Documentation


Fixes problems:
	- An isolated ampersand needed to be converted to the HTML form.
(doc/Bug-form.htm [1.12]: 2000/11/04 17:53:57 lpd)
	- cvs2hist was generating invalid anchor names and not escaping
special charaters. also enable '-r' commandline option, and pass '-b' if
it's not specified for default branch logs only. Changes.htm now passes the
validator. Minor url cleanup in Release.htm (doc/Changes.htm [1.19],
doc/Release.htm [1.23]: 2000/11/10 01:52:35 giles)
	- Removes obsolete documentation reference to -DUsePrinterImages and
clarifies the entry for the replacing -DPrinted. Fixes sourceforge bug #
121907. (doc/Use.htm [1.16]: 2000/11/20 20:34:23 giles)
	- Removes obsolete documentation reference to -DUsePrinterImages and
clarifies the entry for the replacing -DPrinted. Corresponds to revision
1.16 in the main branch. Fixes sourceforge bug # 121907.
(doc/Use.htm [1.13.2.3]: 2000/11/20 20:39:10 giles)
	- Type 1 CharStrings loaded from disk for type 0 CIDFonts were never
freed, except by garbage collection (which doesn't reclaim chunks with only
a small amount of surviving data). In practice, this probably only affects
the PDF writer. (doc/News.htm [1.71]: 2000/11/23 23:34:23 lpd)

Switches the rest of the human-maintained html documentation to the new
stylesheet. We now pass validator.w3.org for HTML 4.01 with the exception of
an error I couldn't figure out in Bug-form.htm. (doc/Bug-form.htm [1.11],
doc/Fonts.htm [1.9], doc/Helpers.htm [1.7], doc/Hershey.htm [1.3],
doc/Htmstyle.htm [1.7], doc/Humor.htm [1.2], doc/Install.htm [1.10],
doc/Language.htm [1.21], doc/Lib.htm [1.4], doc/Maintain.htm [1.7],
doc/Make.htm [1.17], doc/New-user.htm [1.15], doc/Projects.htm [1.20],
doc/Ps-style.htm [1.4], doc/Ps2epsi.htm [1.5], doc/Ps2pdf.htm [1.19],
doc/Psfiles.htm [1.10], doc/Public.htm [1.4], doc/Readme.htm [1.15],
doc/Release.htm [1.22], doc/Source.htm [1.4], doc/Tester.htm [1.4],
doc/Unix-lpr.htm [1.4], doc/Use.htm [1.15], doc/Xfonts.htm [1.4],
doc/index.html [1.2]: 2000/11/04 05:00:00 giles)

updated News.htm to use gs.css (doc/News.htm [1.70]: 2000/11/04 21:59:35
giles)

Adds documentation on Ghostscript's object-oriented coding conventions, and
more documentation on error generation. (doc/C-style.htm [1.7],
doc/Develop.htm [1.19]: 2000/11/07 07:34:57 lpd)

Expands the description of the color mapping process.
(doc/Develop.htm [1.20]: 2000/11/09 05:41:46 lpd)

Updates version number to 6.32 in preparation for release.
(doc/Bug-form.htm [1.9.2.2], doc/Bug-info.htm [1.7.2.2],
doc/C-style.htm [1.4.2.2], doc/Commprod.htm [1.3.2.2],
doc/Copying.htm [1.3.2.2], doc/Current.htm [1.3.2.2], doc/DLL.htm [1.3.2.2],
doc/Devices.htm [1.13.2.2], doc/Drivers.htm [1.8.2.2],
doc/Fonts.htm [1.8.2.2], doc/Helpers.htm [1.6.2.2],
doc/History1.htm [1.3.2.2], doc/History2.htm [1.3.2.2],
doc/History3.htm [1.3.2.2], doc/History4.htm [1.3.2.2]: 2000/11/28 01:31:21
raph)

