Update changelog.

Update documentation.
Fix missing ivlen initializer for 40-bit RC4 (Issue #51 )
2025-07-13 06:24:25 +02:00 · 2023-11-15 10:14:02 -05:00 · 2023-11-15 09:28:57 -05:00 · 2023-11-15 08:43:07 -05:00 · 2023-11-15 08:38:47 -05:00 · 2023-11-14 18:38:26 -05:00
37 changed files with 1798 additions and 304 deletions
--- a/.github/codeql.yml
+++ b/.github/codeql.yml
@ -0,0 +1,22 @@
+paths-ignore:
+  - testpdfio.c
+
+query-filters:
+  - exclude:
+      id: cpp/commented-out-code
+  - exclude:
+      id: cpp/toctou-race-condition
+  - exclude:
+      id: cpp/weak-cryptographic-algorithm
+  - exclude:
+      id: cpp/world-writable-file-creation
+  - exclude:
+      id: cpp/uncontrolled-allocation-size
+  - exclude:
+      id: cpp/path-injection
+  - exclude:
+      id: cpp/stack-address-escape
+  - exclude:
+      id: cpp/loop-variable-changed
+  - exclude:
+      id: cpp/long-switch
--- a/.github/workflows/codeql.yml
+++ b/.github/workflows/codeql.yml
@ -0,0 +1,50 @@
+name: "CodeQL"
+
+on:
+  push:
+    branches: [ "master" ]
+  pull_request:
+    branches: [ "master" ]
+  schedule:
+    - cron: "46 3 * * 0"
+
+jobs:
+  analyze:
+    name: Analyze
+    runs-on: ubuntu-latest
+    permissions:
+      actions: read
+      contents: read
+      security-events: write
+
+    strategy:
+      fail-fast: false
+      matrix:
+        language: [ cpp ]
+
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v3
+        with:
+          submodules: recursive
+
+      - name: Update build environment
+        run: sudo apt-get update --fix-missing -y
+
+      - name: Install prerequisites
+        run: sudo apt-get install -y zlib1g-dev
+
+      - name: Initialize CodeQL
+        uses: github/codeql-action/init@v2
+        with:
+          languages: ${{ matrix.language }}
+          config-file: ./.github/codeql.yml
+          queries: +security-and-quality
+
+      - name: Autobuild
+        uses: github/codeql-action/autobuild@v2
+
+      - name: Perform CodeQL Analysis
+        uses: github/codeql-action/analyze@v2
+        with:
+          category: "/language:${{ matrix.language }}"
--- a/.gitignore
+++ b/.gitignore
@ -8,6 +8,10 @@
 /doc/pdfio.epub
 /packages
 /pdfio.xcodeproj/xcshareddata
+/pdfio-*.tar.gz*
+/pdfio-*.zip*
+/pdfiototext
 /testpdfio
 /testpdfio-*.pdf
+/testttf
 /x64
--- a/.lgtm.yml
+++ b/.lgtm.yml
@ -1,2 +0,0 @@
-queries:
-  - exclude: cpp/toctou-race-condition
--- a/CHANGES.md
+++ b/CHANGES.md
@ -2,8 +2,67 @@ Changes in PDFio
 ================


-v1.0rc1 (Month DD, YYYY)
------------------------
+v1.1.3 (November 15, 2023)
+--------------------------
+
+- Fixed Unicode font support (Issue #16)
+- Fixed missing initializer for 40-bit RC4 encryption (Issue #51)
+
+
+v1.1.2 (October 10, 2023)
+-------------------------
+
+- Updated `pdfioContentSetDashPattern` to support setting a solid (0 length)
+  dash pattern (Issue #41)
+- Fixed an issue with broken PDF files containing extra CR and/or LF separators
+  after the object stream token (Issue #40)
+- Fixed an issue with PDF files produced by Crystal Reports (Issue #45)
+- Fixed an issue with PDF files produced by Microsoft Reporting Services
+  (Issue #46)
+- Fixed support for compound filters where the filter array consists of a
+  single named filter (Issue #47)
+- Fixed builds on Windows - needed windows.h header for temporary files
+  (Issue #48)
+
+
+v1.1.1 (March 20, 2023)
+-----------------------
+
+- CVE-2023-28428: Fixed a potential denial-of-service with corrupt PDF files.
+- Fixed a few build issues.
+
+
+v1.1.0 (February 6, 2023)
+-------------------------
+
+- CVE-2023-24808: Fixed a potential denial-of-service with corrupt PDF files.
+- Added `pdfioFileCreateTemporary` function (Issue #29)
+- Added `pdfioDictIterateKeys` function (Issue #31)
+- Added `pdfioContentPathEnd` function.
+- Added protection against opening multiple streams in the same file at the
+  same time.
+- Documentation updates (Issue #37)
+- Fixed "install-shared" target (Issue #32)
+- Fixed `pdfioFileGet...` metadata APIs (Issue #33)
+- Fixed `pdfioContentMatrixRotate` function.
+
+
+v1.0.1 (March 2, 2022)
+----------------------
+
+- Added missing `pdfioPageGetNumStreams` and `pdfioPageOpenStream` functions.
+- Added demo pdfiototext utility.
+- Fixed bug in `pdfioStreamGetToken`.
+
+
+v1.0.0 (December 14, 2021)
+--------------------------
+
+- First stable release.
+
+
+v1.0rc1 (November 30, 2021)
+---------------------------

 - Fixed a few stack/buffer overflow bugs discovered via fuzzing.

