolor. * pdfdev.c: CTM affects box calculation in dev_expand_box(). Special bann/eann set correct bounding box for rotated text. * New keyword "unicode" in fontmap: The keyword "unicode" can be specified as encoding in fontmap. With this option dvipdfmx creates Unicode-to-GID mapping CMap and GID to Unicode (ToUnicode) CMap from TrueType cmap table (format 4 Windows-UCS2 and format 12 Windows-UCS4 subtable is supported) for TrueType font. All font with encoding "unicode" is treated as Type0 font. Currently supported only for Unicode TrueType font. 2004-07-27 Shunsaku Hirata * Many bug fixes. * fontmap.c: Bug fix. Do not increment num_fontmap when replacing duplicate entries. * mpost.c: Support PS operator findfont etc. Not compatible with dvips. 2004-07-18 Shunsaku Hirata * fontmap.c: Add/replace/remove fontmap entries. Added fontmap_add_map_record(). Scan single mapfile line. * fontmap.c, pdfdev.c: fontmap_tex_name(), ... abort if invalid value is passed. * psspecial.c: bmpfile special This is not PS special... * pdfobj.[ch], pdfspecial.c: Added pdf_copy_object(): Copy object data without changing object label. Dvipdfmx now can handle pdf specials like: \special{pdf:obj @foo << /Bar @bar >>} \special{pdf:obj @bar << /Foo @foo >>} Object reference can appear before definition of object. 2004-07-15 Shunsaku Hirata * cidtype2.c, type0.c: Write CID glyph metrics for non-embedded TrueType fonts. * dvipdfmx.c, cidtype0.[ch], cidtype2.[ch], cid.[ch]: Treat all CIDFont as fixed pitch font if bit 0x0004 is set in "-C" option. (for backward compatibility) 2004-07-14 Shunsaku Hirata * New files: bmpimage.c, bmpimage.h Windows BMP image format support. (not tested well...) 1/4/8/24-bits per pixel images, uncompressed or 4-bit/8-bit RLE compressed images are supported. * pdfparse.c: Added dictionary key "T" to the list of strings to be converted to Unicode... 2004-07-12 Shunsaku Hirata * dvipdfmx.c: Added new option "-C". Thumbnail support available again. * tpic.c, tpic.h: Spline path ("sp" command) was not handled correctly. Use (semi-)transparent color for shading, enabled by option -C 2. (not enabled by default since it requires PDF version 1.4+) Apply shading only to closed paths. * psspecial.c: epsbox support was broken. * pdfdoc.c: Thumbnail supported again 2004-07-04 Shunsaku Hirata * configure, ...: HAVE_KPSE_OPENTYPE_FORMAT and HAVE_KPSE_ENC_FORMAT is not correctly set when compiled with gcc3. 2004-06-28 Shunsaku Hirata * pdfximage.c: Do not clip PDF images when "clip" option is not given. 2004-06-27 Shunsaku Hirata * Workaround for "three dot leader" problem. 2004-06-16 Shunsaku Hirata * Do not adjust refernece point of included images if users do not supply bounding box. * Look for .ofm first. Workaround for incompatibility introduced on change 2004-01-20. * pdfdoc.c: Fixed bug that dvipdfmx crushes with null string key in name tree. 2004-06-02 Jin-Hwan Cho * pdfspecial.c: Bug fix on cropbox in do_bxobj(). 2004-05-03 Shunsaku Hirata * jpegimage.c: Uninitialized variable. 2004-04-09 Shunsaku Hirata * README, index.html: update * contrib/patches/glyphlist-cm-add.txt: new file Additional glyphlist for Computer Modern font. * NEWS: new file Empty. * pdfdev.c: Do not increase line width. 2004-04-03 Shunsaku Hirata * README: update 2004-03-25 Shunsaku Hirata * jpegimage.c: ICC profile Added APP2 ICC_PROFILE support. ICC profile is not deleted from JPEG stream. * Embedded ICC profile support for JPEG and PNG enabled. ICC profiles are not cached. Alternate color space? 2004-03-24 Shunsaku Hirata * Thumbnail support removed. * index.html: Quick check against WCAG 1.0. Lowercasified all HTML tags for migration to XHTML. * pdfdoc.c: Dvipdfm/dvipdfmx copyright notice remved from PDF Docinfo's /Producer string, simplified to "dvipdfmx (VERSION)". * configure.in, configure: Minor change. 