Updates version number to 6.32 in preparation for release.
(doc/History5.htm [1.5.2.2], doc/History6.htm [1.14.2.3],
doc/Htmstyle.htm [1.6.2.2], doc/Install.htm [1.9.2.2],
doc/Language.htm [1.19.2.2], doc/Lib.htm [1.3.2.2],
doc/Maintain.htm [1.4.2.2], doc/Make.htm [1.15.2.2],
doc/New-user.htm [1.13.2.2], doc/News.htm [1.65.2.2],
doc/Projects.htm [1.19.2.2], doc/Ps2epsi.htm [1.4.2.2],
doc/Ps2pdf.htm [1.18.2.2], doc/Psfiles.htm [1.8.2.2],
doc/Readme.htm [1.13.2.2], doc/Release.htm [1.21.2.2],
doc/Source.htm [1.3.2.2], doc/Tester.htm [1.3.2.2],
doc/Unix-lpr.htm [1.3.2.2], doc/Use.htm [1.13.2.4],
doc/Xfonts.htm [1.3.2.2], doc/gs-vms.hlp [1.3.2.2], man/dvipdf.1 [1.1.2.2],
man/font2c.1 [1.1.2.2], man/gs.1 [1.1.2.2], man/gslp.1 [1.1.2.2],
man/gsnd.1 [1.1.2.2], man/pdf2dsc.1 [1.1.2.2], man/pdf2ps.1 [1.3.2.2],
man/pdfopt.1 [1.1.2.2], man/pf2afm.1 [1.1.2.2], man/pfbtopfa.1 [1.2.2.2],
man/printafm.1 [1.1.2.2], man/ps2ascii.1 [1.1.2.2], man/ps2epsi.1 [1.1.2.2],
man/ps2pdf.1 [1.5.2.2], man/ps2pdfwr.1 [1.2.2.2], man/ps2ps.1 [1.7.2.2],
man/wftopfa.1 [1.1.2.2]: 2000/11/28 01:31:24 raph)

Procedures


Fixes problems:
	- WMode was generally ignored in handling the metrics when rendering
a Type 1 or Type 2 CharString; also, the default metrics for WMode = 1 were
the same as the default metrics for WMode = 0, whereas (at least for
CID-keyed fonts) they should be computed from the FontBBox. (fix from
igor@artifex.com) (int.mak [1.37.2.2], lib.mak [1.36.2.1]: 2000/11/09
21:56:58 rayjj)
	- cvs2hist was generating invalid anchor names and not escaping
special charaters. also enable '-r' commandline option, and pass '-b' if
it's not specified for default branch logs only. Changes.htm now passes the
validator. Minor url cleanup in Release.htm (toolbin/cvs2hist.py [1.4]:
2000/11/10 01:52:35 giles)

Collect some stray configuration related externs into iconf.h
(int.mak [1.40]: 2000/11/04 06:12:02 rayjj)

Adds DSC parsing capability to the PostScript interpreter, with an optional
'feature' to use the DSC Orientation comments to compensate for sloppy
PostScript producers that don't emit an appropriate setpagedevice command.
(int.mak [1.41]: 2000/11/05 18:33:53 lpd)

Adds the 'hook' for passing DSC comments to the PDF writer, but doesn't
actually process any of the comments. NOTE: Part of this code was
accidentally included in the recent fix for local/global VM problems in the
PostScript code for using the DSC parser. (devs.mak [1.36]: 2000/11/07
18:54:26 lpd)

Collect some stray configuration related externs into iconf.h
(int.mak [1.37.2.3]: 2000/11/09 22:39:35 rayjj)

Adds DSC parsing capability to the PostScript interpreter, with an optional
'feature' to use the DSC Orientation comments to compensate for sloppy
PostScript producers that don't emit an appropriate setpagedevice command.
(int.mak [1.37.2.4]: 2000/11/09 23:53:36 rayjj)

Reorganizes some text-writing code in anticipation of adding CIDFont and
CMap support. (devs.mak [1.37]: 2000/11/17 22:11:44 lpd)

Avoids exporting empty LD_RUN_PATH environment variable to the link process.
This was causing binaries to search the current directory for libraries,
which was a security hole. Fixes SourceForge bug 122532.
(ugcclib.mak [1.5.2.1], unixlink.mak [1.2.2.1]: 2000/11/28 00:14:34 raph)

Avoids exporting empty LD_RUN_PATH environment variable to the link process.
This was causing binaries to search the current directory for libraries,
which was a security hole. Fixes SourceForge bug 122532. (ugcclib.mak [1.6],
unixlink.mak [1.3]: 2000/11/28 00:14:48 raph)

Updates version number to 6.32 in preparation for release.
(version.mak [1.22.2.2]: 2000/11/28 01:31:27 raph)