--- a/CONTRIBUTING.md
+++ b/CONTRIBUTING.md
@ -118,7 +118,7 @@ the source file and the copyright and licensing notice:
    //
    // Description of file contents.
    //
-    // Copyright YYYY by AUTHOR.
+    // Copyright © YYYY by AUTHOR.
    //
    // Licensed under Apache License v2.0.  See the file "LICENSE" for more
    // information.
@ -330,7 +330,7 @@ typedef, for example:

 All constant names are uppercase with underscores between words, e.g.,
 `PDFIO_THIS_CONSTANT`, `PDFIO_THAT_CONSTANT`, etc.  Constants begin with the
-"PDFio\_" prefix to avoid conflicts with system constants.  Private constants
+"PDFIO\_" prefix to avoid conflicts with system constants.  Private constants
 start with an underscore, e.g., `_PDFIO_THIS_CONSTANT`,
 `_PDFIO_THAT_CONSTANT`, etc.

@ -369,11 +369,12 @@ extensions MUST NOT be used.
 The following variables are defined in the makefile:

 - `AR`; the static library archiver command,
- `ARFLAGS`; options for the static library archiver command,
+- `ARFLAGS`; options for the static library archiver,
 - `CC`; the C compiler command,
- `CFLAGS`; options for the C compiler command,
+- `CFLAGS`; options for the C compiler,
 - `CODESIGN_IDENTITY`: the code signing identity,
 - `COMMONFLAGS`; common compiler optimization options,
+- `CPPFLAGS`; options for the C preprocessor,
 - `DESTDIR`/`DSTROOT`: the destination root directory when installing.
 - `DSO`; the shared library building command,
 - `DSOFLAGS`; options for the shared library building command,
@ -395,4 +396,5 @@ The following standard targets are defined in the makefile:
  with debug printfs and the clang address sanitizer enabled.
 - `install`; installs all distribution files in their corresponding locations.
 - `install-shared`; same as `install` but also installs the shared library.
+- `macos`; same as `all` but creates a Universal Binary (X64 + ARM64).
 - `test`; runs the unit test program, building it as needed.
--- a/52
+++ b/52
@ -1,7 +1,7 @@
 #
 # Makefile for PDFio.
 #
-# Copyright © 2021 by Michael R Sweet.
+# Copyright © 2021-2023 by Michael R Sweet.
 #
 # Licensed under Apache License v2.0.  See the file "LICENSE" for more
 # information.
@ -10,7 +10,10 @@
 # POSIX makefile
 .POSIX:

-# Variables...
+# Build silently
+.SILENT:
+
+# Variables
 AR		=	ar
 ARFLAGS		=	cr
 CC		=	cc
@ -18,7 +21,7 @@ CFLAGS		=
 CODESIGN_IDENTITY =	Developer ID
 COMMONFLAGS	=	-Os -g
 #COMMONFLAGS	=	-O0 -g -fsanitize=address
-CPPFLAGS	=	'-DPDFIO_VERSION="$(VERSION)"'
+CPPFLAGS	=
 DESTDIR		=	$(DSTROOT)
 DSO		=	cc
 DSOFLAGS	=
@ -26,13 +29,14 @@ DSONAME		=
 LDFLAGS		=
 LIBS		=	-lm -lz
 RANLIB		=	ranlib
-VERSION		=	1.0b2
+VERSION		=	1.1.3
 prefix		=	/usr/local


 # Base rules
 .SUFFIXES:	.c .h .o
 .c.o:
+	echo Compiling $<...
 	$(CC) $(CFLAGS) $(CPPFLAGS) $(COMMONFLAGS) -c $<


@ -62,11 +66,15 @@ LIBOBJS		=	\
 			ttf.o
 OBJS		=	\
 			$(LIBOBJS) \
-			testpdfio.o
+			pdfiototext.o \
+			testpdfio.o \
+			testttf.o
 TARGETS		=	\
 			$(DSONAME) \
 			libpdfio.a \
-			testpdfio
+			pdfiototext \
+			testpdfio \
+			testttf


 # Make everything
@ -82,6 +90,9 @@ all-shared:
 debug:
 	$(MAKE) -$(MAKEFLAGS) COMMONFLAGS="-g -fsanitize=address -DDEBUG=1" clean all