2004-03-22 Shunsaku Hirata * pngimage.c: Minor changes. Remaining things for complete PNG 1.1 support is: Complete iCCP support. Support tEXt and other metadata chunks using XML Metadata stream. pHYS (and oFFs) support (in ebb). Preserve filter (PNG predictor) if possible, but some softwares using filters very inefficiently. 16-bit RGB/Gray (PDF-1.5). JPEG (JFIF) can also store ICC profile with APP2. Implementing very basic ICC profile support might be useful. 2004-03-21 Shunsaku Hirata * README: updated * data/config/cid-x.map: "slant" option -s supported in vertical mode. * pngimage.c: Fixed a serious bug, more chunks supported. Gamma obtained by png_get_gAMA() is inverse of screen gamma that the image is prepared. Gamma is sometimes inverted... Approximated CalRGB color space used along with /Intent if sRGB chunk is found (not sure how to treat it). Experimental iCCP support --enable-iccp. Doesn't validate ICC profile embedded in PNG. Range? Rendering Intent? * pdfdev.c: A line width of zero means thinnest line drawable in target device. dev_rule() was wrongly changed to increase line width if the width is too thin. 2004-03-20 Shunsaku Hirata * unicode.c, agl.c, type0.c: ToUnicode CMap support Always create ToUnicode CMap for simple fonts regardless of the existence of glyph list file. If 10% of glyph names can not be converted to Unicode, ToUnicode CMap is not embedded. The reason for the (non-zero) limit 10% is; ToUnicode CMap is shared among fonts using same .enc file. In this case, Unicode mapping is created for all glyphs in the .enc file where non-standard glyph names like "visiblespace" is often expected. But those glyphs are usually not used in the TeX's output and hence it should be simply ignored if mapping is not available. Sharing ToUnicode CMap is not so efficient as expected for simple fonts, "subsetted" ToUnicode CMap might be better. type0.c search ToUnicode CMap named as REGISTRY-ORDERING-UCS2 or FONTNAME-UCS2, the latter case is for CIDFontType 2 font using font's internal glyph ordering. It does not care about the content of CMap at all, especially usecmap is not checked. Don't use it. * README, data/config/cid-x.map: Description about -m option removed from README. This option is no longer necessary for usual cases, and use of it is not recommended as the original dvipdfm fails to read fontmap file with this option. * cmap.c: Increment n-1 th byte if n-th byte overflow in bfrange like: <00> <01> <00 FF> The dstCode string is treated like an integer, but the number of bytes required for that character code is not increased. Acrobat's "ToUnicode" CMap contains such bfrange entries and causing a serious problem in Korean TrueType fonts. 2004-03-18 Shunsaku Hirata * pdfspecial.c: pdf:literal direct should not translate origin. 2004-03-17 Shunsaku Hirata * type1.c, type1c.c, cmap.c, ...: Memory leak. 2004-03-16 Shunsaku Hirata * mpost.c, psspecialc.: Fixed bugs. 2004-03-14 Shunsaku Hirata * dvipdfmx.c: Scan first page's paper size special first. * pdfparse.c: \x in PDF string - broken This (undocumented) extension to the PDF string object disturbs and interrupts PDF object parsing and subsequent text re-encoding specified by tounicode special (another extension). And it makes impossible to separate parsing and re-encoding process ("x" after backslash representing Unicode hex value...). Support for this extension was removed since it is desirable to move (or isolate) tounicode special to the appropriate place, not pdfparse but pdfspecial, if possible. Note that this kind of extension might be incompatible with all other PDF applications, "\x" should