Utilities


Fixes problems:
	- A { } was omitted around an arm of a conditional, causing a
compilation error. (fix by ghost@aladdin.com corresponds to rev. 1.5)
(gp_vms.c [1.4.2.1]: 2000/11/09 21:13:04 rayjj)
	- The simple platform benchmarking program had gotten out of date,
and would no longer compile and link. (fix by ghost@aladdin.com corresponds
to rev. 1.4) (bench.c [1.3.2.1]: 2000/11/10 00:04:02 rayjj)

Removes some old embryonic DSC parser code, since a complete third-party DSC
parser is nearly ready for use. (corresponds to rev 1.2 action by
ghost@aladdin.com) (gsdsc.c [1.1.2.1]: 2000/11/09 23:41:28 rayjj)

Drivers


Fixes problems:
	- The margin settings for the Color LaserJet were incorrect. (fix
from henrys@artifex.com) (gdevcljc.c [1.2.2.1]: 2000/11/09 20:31:45 rayjj)
	- PDF patterns and masks could be compressed with the DCT filter,
causing possible color distortion and loss of detail. (fix by
ghost@aladdin.com) (gdevpsdf.h [1.8.2.1], gdevpsdi.c [1.12.2.1]: 2000/11/09
23:24:18 rayjj)
	- In PDF output, color values of 1.0 were usually written as 0.999.
(bug introduced in version 6.22) (gdevpsdu.c [1.8]: 2000/11/12 07:13:26 lpd)
	- In PDF output, color values of 1.0 were usually written as 0.999.
(bug introduced in version 6.22) (by ghost@aladdin.com corresponds to rev
1.8) (gdevpsdu.c [1.7.2.1]: 2000/11/13 05:28:54 rayjj)
	- Indirect area for StripByteCounts and StripOffsets was put in file
even if there was only one strip so the value was direct.
(gdevtifs.c [1.3.2.1]: 2000/11/16 01:39:55 rayjj)
	- Invalid (uninitialized) StripByteCount was being left in file for
the tiff12nc and tiff24nc devices. Needed to call gdev_tiff_end_strip().
(gdevtfnx.c [1.3.2.1]: 2000/11/16 01:41:49 rayjj)

PDF writer


Fixes problems:
	- kshow caused the PDF writer to revert to bitmapped fonts. Fixes
SourceForge bug # 104118. (gdevpdft.c [1.21]: 2000/11/06 04:17:14 lpd)
	- Writing a Function with subfunctions (Functions array) could
produce garbage output, because the subfunctions could be freed before the
containing Function was written. (fix by ghost@aladdin.com corresponds to
rev. 1.22) (gdevpdf.c [1.21.2.1]: 2000/11/09 20:34:17 rayjj)
	- The PDF writer could write embedded Type 1 fonts with lenIV = -1,
which Acrobat Reader 3 accepts but Acrobat Reader 4 doesn't. (fix by
ghost@aladdin.com) (gdevpdfe.c [1.4.2.1]: 2000/11/09 20:37:29 rayjj)
	- Incrementally downloaded Type 1 fonts weren't embedded properly in
PDF output. (This might have been the case for some TrueType fonts as well,
although we never observed it.) 9fix by ghost@aladdin.com corresponds to
rev. 1.9) (gdevpdff.c [1.8.2.1]: 2000/11/09 20:40:29 rayjj)
	- The algorithm for reducing the resolution of PDF output to deal
with Acrobat Reader's 32K limit on user space coordinates was too
conservative (reduced the resolution more than necessary). (fix by
ghost@aladdin.com corresponds to rev. 1.12) (gdevpdfp.c [1.10.2.2]:
2000/11/09 23:19:36 rayjj)
	- Writing a shading with ShadingType >= 4 and data given by a stream
rather than an array could cause a crash. (fix by ghost@aladdin.com
corresponds to rev 1.4) (gdevpdfv.c [1.3.2.1]: 2000/11/09 23:21:47 rayjj)
	- PDF patterns and masks could be compressed with the DCT filter,
causing possible color distortion and loss of detail. (fix by
ghost@aladdin.com) (gdevpdfb.c [1.2.2.1], gdevpdfv.c [1.3.2.2]: 2000/11/09
23:24:18 rayjj)
	- The PDF writer didn't handle stroked fonts correctly (they were
filled instead of stroked). (gdevpdf.c [1.23], gdevpdfd.c [1.11],
gdevpdft.c [1.22], gdevpdfx.h [1.23]: 2000/11/12 06:52:08 lpd)
	- When mixing filled and stroked fonts, the color wasn't set
correctly. (gdevpdft.c [1.23]: 2000/11/12 07:14:30 lpd)
	- The PDF writer incorrectly assumed that in PDF 1.3, the standard
Latin character set was the PostScript 3 extended set (it's actually the old
PostScript Level 2 set). (gdevpdft.c [1.28]: 2000/11/20 18:13:55 lpd)