+macos:
+	$(MAKE) -$(MAKEFLAGS) COMMONFLAGS="-Os -mmacosx-version-min=11 -arch x86_64 -arch arm64" clean all
+

 # Clean everything
 clean:
@ -90,8 +101,10 @@ clean:

 # Install everything
 install:	$(TARGETS)
+	echo Installing header files to $(DESTDIR)$(prefix)/include...
 	-mkdir -p $(DESTDIR)$(prefix)/include
 	cp $(PUBHEADERS) $(DESTDIR)$(prefix)/include
+	echo Installing library files to $(DESTDIR)$(prefix)/lib...
 	-mkdir -p $(DESTDIR)$(prefix)/lib
 	cp libpdfio.a $(DESTDIR)$(prefix)/lib
 	$(RANLIB) $(DESTDIR)$(prefix)/lib/libpdfio.a
@ -103,26 +116,30 @@ install:	$(TARGETS)
 		codesign -s "$(CODESIGN_IDENTITY)" -o runtime --timestamp $(DESTDIR)$(prefix)/lib/libpdfio.1.dylib; \
 		ln -sf libpdfio.1.dylib $(DESTDIR)$(prefix)/lib/libpdfio.dylib; \
 	fi
+	echo Installing pkg-config files to $(DESTDIR)$(prefix)/lib/pkgconfig...
 	-mkdir -p $(DESTDIR)$(prefix)/lib/pkgconfig
 	echo 'prefix="$(prefix)"' >$(DESTDIR)$(prefix)/lib/pkgconfig/pdfio.pc
 	echo 'Version: $(VERSION)' >>$(DESTDIR)$(prefix)/lib/pkgconfig/pdfio.pc
 	cat pdfio.pc.in >>$(DESTDIR)$(prefix)/lib/pkgconfig/pdfio.pc
+	echo Installing documentation to $(DESTDIR)$(prefix)/share/doc/pdfio...
 	-mkdir -p $(DESTDIR)$(prefix)/share/doc/pdfio
 	cp doc/pdfio.html doc/pdfio-512.png LICENSE NOTICE $(DESTDIR)$(prefix)/share/doc/pdfio
+	echo Installing man page to $(DESTDIR)$(prefix)/share/man/man3...
 	-mkdir -p $(DESTDIR)$(prefix)/share/man/man3
 	cp doc/pdfio.3 $(DESTDIR)$(prefix)/share/man/man3

 install-shared:
 	if test `uname` = Darwin; then \
 		$(MAKE) DSONAME="libpdfio.1.dylib" -$(MAKEFLAGS) install; \
-	else
+	else \
 		$(MAKE) DSONAME="libpdfio.so.1" -$(MAKEFLAGS) install; \
 	fi


 # Test everything
 test:	testpdfio
-	./testpdfio
+	./testttf 2>test.log
+	./testpdfio 2>test.log

 valgrind:	testpdfio
 	valgrind --leak-check=full ./testpdfio
@ -130,13 +147,16 @@ valgrind:	testpdfio

 # pdfio library
 libpdfio.a:		$(LIBOBJS)
+	echo Archiving $@...
 	$(AR) $(ARFLAGS) $@ $(LIBOBJS)
 	$(RANLIB) $@

 libpdfio.so.1:		$(LIBOBJS)
+	echo Linking $@...
 	$(CC) $(DSOFLAGS) $(COMMONFLAGS) -shared -o $@ -Wl,-soname,$@ $(LIBOBJS) $(LIBS)

 libpdfio.1.dylib:	$(LIBOBJS)
+	echo Linking $@...
 	$(CC) $(DSOFLAGS) $(COMMONFLAGS) -dynamiclib -o $@ -install_name $(prefix)/lib/$@ -current_version $(VERSION) -compatibility_version 1.0 $(LIBOBJS) $(LIBS)


@ -154,24 +174,38 @@ pdfio1.def: $(LIBOBJS) Makefile
 		grep -v '^_ttf' | sed -e '1,$$s/^_//' | sort >>$@


+# pdfio text extraction (demo, doesn't handle a lot of things yet)
+pdfiototext:		pdfiototext.o libpdfio.a
+	echo Linking $@...
+	$(CC) $(LDFLAGS) $(COMMONFLAGS) -o $@ pdfiototext.o libpdfio.a $(LIBS)
+
+
 # pdfio test program
 testpdfio:		testpdfio.o libpdfio.a
+	echo Linking $@...
 	$(CC) $(LDFLAGS) $(COMMONFLAGS) -o $@ testpdfio.o libpdfio.a $(LIBS)

+# TTF test program
+testttf:	ttf.o testttf.o
+	echo Linking $@...
+	$(CC) $(LDFLAGS) $(COMMONFLAGS) -o testttf ttf.o testttf.o $(LIBS)
+

 # Dependencies
 $(OBJS):		pdfio.h pdfio-private.h Makefile
 pdfio-content.o:	pdfio-content.h ttf.h
+testttf.o:		ttf.h
 ttf.o:			ttf.h