be treated as a single "x" in literal strings (like PostScript). * pdfspecial.c: Special image/uxobj is partially broken. Special uxobj doesn't work for external images after the change made on 2003-08-12. It only works for the grabbed page contents created by special bxobj/exobj. Access to XObject reference (but can't be placed in the document) is now available again. Supporting uxobj is usually not necessary for external images because they are always reused as a document resource but allowing access to the (form) XObject's indirect reference by name like @foo is usefull for using them as an entry of appearance dictionary in annotations. * pdfspecial.c, pdfdev.c: Special bann/eann does not take account of current transformation matrix correctly. It result in incorrect /Rect for all annotations created by special bann/eann if transformation is applied. Not fixed yet. * pdfparse.c: "ToUnicode" special Illegal PDF string is expected within pdf: specials. Many users try to use raw Shift-JIS text without properly escaping backslash appears in the 2nd byte of double-byte character. The same problem occurs in many other encodings. As a workaround for this problem, backslashes are not treated as "escape" in some situation when the current "ToUnicode" CMap name contains a substring RKSJ, B5, GBK, or KSC. * cmap.c, ...: 2004-03-12 Shunsaku Hirata * mpost.c: Skip until %%EndProlog. * pdfspecial.c, pdfdev.c, pdfdoc.c: pdf_doc_set_mediasize()/pdf_doc_get_mediasize(). * tfm.c tfm.c was highly wasteful and inflexible. Preparing for 31-bits character code support. 2004-03-11 Shunsaku Hirata * dvipdfmx.c, dvi.[ch], pdfdev.[ch], pdfdoc.[ch],...: An effort to eliminate complicated initilization chain... * pdfximage.c, pdfximage.h: New file. An effort to make image support more flexible. Included images and grabbed page contents (forms) should be managed by pdf_ximage as document resource (like font). * pdfspecial.c, psspecial.c: Image support moved to pdfximage. (might be broken) * mpost.[ch], epdf.[ch], pngimage.[ch], jpegimage.[ch], psimage.[ch]: Image support moved to pdfximage. * dvipdfmx.c, mpost.[ch]: Experimental "mpspdfmx" mode -M. * pdfspecial.c, pdfdev.c, pdfdoc.c: Hard to explain... There are many misleading/illogical things. * dvipdfmx.c, dvi.[ch]: DVI specials related to page/paper size is pre-scanned. * dvipdfmx.c: Page selection option "-s" accept page labels. For example, command line option "-s 1-10/r,11-12/D:2:empty-page-,13-" set page labels for page 1-10 to lowercase roman numerals, page 11-12 to arabic numerals starting with 2 and with prefix "empty-page-",... * dvipdfmx.c: An option for "bookmark open lavel" -O. * pdfdoc.c: Undocumented watermark/page decoration feature. Not completed yet. * thumbnail.c: I can hardly think that thumbnail support is useful. Acrobat Reader can create thumbnail image and other PDF viewers like gs* and Xpdf doesn't support it. Currently broken due to changes in pngimage.c. * dvipdfmx.c, pdfdev.c: The default precision set back to 2 fractional digits. Most sensitive part is dev_set_string() but the accuracy of this part is essentially limited to 0.01 point in the most cases. Increasing precision above 2 does not give good result (except DVI rules) but merely increases file size. 2004-03-05 Shunsaku Hirata * pdfdoc.c: Fixed a critical bug introduced on change 2004-02-29: Consistency check was done after pages are free-ed. * pdfdoc.c, pdfspecial.c: special pdf:names. Add name tree entry. * pdfobj.c, pdfparse.c: Few functions added for manipulating PDF array. 2004-03-04 Shunsaku Hirata * pdfparse.c: Added missing () in macro PDF_TOKEN_END() * pdffont.c: Compiler warning on uninitialized variable. 