Adds the 'hook' for passing DSC comments to the PDF writer, but doesn't
actually process any of the comments. NOTE: Part of this code was
accidentally included in the recent fix for local/global VM problems in the
PostScript code for using the DSC parser. (gdevpdfp.c [1.13]: 2000/11/07
18:54:26 lpd)

Updates the list of open PDF writer tasks to reflect recent progress.
(change by ghost@aladdin.com corresponds to rev 1.11)
(gdevpdfp.c [1.10.2.1]: 2000/11/09 23:17:37 rayjj)

Changes the internal indication of whether to allow font subsetting from a
Boolean to OK, YES, NO. (Unfortunately, I no longer remember why this change
is needed.) (gdevpdfe.c [1.6], gdevpdff.c [1.10], gdevpdff.h [1.5],
gdevpdft.c [1.24]: 2000/11/14 06:57:21 lpd)

When writing PDF output, adds the ability to convert text in Type 0 fonts to
text in the descendant leaf fonts. This completes the handling of text
output except for CID-keyed fonts. (gdevpdft.c [1.25]: 2000/11/14 16:47:09
lpd)

Reorganizes some text-writing code in anticipation of adding CIDFont and
CMap support. (gdevpdff.c [1.11], gdevpdff.h [1.6], gdevpdft.c [1.26]:
2000/11/17 22:11:44 lpd)

Makes glyphshow no longer revert to bitmapped fonts. Also adds a CIDFont
pseudo-resource, in anticipation of adding CID-keyed font support.
(gdevpdff.c [1.12], gdevpdff.h [1.7], gdevpdft.c [1.27], gdevpdfx.h [1.24]:
2000/11/20 08:31:20 lpd)

Interpreter (PDF)


Fixes problems:
	- The PDF interpreter caused an error if the output device wasn't a
page device. (fix by ghost@aladdin.com corresponds to rev. 1.22)
(lib/pdf_main.ps [1.19.2.3]: 2000/11/09 19:21:48 rayjj)

Improves font substitution by taking the "Narrow" property into account.
Fixes SourceForge bug 116460. Fix due to Leon Bottou.
(lib/pdf_font.ps [1.12.2.1]: 2000/11/05 17:35:46 raph)

Improves font substitution by taking the "Narrow" property into account.
Fixes SourceForge bug 116460. Fix due to Leon Bottou.
(lib/pdf_font.ps [1.13]: 2000/11/05 17:36:28 raph)

Wraps the showpagecontents call in gsave/grestore, so that the graphics
state for drawing annotations is pristine. Fixes Sourceforge bug 123310.
(lib/pdf_main.ps [1.19.2.4]: 2000/11/25 21:58:19 raph)

Wraps the showpagecontents call in gsave/grestore, so that the graphics
state for drawing annotations is pristine. Fixes Sourceforge bug 123310.
(lib/pdf_main.ps [1.23]: 2000/11/25 22:00:21 raph)