 # Make documentation using Codedoc <https://www.msweet.org/codedoc>
 DOCFLAGS	=	\
 			--author "Michael R Sweet" \
-			--copyright "Copyright (c) 2021 by Michael R Sweet" \
+			--copyright "Copyright (c) 2021-2023 by Michael R Sweet" \
 			--docversion $(VERSION)

 .PHONY: doc
 doc:
+	echo Generating documentation...
 	codedoc $(DOCFLAGS) --title "PDFio Programming Manual v$(VERSION)" $(PUBHEADERS) $(PUBOBJS:.o=.c) --body doc/pdfio.md --coverimage doc/pdfio-512.png pdfio.xml >doc/pdfio.html
 	codedoc $(DOCFLAGS) --title "PDFio Programming Manual v$(VERSION)" --body doc/pdfio.md --coverimage doc/pdfio-epub.png pdfio.xml --epub doc/pdfio.epub
 	codedoc $(DOCFLAGS) --title "pdf read/write library" --man pdfio --section 3 --body doc/pdfio.md pdfio.xml >doc/pdfio.3
--- a/2
+++ b/2
@ -1,6 +1,6 @@
 PDFio - PDF Read/Write Library

-Copyright © 2021 by Michael R Sweet.
+Copyright © 2021-2023 by Michael R Sweet.

 (Optional) Exceptions to the Apache 2.0 License:
 ================================================
--- a/README.md
+++ b/README.md
@ -3,10 +3,8 @@ pdfio - PDF Read/Write Library

 ![Version](https://img.shields.io/github/v/release/michaelrsweet/pdfio?include_prereleases)
 ![Apache 2.0](https://img.shields.io/github/license/michaelrsweet/pdfio)
-[![Build](https://github.com/michaelrsweet/pdfio/workflows/Build/badge.svg)](https://github.com/michaelrsweet/pdfio/actions/workflows/build.yml)
+[![Build Status](https://img.shields.io/github/workflow/status/michaelrsweet/pdfio/Build)](https://github.com/michaelrsweet/pdfio/actions/workflows/build.yml)
 [![Coverity Scan Status](https://img.shields.io/coverity/scan/22385.svg)](https://scan.coverity.com/projects/michaelrsweet-pdfio)
-[![LGTM Grade](https://img.shields.io/lgtm/grade/cpp/github/michaelrsweet/pdfio)](https://lgtm.com/projects/g/michaelrsweet/pdfio/context:cpp)
-[![LGTM Alerts](https://img.shields.io/lgtm/alerts/github/michaelrsweet/pdfio)](https://lgtm.com/projects/g/michaelrsweet/pdfio/)

 PDFio is a simple C library for reading and writing PDF files.  The primary
 goals of PDFio are:
@ -115,13 +113,13 @@ generates a static library that will be installed under "/usr/local" with:

 You can reproduce this with the makefile using:

-    sudo make COMMONFLAGS="-Os -mmacosx-version-min=10.14 -arch x86_64 -arch arm64" install
+    sudo make macos install


 Legal Stuff
 -----------

-PDFio is Copyright © 2021 by Michael R Sweet.
+PDFio is Copyright © 2021-2023 by Michael R Sweet.

 This software is licensed under the Apache License Version 2.0 with an
 (optional) exception to allow linking against GPL2/LGPL2 software.  See the
--- a/SECURITY.md
+++ b/SECURITY.md
@ -5,6 +5,25 @@ This file describes how security issues are reported and handled, and what the
 expectations are for security issues reported to this project.


+Reporting a Security Bug
+------------------------
+
+For the purposes of this project, a security bug is a software defect that
+allows a *local or remote user* to gain unauthorized access or privileges on the
+host computer or to cause the software to crash.  Such defects should be
+reported to the project security advisory page at
+<https://github.com/michaelrsweet/pdfio/security/advisories>.
+
+Alternately, security bugs can be reported to "security AT msweet.org" using the
+PGP public key below.  Expect a response within 5 business days.  Any proposed
+embargo date should be at least 30 days and no more than 90 days in the future.
+
+> *Note:* If you've found a software defect that allows a *program* to gain
+> unauthorized access or privileges on the host computer or causes the program
+> to crash, that defect should be reported as an ordinary project issue at
+> <https://github.com/michaelrsweet/pdfio/issues>.
+
+
 Responsible Disclosure
 ----------------------

@ -50,14 +69,6 @@ example:
    1.0rc1


-Reporting a Vulnerability
-------------------------
-
-Report all security issues to "security AT msweet.org".  Expect a response
-within 5 business days.  Any proposed embargo date should be at least 30 days
-and no more than 90 days in the future.
-
-
 PGP Public Key
 --------------