2004-03-02 Shunsaku Hirata * Groff support: Groff support was essentially broken. Dvipdfm(x) generated very inaccurate result due to incorrect choice of scale/unit. It should be fixed now. 2004-02-29 Shunsaku Hirata * pdfspecial, pdfdoc, pdfobj, pdfparse: Partially rewrote... * pdfdoc.c: Moved static ... to static struct { ... } docXXX. Removed (strange) optimization. Outline item open/close. * tfm.c: Partially rewote... 2004-02-25 Shunsaku Hirata * Too many warnings when compiled with -Wshadow... pdfdoc.c need to be fixed. * -Wmissing-prototypes, -Wmissing-declarations 2004-02-21 Shunsaku Hirata * type0.c: Changes for New TDS/kpathsea. Fixed a bug in get_tounicode_cmap()... 2004-02-19 Shunsaku Hirata * cidtype0.c, type0.c: Write CID glyph metrics for non-embedded CFF/OpenType fonts. 2004-02-18 Shunsaku Hirata * tpic.c: Ignore unpainted path in flush_path(). show_path() output 'n' operator for path without stroking nor filling to avoid Acrobat error (bug?) "Illegal operation inside a path.". 2004-02-17 Shunsaku Hirata * pkfont.c, pkfont.h, pdfdev.c: Removed dependency to TFM in pkfont. * type1.[ch], type1c.[ch], type0.[ch], ttf.[ch]: No longer need to know font resource name and remap. * pdffont.c, pdffont.h: Added new file. * type1.[ch], type1c.[ch], ttf.[ch], pkfont.[ch], pdfdev.c, dvipdfmx.c: Simple font support will be unified to pdffont.c 2004-02-16 Shunsaku Hirata * type1.c, type1c.c, cidtype0.c: Remove UniqueID and XUID from CFF Top dict. * type1.c: Keep original FontBBox instead of replacing it with tight one obtained from glyph's charstrings to prevent Acrobat 4 from greeking text. Acrobat 4 may display text drawn with fonts which only contain small glyphs like period and minus-sign with greeked text. * tt_build.c: TrueType output slightly optimized. * ttf.c, cidtype2.c: Preserve 'name' table when embedding TrueType font. 2004-02-15 Shunsaku Hirata * type1c.c, cs_type2.c, cs_type2.h, cidtype0.c: Dependency to TFM removed. Widths are obtained from font file. * type1c.c: Delete unused strings from CFF string index. * tfm.c, tfm.h: Added tfm_get_codingscheme(). CODINGSCHEME from TFM header. * agl.c, agl.h, unicode.c, unicode.h, ttf.c: Handle decomposed (in Unicode) glyphs in AGL and glyph names of a form XXX_YYY_ZZZ in ToUnicode CMap. The glyph name with suffix, like "x.y", is sotred as a glyph "x" with a suffix that implies a variant "y" of "x" in AGL table, not as "x.y" glyph. (for future use) 2004-02-13 Shunsaku Hirata * type1.c: Bug fix. Skip first four bytes of encrypted PFB section. Not warn about "dotsection" operator which exist only for compatibility to Type 1 font rasterizer that can't do hint replacement. * tfm.c, tfm.h: Fixed a bug that calculation of tfm_string_height() was incorrect. ASCII pTeX specific things are put within #ifndef WITHOUT_ASCII_PTEX. 2004-02-12 Shunsaku Hirata * type1.c: Another bug in flex handling. hmoveto and vmoveto allowed. * type1c.c, cmap.c, cidtype0.c, encodings.c, fontmap.c: Changes for new TDS and teTeX/kpathsea. * encodings.c, encodings.h, pdfdev.c: Removed get_encoding(). 2004-02-10 Shunsaku Hirata * type1.c: Bug fix. Compression of vhcurveto-hvcurveto was not actually done. One more bug in flex handling. 2004-02-08 Shunsaku Hirata * dvipdfmx.c: paperinit() and paperdone(). * tfm.c: Bug fix. Read OFM level 1 header, required by get_design_size(). * dvi.c, dvi.h, vf.c: DVI specials in virtual fonts (patch by P. TJ MacPhee). 2004-02-05 Shunsaku Hirata * cid.c, cidtype2.c: Use font's internal glyph ordering if CSI Adobe-Identity is specified for TrueType font. This feature is intended to be used with Identity CMap and /I option in fontmap. * configure, acinclude.m4, acconfig.h, config.h.in,...: Check kpse_{enc,cmap,sfd,opentype}_format. * type1.c: Fixed a bug in flex handling. 