Interpreter


Fixes problems:
	- The recently updated code for activating the DSC comment parser
had several local/global VM problems, causing invalidaccess errors.
(lib/gs_dps2.ps [1.3], lib/gs_init.ps [1.27], lib/gs_pdfwr.ps [1.7]:
2000/11/07 18:51:24 lpd)
	- There was still a bug in the code for passing DSC comments to the
PDF writer, causing typecheck errors. (lib/gs_pdfwr.ps [1.8]: 2000/11/07
19:27:14 lpd)
	- If an error occurred in a page device Install procedure, the
stacks weren't restored properly, possibly leading to other errors. (fix by
ghost@aladdin.com corresponds to rev. 1.6) (lib/gs_setpd.ps [1.5.2.1]:
2000/11/09 18:46:55 rayjj)
	- resourceforall was a procedure rather than an operator. Also adds
an explicit acknowledgement that the method used for this fix is a hack.
(fix by ghost@aladdin.com) (lib/gs_res.ps [1.13.2.1],
lib/gs_setpd.ps [1.5.2.2]: 2000/11/09 18:54:57 rayjj)
	- CIDFontType 2 fonts with a GlyphDirectory and string or string
array CIDMap ignored GDBytes, causing an arithmetic error (division by
zero). (SourceForge bug # 115942) (fix by ghost@aladdin.com)
(zfcid.c [1.10.2.1], zfcid1.c [1.2.2.1]: 2000/11/09 23:04:44 rayjj)
	- Most of the procedures in this file didn't use 'bind' (performance
bug only). Fixing this sped up interpretation of PLRM.pdf by 40%! (fix by
ghost@aladdin.com corresponds to rev 1.8) (lib/gs_cff.ps [1.7.2.1]:
2000/11/09 23:08:47 rayjj)
	- If the procedure of a FunctionType 4 Function was syntactically
valid but exceeded the maximum stack depth, an attempt was made to free a
string as an object, causing memory corruption. (zfunc4.c [1.6]: 2000/11/11
17:34:06 lpd)
	- If the procedure of a FunctionType 4 Function was syntactically
valid but exceeded the maximum stack depth, an attempt was made to free a
string as an object, causing memory corruption. (fix by ghost@aladdin.com
corresponds to rev. 1.6) (zfunc4.c [1.5.2.1]: 2000/11/12 21:33:46 rayjj)
	- Disk-based TrueType fonts caused a memory leak. (Currently, there
are no such fonts, but CIDFontType 2 fonts, and directly-accessed TrueType
fonts, should be made disk-based.) (zfcid1.c [1.4]: 2000/11/22 08:32:07 lpd)
	- Type 1 CharStrings loaded from disk for type 0 CIDFonts were never
freed, except by garbage collection (which doesn't reclaim chunks with only
a small amount of surviving data). In practice, this probably only affects
the PDF writer. (zfcid0.c [1.5]: 2000/11/23 23:34:22 lpd)

Collect some stray configuration related externs into iconf.h
(imain.c [1.3], imainarg.c [1.5]: 2000/11/04 06:12:02 rayjj)

branches: 1.1.2; Collect some stray configuration related externs into
iconf.h (iconf.h [1.1]: 2000/11/04 06:12:02 rayjj)

Add COMPILE_INITS information line to -h help output. (imainarg.c [1.6]:
2000/11/04 06:13:42 rayjj)

Adds DSC parsing capability to the PostScript interpreter, with an optional
'feature' to use the DSC Orientation comments to compensate for sloppy
PostScript producers that don't emit an appropriate setpagedevice command.
(lib/gs_dscp.ps [1.1], zdscpars.c [1.1]: 2000/11/05 18:33:52 lpd)

branches: 1.1.2; Adds DSC parsing capability to the PostScript interpreter,
with an optional 'feature' to use the DSC Orientation comments to compensate
for sloppy PostScript producers that don't emit an appropriate setpagedevice
command. (dscparse.c [1.1], dscparse.h [1.1]: 2000/11/05 18:33:53 lpd)

Polishes the DSC parser interface: improves comments and formatting, and
fixes a couple of very minor glitches. (zdscpars.c [1.2]: 2000/11/06
07:34:21 lpd)

branches: 1.3.2; ... and fixes an overlooked const discrepancy.
(zdscpars.c [1.3]: 2000/11/06 07:35:51 lpd)

Add to GS_6_5 branch. (zdscpars.c [1.3.2.1]: 2000/11/07 18:26:52 rayjj)

branches: 1.2.2; Fix: The recently updated code for activating the DSC
comment parser had several local/global VM problems, causing invalidaccess
errors. (lib/gs_dscp.ps [1.2]: 2000/11/07 18:51:25 lpd)

Return code variables (code and comment_code) in zparse_dsc_comments(). need
to be signed. (zdscpars.c [1.4]: 2000/11/08 00:54:48 dancoby)

Return code variables (code and comment_code) in zparse_dsc_comments(). need
to be signed. (zdscpars.c [1.3.2.2]: 2000/11/08 01:25:58 dancoby)

Rewrites .parse_dsc_comment to use the parameter list machinery, removing a
little functionally duplicated code; changes two incorrect uints to ints
(fix from dan@artifex.com); adds additional comment types (BoundingBox,
PageBoundingBox, BeginDefaults, EndDefaults) to the interface.
(zdscpars.c [1.5]: 2000/11/08 07:07:32 lpd)