--- a/doc/pdfio.3
+++ b/doc/pdfio.3
@ -1,4 +1,4 @@
-.TH pdfio 3 "pdf read/write library" "2021-10-25" "pdf read/write library"
+.TH pdfio 3 "pdf read/write library" "2023-11-15" "pdf read/write library"
 .SH NAME
 pdfio \- pdf read/write library
 .SH Introduction
@ -34,7 +34,7 @@ PDFio is
 .I not
 concerned with rendering or viewing a PDF file, although a PDF RIP or viewer could be written using it.
 .PP
-PDFio is Copyright \[co] 2021 by Michael R Sweet and is licensed under the Apache License Version 2.0 with an (optional) exception to allow linking against GPL2/LGPL2 software. See the files "LICENSE" and "NOTICE" for more information.
+PDFio is Copyright \[co] 2021\-2023 by Michael R Sweet and is licensed under the Apache License Version 2.0 with an (optional) exception to allow linking against GPL2/LGPL2 software. See the files "LICENSE" and "NOTICE" for more information.
 .SS Requirements
 .PP
 PDFio requires the following to build the software:
@ -156,7 +156,7 @@ There is also an Xcode project ("pdfio.xcodeproj") you can use on macOS which ge
 You can reproduce this with the makefile using:
 .nf

-    sudo make COMMONFLAGS="\-Os \-mmacosx\-version\-min=10.14 \-arch x86_64 \-arch arm64" install
+    sudo make macos install
 .fi
 .SS Detecting PDFio
 .PP
@ -254,7 +254,7 @@ Each PDF file contains one or more pages. The pdfioFileGetNumPages function retu
    }
 .fi
 .PP
-Each page is represented by a "page tree" object (what pdfioFileGetPage returns) that specifies information about the page and one or more "content" objects that contain the images, fonts, text, and graphics that appear on the page.
+Each page is represented by a "page tree" object (what pdfioFileGetPage returns) that specifies information about the page and one or more "content" objects that contain the images, fonts, text, and graphics that appear on the page. Use the pdfioPageGetNumStreams and pdfioPageOpenStream functions to access the content streams for each page.
 .PP
 The pdfioFileClose function closes a PDF file and frees all memory that was used for it:
 .nf
@ -324,6 +324,14 @@ Some PDF objects have an associated data stream, such as for pages, images, ICC
 .PP
 The first argument is the object pointer. The second argument is a boolean value that specifies whether you want to decode (typically decompress) the stream data or return it as\-is.
 .PP
+When reading a page stream you'll use the pdfioPageOpenStream function instead:
+.nf
+
+    pdfio_file_t *pdf = pdfioFileOpen(...);
+    pdfio_obj_t *obj = pdfioFileGetPage(pdf, number);
+    pdfio_stream_t *st = pdfioPageOpenStream(obj, 0, true);
+.fi
+.PP
 Once you have the stream open, you can use one of several functions to read from it:
 .IP \(bu 5
 .PP
@ -353,12 +361,21 @@ To create a stream for a new object, call the pdfioObjCreateStream function:
 .nf

    pdfio_file_t *pdf = pdfioFileCreate(...);
-    pdfio_obj_t *pdfioFileCreateObj(pdf, ...);
-    pdfio_stream_t *pdfioObjCreateStream(obj, PDFIO_FILTER_FLATE);
+    pdfio_obj_t *obj = pdfioFileCreateObj(pdf, ...);
+    pdfio_stream_t *st = pdfioObjCreateStream(obj, PDFIO_FILTER_FLATE);
 .fi
 .PP
 The first argument is the newly created object. The second argument is either PDFIO_FILTER_NONE to specify that any encoding is done by your program or PDFIO_FILTER_FLATE to specify that PDFio should Flate compress the stream.
 .PP
+To create a page content stream call the pdfioFileCreatePage function:
+.nf
+
+    pdfio_file_t *pdf = pdfioFileCreate(...);
+    pdfio_dict_t *dict = pdfioDictCreate(pdf);
+    \... set page dictionary keys and values ...
+    pdfio_stream_t *st = pdfioFileCreatePage(pdf, dict);
+.fi
+.PP
 Once you have created the stream, use any of the following functions to write to the stream:
 .IP \(bu 5
 .PP
@ -1340,7 +1357,7 @@ bool  pdfioContentFillAndStroke (
 .fi
 .SS pdfioContentMatrixConcat
 Concatenate a matrix to the current graphics
-state.
+                               state.
 .PP
 .nf
 bool  pdfioContentMatrixConcat (
@ -1423,6 +1440,14 @@ bool  pdfioContentPathCurve23 (
    double y3
 );
 .fi
+.SS pdfioContentPathEnd
+Clear the current path.
+.PP
+.nf
+bool  pdfioContentPathEnd (
+    pdfio_stream_t *st
+);
+.fi
 .SS pdfioContentPathLineTo
 Add a straight line to the current path.
 .PP
@ -1482,6 +1507,9 @@ bool  pdfioContentSetDashPattern (
    double off
 );
 .fi
+.PP
+This function sets the stroke pattern when drawing lines.  If "on" and "off"
+are 0, a solid line is drawn.
 .SS pdfioContentSetFillColorDeviceCMYK
 Set device CMYK fill color.
 .PP
@ -1933,6 +1961,32 @@ pdfio_valtype_t  pdfioDictGetType (
    const char *key
 );
 .fi
+.SS pdfioDictIterateKeys
+Iterate the keys in a dictionary.
+.PP
+.nf
+void pdfioDictIterateKeys (
+    pdfio_dict_t *dict,
+    pdfio_dict_cb_t cb,
+    void *cb_data
+);
+.fi
+.PP
+This function iterates the keys in a dictionary, calling the supplied
+function "cb":
+.PP
+.nf
+    bool
+    my_dict_cb(pdfio_dict_t *dict, const char *key, void *cb_data)
+    {
+    ... "key" contains the dictionary key ...
+    ... return true to continue or false to stop ...
+    }
+
+.fi
+
+The iteration continues as long as the callback returns \fBtrue\fR or all keys
+have been iterated.
 .SS pdfioDictSetArray
 Set a key array in a dictionary.
 .PP
@ -2255,10 +2309,11 @@ written:
    ssize_t
    output_cb(void *output_ctx, const void *buffer, size_t bytes)
    {
-    // Write buffer to output and return the number of bytes written
+      // Write buffer to output and return the number of bytes written
    }