2004-02-04 Shunsaku Hirata * dvipdfmx.c: libpaper support. * configure, acinclude.m4, acconfig.h, config.h.in: --with-paper added for configure option. HAVE_LIBPAPER in config.h. * pdfspecial.c: Added "bp" and "truebp" in units[]. * Accept w,h for papersize option. 2004-02-03 Shunsaku Hirata * type1.c: Counter control (othersubr #12 and #13) supported. 2004-02-01 Shunsaku Hirata * tt_build.c, tt_build.h, cidtype2.c, ttf.c: Workaround for Ghostscript's bug. Ghostscript doesn't handle CIDToGIDMap correctly. Use Identity CIDToGIDMap. * ttf.c: Removed dependency to TFM. Glyph metric is obtained from font file. * type1.c: Warn about "seac" operator. 2004-01-31 Shunsaku Hirata * type1.c: Bug fix. * Added many 'const'. 2004-01-29 Shunsaku Hirata * type1.c: Type 1 -> Type 1C conversion PostScript Type 1 font is converted to Type 1C (CFF). There are few limitations: Ghost hint is not converted to edge hint. The Type 1 charstring operator "dotsection" is ignored. The ExpertEncoding is not supported. The "seac" operator is not supported. Counter control is not supported. * Makefile.in, Makefile.am, standardenc.h: standardenc.h removed. * pst.c: Fixed a bug in skip_comments() that cause infinite loop. 2004-01-27 Shunsaku Hirata * cff.c, cff.h, cff_dict.c, cff_dict.h: Added few functions for Type 1 -> Type 1C conversion. 2004-01-24 Shunsaku Hirata * pkfont.c: Acrobat and MacOS X's Preview have some problem with Type 3 font. Acrobat cannot handle non-invertible transformation matrix which happens when PK glyph with width and/or height equals to zero is embedded as Type 3 font. The preview app. of MacOS X 10.3 cannot display few glyphs generated by dvipdfm with unknown reasons. 2004-01-21 Shunsaku Hirata * cs_type2.c: Don't use rint(). 2004-01-21 Shunsaku Hirata * psimage.c, configure.in, config.h.in: Use mkstemp() if available. 2004-01-20 Shunsaku Hirata * tt_table.[ch], tt_build.[ch], cidtype[02].c: CIDFont glyph metrics for embedded font. 2004-01-17 Shunsaku Hirata * pdfobj.[ch], pdfparse.[ch], mpost.c, pdfencrypt.c, pngimage.c: Removed pdf_*_stringx(). * pdfobj.c, pdfparse.c: Changes for PDF-1.3 compliance. 2004-01-14 Shunsaku Hirata * ToUnicode support for CFF/OpenType + built-in encoding. * Bug fix in cff.c, type1c.c * Workaround for Shift-JIS string parsing in pdfparse.c. * pngimage.c don't require PNG-write/MNG-feature/progressive-read. * Unicode: unsigned short --> long. * Removed winansi.h 2004-01-07 Shunsaku Hirata * IMPORTANT CHANGE: Searching order is reversed in fontmap.c. Last match takes precedence in font mapping now. * Fixed bug in cff.c, cmap.c, and type1c.c. 2003-12-28 Shunsaku Hirata * Bug fixes in pdfparse.c and cid_basefont.h. 2003-12-15 Shunsaku Hirata * Minor fixes. * Added PDF special "fstream" - create PDF stream object from file. This allows users to attach files in PDF. 2003-12-07 Shunsaku Hirata * JPEG moved to jpegimage.c. * Bug fixes. 2003-12-06 Shunsaku Hirata * Changes in pdfspecail.c. * Fixed a bug in mpost.c. 2003-12-02 Shunsaku Hirata * Added a patch for dvipdfm's ebb.c. (PDF ArtBox support) 2003-11-29 Shunsaku Hirata * Bug fixes. * CFF/OpenType support for single-byte font. 2003-11-26 Shunsaku Hirata * Little enhancement to epdf.c. * Merged epsbox.sty patch by Noda. 2003-11-25 Shunsaku Hirata * Many part of font/encoding related code rewrote. * Unicode support for single-byte font. 2003-11-10 Jin-Hwan Cho * Fixed a bug on the algorithm converting arc to cubic bezier curves in arc() (tpic.c), which is reported by Werner Lemberg. 2003-11-04 Shunsaku Hirata * PNG support completely rewritten. Color pallete images are now not converted to 24-bit RGB, colorspace Indexed is used instead. Added tRNS and alpha chunnel (transparency) support. Experimental gAMA, cHRM and sRGB support. 