Collect some stray configuration related externs into iconf.h
(iconf.h [1.1.2.1], imain.c [1.2.2.1], imainarg.c [1.4.2.1]: 2000/11/09
22:39:35 rayjj)

Add COMPILE_INITS information line to -h help output. (imainarg.c [1.4.2.2]:
2000/11/09 22:45:17 rayjj)

Speeds up loading embedded CFF fonts some more, by adding some missing
'bind's and similar tactics. (fix by ghost@aladdin.com corresponds to rev
1.9) (lib/gs_cff.ps [1.7.2.2]: 2000/11/09 23:11:33 rayjj)

Rewrites .parse_dsc_comment to use the parameter list machinery, removing a
little functionally duplicated code; changes two incorrect uints to ints
(fix from dan@artifex.com); adds additional comment types (BoundingBox,
PageBoundingBox, BeginDefaults, EndDefaults) to the interface. (fix by
ghost@aladdin.com corresponds to rev. 1.5) (zdscpars.c [1.3.2.3]: 2000/11/09
23:43:39 rayjj)

Add to GS_6_5 branch (lib/gs_dscp.ps [1.2.2.1]: 2000/11/09 23:47:04 rayjj)

Adds DSC parsing capability to the PostScript interpreter, with an optional
'feature' to use the DSC Orientation comments to compensate for sloppy
PostScript producers that don't emit an appropriate setpagedevice command.
(dscparse.c [1.1.2.1], dscparse.h [1.1.2.1]: 2000/11/09 23:53:36 rayjj)

DSC parser changes: adds %%ViewerOrientation to DSC parser; doesn't flag an
error if %%Pages is absent but one %%Page is found; frees memory if
initialization fails; recognizes general whitespace in some places where
only spaces were recognized before. (changes from gsview@ghostgum.com.au)
(dscparse.c [1.2], dscparse.h [1.2]: 2000/11/13 05:44:58 lpd)

DSC parser changes: adds %%ViewerOrientation to DSC parser; doesn't flag an
error if %%Pages is absent but one %%Page is found; frees memory if
initialization fails; recognizes general whitespace in some places where
only spaces were recognized before. (changes from gsview@ghostgum.com.au)
(fix corresponds to rev 1.2 of both files). (dscparse.c [1.1.2.2],
dscparse.h [1.1.2.2]: 2000/11/13 06:01:30 rayjj)

Adds DSC comment tracing to ps2pdf, if -dDEBUG is specified on the command
line. (lib/gs_pdfwr.ps [1.9]: 2000/11/13 17:03:51 lpd)

Adds %%EndDefaults, %%For:, and %%ViewerOrientation: to the set of DSC
comments handled by the .parse_dsc_comments operator. (zdscpars.c [1.6]:
2000/11/13 17:06:18 lpd)

Slightly simplifies the code for creating the internal representation of
CMaps. (lib/gs_cmap.ps [1.8]: 2000/11/16 00:32:00 lpd)

Changes .buildshading to always apply a ReusableStreamDecode filter to file
and stream DataSources, so that multiple passes over the stream work. This
patch closes SourceForge bug 121810. (lib/gs_ll3.ps [1.6.2.1]: 2000/11/16
00:37:03 raph)

Changes .buildshading to always apply a ReusableStreamDecode filter to file
and stream DataSources, so that multiple passes over the stream work. This
patch closes SourceForge bug 121810. (lib/gs_ll3.ps [1.7]: 2000/11/16
00:37:32 raph)

Updates version number to 6.32 in preparation for release.
(lib/gs_init.ps [1.25.2.2]: 2000/11/28 01:31:26 raph)

Streams


Fixes problems:
	- The ASCII85Decode filter sometimes asked for too much output
space, which caused a limitcheck error on ASCII85 strings whose decoded size
was 65533 - 65535 bytes. (fix by ghost@aladdin.com corresponds to rev 1.3)
(sa85d.c [1.2.2.1]: 2000/11/09 22:57:10 rayjj)