 .fi
+
 The "version" argument specifies the PDF version number for the file or
 \fBNULL\fR for the default ("2.0").
 .PP
@ -2285,6 +2340,20 @@ pdfio_stream_t * pdfioFileCreatePage (
    pdfio_dict_t *dict
 );
 .fi
+.SS pdfioFileCreateTemporary
+
+.PP
+.nf
+pdfio_file_t * pdfioFileCreateTemporary (
+    char *buffer,
+    size_t bufsize,
+    const char *version,
+    pdfio_rect_t *media_box,
+    pdfio_rect_t *crop_box,
+    pdfio_error_cb_t error_cb,
+    void *error_data
+);
+.fi
 .SS pdfioFileFindObj
 Find an object using its object number.
 .PP
@ -2552,7 +2621,7 @@ double  pdfioImageGetWidth (
 .fi
 .SS pdfioObjClose
 Close an object, writing any data as needed to the PDF
-file.
+                    file.
 .PP
 .nf
 bool  pdfioObjClose (
@ -2693,6 +2762,24 @@ bool  pdfioPageDictAddImage (
    pdfio_obj_t *obj
 );
 .fi
+.SS pdfioPageGetNumStreams
+Get the number of content streams for a page object.
+.PP
+.nf
+size_t  pdfioPageGetNumStreams (
+    pdfio_obj_t *page
+);
+.fi
+.SS pdfioPageOpenStream
+Open a content stream for a page.
+.PP
+.nf
+pdfio_stream_t * pdfioPageOpenStream (
+    pdfio_obj_t *page,
+    size_t n,
+    bool decode
+);
+.fi
 .SS pdfioStreamClose
 Close a (data) stream in a PDF file.
 .PP
@ -2720,6 +2807,13 @@ bool  pdfioStreamGetToken (
    size_t bufsize
 );
 .fi
+.PP
+This function reads a single PDF token from a stream.  Operator tokens,
+boolean values, and numbers are returned as-is in the provided string buffer.
+String values start with the opening parenthesis ('(') but have all escaping
+resolved and the terminating parenthesis removed.  Hexadecimal string values
+start with the opening angle bracket ('<') and have all whitespace and the
+terminating angle bracket removed.
 .SS pdfioStreamPeek
 Peek at data in a stream.
 .PP
@ -2840,6 +2934,12 @@ Standard color spaces
 .nf
 typedef enum pdfio_cs_e pdfio_cs_t;
 .fi
+.SS pdfio_dict_cb_t
+Dictionary iterator callback
+.PP
+.nf
+typedef bool(*)(pdfio_dict_t *dict, const char *key, void *cb_data) pdfio_dict_cb_t;
+.fi
 .SS pdfio_dict_t
 Key/value dictionary
 .PP
@ -2947,4 +3047,4 @@ typedef uint8_t state_t[4][4];
 Michael R Sweet
 .SH COPYRIGHT
 .PP
-Copyright (c) 2021 by Michael R Sweet
+Copyright (c) 2021-2023 by Michael R Sweet
--- a/doc/pdfio.html
+++ b/doc/pdfio.html
@ -1,13 +1,13 @@
 <!DOCTYPE html>
 <html lang="en-US">
 <head>
-<title>PDFio Programming Manual v1.0.0</title>
+<title>PDFio Programming Manual v1.1.3</title>
 <meta http-equiv="Content-Type" content="text/html;charset=utf-8">
 <meta name="generator" content="codedoc v3.7">
 <meta name="author" content="Michael R Sweet">
 <meta name="language" content="en-US">
-<meta name="copyright" content="Copyright © 2021 by Michael R Sweet">
-<meta name="version" content="1.0.0">
+<meta name="copyright" content="Copyright © 2021-2023 by Michael R Sweet">
+<meta name="version" content="1.1.3">
 <style type="text/css"><!--
 body {
  background: white;
@ -245,9 +245,9 @@ span.string {
 <body>
 <div class="header">
 <p><img class="title" src="pdfio-512.png"></p>
-<h1 class="title">PDFio Programming Manual v1.0.0</h1>
+<h1 class="title">PDFio Programming Manual v1.1.3</h1>
 <p>Michael R Sweet</p>
-<p>Copyright © 2021 by Michael R Sweet</p>
+<p>Copyright © 2021-2023 by Michael R Sweet</p>
 </div>
 <div class="contents">
 <h2 class="title">Contents</h2>
@ -307,6 +307,7 @@ span.string {
 <li><a href="#pdfioContentPathCurve">pdfioContentPathCurve</a></li>