2003-11-03 Shunsaku Hirata * Fixed bug that dvipdfmx stops with error "Invalid Subr Index" when embedding CFF OpenType font. 2003-10-09 Jin-Hwan Cho * Add a missing line including pdfparse.h in dvi.c, which is reported by Akira Kakuto. * Changed the default font map files for teTeX 2.x as suggested by Haruhiko Okumura. 2003-10-08 Jin-Hwan Cho * A bug on the paper size checking routine of dvi_paper_size() in dvi.c is fixed, which is reported by Haruhiko Okumura. 2003-10-06 Jin-Hwan Cho * A bug on the kerning routine of dev_set_string() in pdfdev.c is fixed, which is reported by Edward G.J. Lee. 2003-08-12 Jin-Hwan Cho * Changed to embed images only one time. * Updated configure.in and config.h.in to support WORDS_BIGENDIAN and SIZEOF_SHORT, SIZEOF_INT, SIZEOF_LONG. 2003-08-11 Jin-Hwan Cho * A faked font name was used for TrueType fonts without any PS font name as suggested by Hin-Tak Leung. 2003-08-05 Jin-Hwan Cho * Built-in codes (originally from libgcrypt-1.1.42) was used for PDF encryption instead of the OpenSSL library as suggested by Thomas Esser. 2003-07-28 Jin-Hwan Cho * A significant bug reported from Haruhiko Okumura is fixed. The bug occurs only if compiled with gcc 3.x, not gcc 2.x. 2003-03-26 Jin-Hwan Cho * New feature, hexadecimal Unicode, suggested by iNOUE Koich! and Akira Kakuto is implemented. 2003-03-13 Jin-Hwan Cho * Fixed bugs on reading and writing OpenType fonts. 2003-03-12 Jin-Hwan Cho * Applied changes for MiKTeX in ps_include() (psimage.c). 2003-03-11 Jin-Hwan Cho * Fixed a bug in cff_charsets_lookup() (cff.c). 2003-02-06 Jin-Hwan Cho * Fixed a bug in string_mode() (pdfdev.c) reported by iNOUE Koich!. 2002-12-30 Jin-Hwan Cho * Fixed a bug in writing null array. * Fixed a bug in reading CMap and encoding files in MiKTeX. 2002-11-28 Jin-Hwan Cho * Support color shading feature in ConTeXt. 2002-11-25 Jin-Hwan Cho * New feature pdfdecimaldigits (default=3) is implemented. * Fixed a missing line in dvipdfmx.c. * Fixed a bug in page attribute boxes, e.g., CropBox, TrimBox. 2002-11-11 Jin-Hwan Cho * The "pdf: dest" special reads pdf_object not pdf_array. * Fixed a bug in finding encoding files with extension. * Look up the entries in miktex.ini while finding files. 2002-10-31 Jin-Hwan Cho * Fixed a bug in finding files with the kpathsea library 3.3.7. 2002-10-28 Jin-Hwan Cho * The project name was changed to dvipdfmx from dvipdfm-cjk. * Changed the way of finding files (MiKTeX and KPATHSEA) (contributed by Akira Kakuto). * Support embedded MetaPost graphics in ConTeXt using new PDF special \special{pdf:literal [direct|reverse] ...}. 2002-10-25 Jin-Hwan Cho * Removed all files related to the original dvipdfm, and all CMap files from Adobe. * Changed the executable file dvipdfm to dvipdfmx, the configuration file config to dvipdfmx.cfg, the Type0 map file cid.map to cid-x.map. 2002-10-24 Jin-Hwan Cho * It is possible to change the size of each page with the special command \special{pdf:pagesize ...}. Override the previous changes if there are two such commands in a page. 2002-10-17 Jin-Hwan Cho * Fixed a bug in parsing an array in two lines. (mpost.c) * Changed not to lookup the header to find fonts. (mpost.c) 2002-10-09 Jin-Hwan Cho * Duplication is allowed for named_references[]. (pdfspecial.c) 2002-06-05 Jin-Hwan Cho * Fixed a bug in parse_key_val() found by Yasuhiro Okazaki. 2002-05-31 Jin-Hwan Cho * Bug fixed for ToUnicode conversion in PDF outlines(bookmarks), docinfo, and text annotations (contributed by Akira Kakuto). * Included the ToUnicode CMap, EUC-UCS2. 2002-05-22 Jin-Hwan Cho * Support CJK characters in PDF outlines(bookmarks), docinfo, and text annotations (contributed by Akira Kakuto). 