Library


Fixes problems:
	- Several places in the band list code didn't call s_init_state to
initialize locally allocated stream states: this is now required (see the
"Incompatible changes" section above). We suspect there are a few more
places that will need this fix. (gxclbits.c [1.3], gxcldev.h [1.4],
gxclread.c [1.3], gxclutil.c [1.3]: 2000/11/05 00:34:24 lpd)
	- Fixes infinite loops when writing large bitmaps into clist. Now,
cmd_put_bits never requests an allocation in the clist buffer larger than
cbuf_size. (gxclbits.c [1.4]: 2000/11/05 18:44:57 raph)
	- Fixes infinite loops when writing large bitmaps into clist. Now,
cmd_put_bits never requests an allocation in the clist buffer larger than
cbuf_size. (gxclbits.c [1.2.2.1]: 2000/11/05 18:47:27 raph)
	- Works around an Acrobat Reader 4 bug that causes AR4 to draw a
single pixel, rather than nothing, when filling a subpath consisting only of
a moveto. (gdevvec.c [1.8]: 2000/11/07 16:42:28 lpd)
	- The PDF writer could write embedded Type 1 fonts with lenIV = -1,
which Acrobat Reader 3 accepts but Acrobat Reader 4 doesn't. (fix by
ghost@aladdin.com) (gdevpsf.h [1.7.2.1], gdevpsf1.c [1.5.2.1]: 2000/11/09
20:37:29 rayjj)
	- Sampling the spot function for a halftone didn't correctly sample
at the centers of device space pixels, causing moire' effects and irregular
variations in spot size. (fix from igor@artifex.com corresponds to 1.4)
(gshtscr.c [1.3.2.1], gzht.h [1.3.2.1]: 2000/11/09 21:26:51 rayjj)
	- Getting the glyph_info of a Type 2 CharString could cause an
invalid memory access. (SourceForge bug # 121070) (fix by ghost@aladdin.com
corresponds to rev 1.8) (gstype2.c [1.7.2.1]: 2000/11/09 21:28:19 rayjj)
	- TrueType glyphs with more than 200 points caused a crash or
corrupted memory. (fix by ghost@aladdin.com corresponds to rev 1.8)
(gstype42.c [1.7.2.1]: 2000/11/09 21:30:39 rayjj)
	- Type 42 / TrueType fonts didn't clear the pointers to the metrics
tables in the font, possibly producing incorrect metrics or memory access
errors for fonts that lack one or more of hhea, hmtx, vhea, or vmtx. (fix
from henrys@artifex.com corresponds to rev. 1.9) (gstype42.c [1.7.2.2]:
2000/11/09 21:32:30 rayjj)
	- Works around an Acrobat Reader 4 bug that causes AR4 to draw a
single pixel, rather than nothing, when filling a subpath consisting only of
a moveto. (fix by ghost@aladdin.com corresponds to rev. 1.8)
(gdevvec.c [1.7.2.1]: 2000/11/09 23:25:38 rayjj)
	- Several places in the band list code didn't call s_init_state to
initialize locally allocated stream states: this is now required (see the
"Incompatible changes" section). (fix by ghost@aladdin.com)
(gxclbits.c [1.2.2.2], gxcldev.h [1.3.2.1], gxclread.c [1.2.2.1],
gxclutil.c [1.2.2.1]: 2000/11/09 23:36:53 rayjj)
	- The PDF writer didn't handle stroked fonts correctly (they were
filled instead of stroked). (gdevvec.c [1.9], gdevvec.h [1.5]: 2000/11/12
06:52:08 lpd)
	- The pdfwrite and pswrite drivers could emit a lineto or rlineto
without a moveto at the beginning of the path. (bug introduced very
recently, by a workaround for an Acrobat Reader bug) (gdevvec.c [1.10]:
2000/11/14 07:41:08 lpd)
	- The pdfwrite and pswrite drivers could emit a lineto or rlineto
without a moveto at the beginning of the path. (bug introduced very
recently, by a workaround for an Acrobat Reader bug) (fix by
ghost@aladdin.com corresponds to rev. 1.9 to rev 1.10 change)
(gdevvec.c [1.7.2.2]: 2000/11/14 17:20:05 rayjj)
	- Attempting to write a CFF Type 2 font larger than 32K but smaller
than 64K caused a rangecheck. (gdevpsf2.c [1.8]: 2000/11/19 17:48:08 lpd)
	- Attempting to write a CFF Type 2 font larger than 32K but smaller
than 64K caused a rangecheck. (fix by ghost@aladdin.com corresponds to rev.
1.8) (gdevpsf2.c [1.7.2.1]: 2000/11/19 19:03:19 rayjj)
	- Computing the default width or the fixed width for a CIDFont or
TrueType font could do a huge amount of unnecessary work. (Probably only
affects the PDF writer.) (gsfont.c [1.3]: 2000/11/20 21:09:09 lpd)
	- Computing the default width or the fixed width for a CIDFont or
TrueType font could do a huge amount of unnecessary work. (Probably only
affects the PDF writer.) (fix by ghost@aladdin.com corresponds to rev 1.3)
(gsfont.c [1.2.2.1]: 2000/11/21 00:20:28 rayjj)
	- The Type 1 to Type 2 CharString converter (currently used only by
the PDF writer) often didn't detect invalid CharStrings as input.
(gdevpsfx.c [1.6]: 2000/11/21 02:33:55 lpd)
	- The code for writing a CIDFontType0 CFF font mistook Type 2
CharStrings for Type 1, and tried to convert them to Type 2. (This currently
affects only the PDF writer.) (gdevpsf2.c [1.9]: 2000/11/21 02:35:59 lpd)
	- The code for writing a CIDFontType0 CFF font mistook Type 2
CharStrings for Type 1, and tried to convert them to Type 2. (This currently
affects only the PDF writer.) (fix by ghost@aladdin.com corresponds to rev
1.9) (gdevpsf2.c [1.7.2.2]: 2000/11/21 05:46:25 rayjj)
	- The Type 1 to Type 2 CharString converter (currently used only by
the PDF writer) often didn't detect invalid CharStrings as input. (fix by
ghost@aladdin.com corresponds to rev 1.6) (gdevpsfx.c [1.5.2.1]: 2000/11/21
05:52:15 rayjj)
	- Disk-based TrueType fonts caused a memory leak. (Currently, there
are no such fonts, but CIDFontType 2 fonts, and directly-accessed TrueType
fonts, should be made disk-based.) (gdevpsft.c [1.5], gstype42.c [1.10]:
2000/11/22 08:32:07 lpd)
	- Type 1 CharStrings loaded from disk for type 0 CIDFonts were never
freed, except by garbage collection (which doesn't reclaim chunks with only
a small amount of surviving data). In practice, this probably only affects
the PDF writer. (gsfcid.c [1.5], gstype1.c [1.6], gstype2.c [1.9],
gxfont1.h [1.5], gxtype1.c [1.8], gxtype1.h [1.5]: 2000/11/23 23:34:22 lpd)