 <li><a href="#pdfioContentPathCurve13">pdfioContentPathCurve13</a></li>
 <li><a href="#pdfioContentPathCurve23">pdfioContentPathCurve23</a></li>
+<li><a href="#pdfioContentPathEnd">pdfioContentPathEnd</a></li>
 <li><a href="#pdfioContentPathLineTo">pdfioContentPathLineTo</a></li>
 <li><a href="#pdfioContentPathMoveTo">pdfioContentPathMoveTo</a></li>
 <li><a href="#pdfioContentPathRect">pdfioContentPathRect</a></li>
@ -360,6 +361,7 @@ span.string {
 <li><a href="#pdfioDictGetRect">pdfioDictGetRect</a></li>
 <li><a href="#pdfioDictGetString">pdfioDictGetString</a></li>
 <li><a href="#pdfioDictGetType">pdfioDictGetType</a></li>
+<li><a href="#pdfioDictIterateKeys">pdfioDictIterateKeys</a></li>
 <li><a href="#pdfioDictSetArray">pdfioDictSetArray</a></li>
 <li><a href="#pdfioDictSetBinary">pdfioDictSetBinary</a></li>
 <li><a href="#pdfioDictSetBoolean">pdfioDictSetBoolean</a></li>
@ -383,6 +385,7 @@ span.string {
 <li><a href="#pdfioFileCreateObj">pdfioFileCreateObj</a></li>
 <li><a href="#pdfioFileCreateOutput">pdfioFileCreateOutput</a></li>
 <li><a href="#pdfioFileCreatePage">pdfioFileCreatePage</a></li>
+<li><a href="#pdfioFileCreateTemporary">pdfioFileCreateTemporary</a></li>
 <li><a href="#pdfioFileFindObj">pdfioFileFindObj</a></li>
 <li><a href="#pdfioFileGetAuthor">pdfioFileGetAuthor</a></li>
 <li><a href="#pdfioFileGetCreationDate">pdfioFileGetCreationDate</a></li>
@ -425,6 +428,8 @@ span.string {
 <li><a href="#pdfioPageDictAddColorSpace">pdfioPageDictAddColorSpace</a></li>
 <li><a href="#pdfioPageDictAddFont">pdfioPageDictAddFont</a></li>
 <li><a href="#pdfioPageDictAddImage">pdfioPageDictAddImage</a></li>
+<li><a href="#pdfioPageGetNumStreams">pdfioPageGetNumStreams</a></li>
+<li><a href="#pdfioPageOpenStream">pdfioPageOpenStream</a></li>
 <li><a href="#pdfioStreamClose">pdfioStreamClose</a></li>
 <li><a href="#pdfioStreamConsume">pdfioStreamConsume</a></li>
 <li><a href="#pdfioStreamGetToken">pdfioStreamGetToken</a></li>
@ -440,6 +445,7 @@ span.string {
 <li><a href="#TYPES">Data Types</a><ul class="subcontents">
 <li><a href="#pdfio_array_t">pdfio_array_t</a></li>
 <li><a href="#pdfio_cs_t">pdfio_cs_t</a></li>
+<li><a href="#pdfio_dict_cb_t">pdfio_dict_cb_t</a></li>
 <li><a href="#pdfio_dict_t">pdfio_dict_t</a></li>
 <li><a href="#pdfio_encryption_t">pdfio_encryption_t</a></li>
 <li><a href="#pdfio_error_cb_t">pdfio_error_cb_t</a></li>
@ -491,7 +497,7 @@ span.string {
 </li>
 </ul>
 <p>PDFio is <em>not</em> concerned with rendering or viewing a PDF file, although a PDF RIP or viewer could be written using it.</p>
-<p>PDFio is Copyright © 2021 by Michael R Sweet and is licensed under the Apache License Version 2.0 with an (optional) exception to allow linking against GPL2/LGPL2 software. See the files &quot;LICENSE&quot; and &quot;NOTICE&quot; for more information.</p>
+<p>PDFio is Copyright © 2021-2023 by Michael R Sweet and is licensed under the Apache License Version 2.0 with an (optional) exception to allow linking against GPL2/LGPL2 software. See the files &quot;LICENSE&quot; and &quot;NOTICE&quot; for more information.</p>
 <h3 class="title" id="requirements">Requirements</h3>
 <p>PDFio requires the following to build the software:</p>
 <ul>
@ -557,7 +563,7 @@ make install-shared
 <pre><code>sudo xcodebuild install
 </code></pre>
 <p>You can reproduce this with the makefile using:</p>
-<pre><code>sudo make COMMONFLAGS=&quot;-Os -mmacosx-version-min=10.14 -arch x86_64 -arch arm64&quot; install
+<pre><code>sudo make macos install