2002-05-15 Jin-Hwan Cho * PDF encryption support with the OpenSSL library. 2002-04-30 Jin-Hwan Cho * ToUnicode CMap (only Unicode) support for Identity-H(V). * Included the ToUnicode CMap, Adobe-Identity-UCS2. 2002-04-22 Jin-Hwan Cho * Included two CMaps, UniKSCms-UCS2-H(V) which fix a bug in UniKS-UCS-H(V). 2002-04-17 Jin-Hwan Cho * Changed xxx_set_debug() to #ifdef DEBUG. 2002-04-16 Jin-Hwan Cho * Improved reading PNG image files. * Rearranged all files to work with GNU autotools. * Removed HAVE_TTF_FORMATS and HAVE_OFM_FORMATS. 2002-04-09 Jin-Hwan Cho * Minor fix on 'source/pkfont.c'. 2002-04-07 Jin-Hwan Cho * Fixed all warning messages generated with MSVC++. * Rewrote the functions 'xxx_locate_font()' again. 2002-04-04 Jin-Hwan Cho * Changed the order of loading virtual and physical fonts as suggested by Akira Kakuto . If map entries are found, virtual fonts will not be used anymore. 2002-04-02 Jin-Hwan Cho * Implemented no-embedding and stylistic variants for CIDFontType0. 2002-03-31 Jin-Hwan Cho * Added two CMaps, ETenms-B5-H and ETenms-B5-V. 2002-03-30 Shunsaku Hirata * Minor fix on the patch by Takanori Uchiyama. 2002-03-29 Jin-Hwan Cho * Applied the patch by Takanori Uchiyama , supporting the direction command of ASCII pTeX's extension in virtual fonts. * Two CMaps, Identity-H and Identity-V are hard-coded in cmap.c. 2002-03-20 Jin-Hwan Cho * Applied the patch by Jungshik Shin, not embedding the unnecessary character /.notdef in type1 fonts. 2002-03-20 Shunsaku Hirata * Modified cid_basefont.h. 2002-03-05 Jin-Hwan Cho * Support for UTF8 encoding through Identity CMap. 2002-02-25 Jin-Hwan Cho * Implemented no-embedding feature for CIDFontType2. * Implemented Bold, Italic, BoldItalic attributes for CIDFontType2 and CID predefined fonts. 2002-02-22 Jin-Hwan Cho * Implemented subfont approach for CJK and HLaTeX. 2002-01-07 Shunsaku Hirata * Fixed bug introduced on 2002-01-05. 2002-01-05 Shunsaku Hirata * Workaround for some bad TrueType fonts that does not have US-English PostScript font name. 2001-12-27 Shunsaku Hirata * Changed to parse all charstrings. Check for the SEAC character. * Minor changes in CFF related codes. 2001-12-26 Shunsaku Hirata * Merged cmap and cmap_parse. 2001-12-25 Shunsaku Hirata * Cleaned up TrueType and CFF related codes. * Minor bug fixes. 2001-12-23 Shunsaku Hirata * Fixed bug that always requires Private DICT(s) which may not exist. * Cleaned up TrueType and CFF related things. 2001-12-18 Shunsaku Hirata * Biased subroutine number is signed in cs_type2.c. 2001-12-14 Shunsaku Hirata * macglyphs.h regenerated from apple's TTRefMan. 2001-12-13 Shunsaku Hirata * Minor cleanups. * Fixed typo (yacture -> yacute, Islash -> lslash) in macglyphs.h. * Fixed signedness of TureType cmap subtable format 0 data. 2001-12-11 Shunsaku Hirata * Minor fixes. * Fixed bug that interpretion of hmtx table is still wrong. * Fixed fixed(). * More improved FontDescriptor for TrueType/OpenType fonts. * Support for PostScript glyph name in TrueType post table. Full subsettting is also available for non-CJK TrueType fonts. Cyrillic and other TrueType fonts can be used. Requires post table with version 2.0. 2001-12-09 Shunsaku Hirata * Bug fixes. 2001-12-05 Shunsaku Hirata * Fixed bug that last entry of the loca table may not be read. * DW is set to that of missing glyph (CIDFontType 2). * Bug fixes. 2001-12-04 Shunsaku Hirata * Fixes. 2001-12-03 Shunsaku Hirata * Minor charnges, cleanups. * Write CIDSet. 2001-12-02 Shunsaku Hirata * Initial CFF CID-Keyed Font support. Only Type 2 charstring is supported. 2001-11-27 Shunsaku Hirata * Minor fixes, cleanups. 