Improves the documentation of the minimum responsibilities of the 'process'
procedure in a text enumerator. (gxtext.h [1.6]: 2000/11/06 00:44:37 lpd)

Adds a param_string_from_transient_string macro, to parallel
param_string_from_string for non-permanent C strings. (gsparam.h [1.8]:
2000/11/08 06:56:50 lpd)

Works around a bug in the DEC VMS C compiler, which incorrectly defines the
second argument of fdopen as (char *) rather than (const char *). (fix by
ghost@aladdin.com corresponds to rev. 1.8) (gpmisc.c [1.7.2.1]: 2000/11/09
21:18:00 rayjj)

Adds a macro for declaring suffix subclasses with 5 added pointers (in
addition to the existing 1, 2, 3, 4, and 6). (gsstruct.h [1.6]: 2000/11/21
16:46:55 lpd)

Adds a procedure for testing whether a glyph is a "notdef" in a font,
necessary for fixing a memory consumption problem when accessing type 0
CIDFonts. (gsfont.c [1.4], gxfont.h [1.3]: 2000/11/23 23:23:40 lpd)

Speeds up the slow case of fixed_mult_quo by nearly a factor of 2, by using
a tiny bit of numerical analysis. (gsmisc.c [1.9], gxfixed.h [1.3]:
2000/11/24 06:53:23 lpd)

Improves the performance of embedded font writing by freeing strings when no
longer needed, and by removing redundant computation. (Currently relevant
only to the PDF writer.) (gdevpsf1.c [1.7], gdevpsf2.c [1.10],
gdevpsfu.c [1.5], gdevpsfx.c [1.7]: 2000/11/24 07:10:41 lpd)

Adds a procedure for creating an Identity CMap, needed by the PDF writer.
(gsfcmap.c [1.11], gxfcmap.h [1.6]: 2000/11/26 04:08:41 lpd)