2001-11-26 Shunsaku Hirata * Fixed (serious) bug that hmtx table is incorrectly interpreted. * Fixed bug that encoding ID of RPC and BIG5 are interchanged. * Added missing cast in cidtype2.c. 2001-11-25 Shunsaku Hirata * Bug fix. 2001-11-24 Shunsaku Hirata * Added missing free() (RELEASE) in cidtype2.c. * Minor fixes. * Fixed typos. 2001-11-22 Shunsaku Hirata * Clean up. * Allow `Preview & Print' licensed fonts to be embedded. TrueType OS/2 table is preserved. 2001-11-20 Shunsaku Hirata * Minor cleanups. Fixed typos. * Strip out bad strings form PostScript names; -WIN-RKSJ-H, -WINP-RKSJ-H, -WING-RKSJ-H, and -90pv-RKSJ-H. Some font uses PostScript name one of above strings appended, even if Shift-JIS encoding is not used. 2001-11-18 Shunsaku Hirata * Workaround for Omega: Omega uses setxxx (single-byte char) command while double-byte font is in use. Pass character type to dev_set_string() to distinguish single-byte and double-byte. Maybe, it works with CJK-LaTeX by preparing set of CMaps. * Write double-byte strings as hex strings. This makes bit faster and tend to generate smaller PDF in most cases. * CMap parser uses wrok_buffer in mfileio. * Reserved w_array and w2_array in struct cid_font for future use. * Now write /W array for TrueType, proportional latin and kana works properly. The metric information are taken from font file, not from TFM, to ensure we always have correct widths. The default width (DW) is set to 1000. * Changed sizeof() appeared in sfnt_put_xxx() to expected values in sfnt.h. I do not assume LONG to be exactly 32-bit long, .... 2001-11-17 Shunsaku Hirata * Added Chinese and Korean related stuff again. Chinese and Korean languages are supported through Omega, or by achieving Ascii pTeX compatibility. * Closing CMaps is missing. Added cmap_close_all(), which is invoked by dev_close_all_fonts() in pdfdev.c. * Check length of output string for Type 0 font. It must be multiple of 2. Omega uses 8-bit and 16-bit mixed string even for double-byte fonts. This causes problems. * Abort if missing/invalid character is used. 2001-11-15 Shunsaku Hirata * Bug fixes. 2001-11-12 Shunsaku Hirata * (nearly) Completely rewritten. * Force Identity CMap. (double-byte character codes are mapped to CID) * Added CMap parser. (mkcidcct.pl is no longer needed) * Sync with jpatch. 2001-10-01 Shunsaku Hirata * Should return -1 when encoding_id >= 0 in type0_font(). 2001-08-08 Shunsaku Hirata * Fixed a bug that raster images show up incorrectly in vertical writing mode. 2001-08-07 Shunsaku Hirata * Removed dev_set_string2(), use dev_set_string() instead. * set_dev_vwmode() is renamed to set_dev_wmode(). Added dev_wmode() and mp_mode flag in pdfdev.c. * Support for Japanized MetaPost (pdfparse.h, mpost.c). 2001-07-21 Shunsaku Hirata * Fixed a bug that do not reset text state in vertical writing. 2001-07-20 Shunsaku Hirata * Fixed wrong positioning for slanted non-CJK font in vertical writing. 2001-07-19 Shunsaku Hirata * Support OS/2 table version 0x0002. 2001-07-17 Shunsaku Hirata * Support multiple parents for a single CID-keyed font (incomplete). TODO: check the consistency of charcter collections. * Removed entries for unsupported CMaps. * Updated cid.map. 2001-07-15 Shunsaku Hirata * Better font metric info for CIDFontType2. * Added (partial) support for vertical writing. Added wmode flag to struct a_type0_font and dev_font. TODO: GSUB support in CIDFontType2. /W2 and /DW2. * Added support for vertical version of CMaps in mkcidcct.pl 2001-07-07 Shunsaku Hirata * Minor fixes. * Updated out2uni.c 2001-06-30 Shunsaku Hirata * --> dvipdfm-0.13.2c 2001-06-29 Shunsaku Hirata * Fixed a bug that select incorrect indexToLocaFormat (cidtype2.c). * Fixed several bugs in cidtype2.c. 2001-06-27 Shunsaku Hirata * Added support for CJK-TrueType font embedding. * Cleaned up Type0 and CID font handling code.