 </code></pre>
 <h3 class="title" id="detecting-pdfio">Detecting PDFio</h3>
 <p>PDFio can be detected using the <code>pkg-config</code> command, for example:</p>
@ -621,7 +627,7 @@ pdfio_obj_t  *page;  <span class="comment">// Current page</span>
  <span class="comment">// do something with page</span>
 }
 </code></pre>
-<p>Each page is represented by a &quot;page tree&quot; object (what <a href="#pdfioFileGetPage"><code>pdfioFileGetPage</code></a> returns) that specifies information about the page and one or more &quot;content&quot; objects that contain the images, fonts, text, and graphics that appear on the page.</p>
+<p>Each page is represented by a &quot;page tree&quot; object (what <a href="#pdfioFileGetPage"><code>pdfioFileGetPage</code></a> returns) that specifies information about the page and one or more &quot;content&quot; objects that contain the images, fonts, text, and graphics that appear on the page. Use the <a href="#pdfioPageGetNumStreams"><code>pdfioPageGetNumStreams</code></a> and <a href="#pdfioPageOpenStream"><code>pdfioPageOpenStream</code></a> functions to access the content streams for each page.</p>
 <p>The <a href="#pdfioFileClose"><code>pdfioFileClose</code></a> function closes a PDF file and frees all memory that was used for it:</p>
 <pre><code class="language-c">pdfioFileClose(pdf);
 </code></pre>
@ -663,6 +669,11 @@ pdfio_obj_t *obj = pdfioFileFindObj(pdf, number);
 pdfio_stream_t *st = pdfioObjOpenStream(obj, <span class="reserved">true</span>);
 </code></pre>
 <p>The first argument is the object pointer. The second argument is a boolean value that specifies whether you want to decode (typically decompress) the stream data or return it as-is.</p>
+<p>When reading a page stream you'll use the <a href="#pdfioPageOpenStream"><code>pdfioPageOpenStream</code></a> function instead:</p>
+<pre><code class="language-c">pdfio_file_t *pdf = pdfioFileOpen(...);
+pdfio_obj_t *obj = pdfioFileGetPage(pdf, number);
+pdfio_stream_t *st = pdfioPageOpenStream(obj, <span class="number">0</span>, <span class="reserved">true</span>);
+</code></pre>
 <p>Once you have the stream open, you can use one of several functions to read from it:</p>
 <ul>
 <li><p><a href="#pdfioStreamConsume"><code>pdfioStreamConsume</code></a> reads and discards a number of bytes in the stream</p>
@ -679,10 +690,16 @@ pdfio_stream_t *st = pdfioObjOpenStream(obj, <span class="reserved">true</span>)
 </code></pre>
 <p>To create a stream for a new object, call the <a href="#pdfioObjCreateStream"><code>pdfioObjCreateStream</code></a> function:</p>
 <pre><code class="language-c">pdfio_file_t *pdf = pdfioFileCreate(...);
-pdfio_obj_t *pdfioFileCreateObj(pdf, ...);
-pdfio_stream_t *pdfioObjCreateStream(obj, PDFIO_FILTER_FLATE);
+pdfio_obj_t *obj = pdfioFileCreateObj(pdf, ...);
+pdfio_stream_t *st = pdfioObjCreateStream(obj, PDFIO_FILTER_FLATE);
 </code></pre>
 <p>The first argument is the newly created object. The second argument is either <code>PDFIO_FILTER_NONE</code> to specify that any encoding is done by your program or <code>PDFIO_FILTER_FLATE</code> to specify that PDFio should Flate compress the stream.</p>
+<p>To create a page content stream call the <a href="#pdfioFileCreatePage"><code>pdfioFileCreatePage</code></a> function:</p>
+<pre><code class="language-c">pdfio_file_t *pdf = pdfioFileCreate(...);
+pdfio_dict_t *dict = pdfioDictCreate(pdf);
+... set page dictionary keys <span class="reserved">and</span> values ...
+pdfio_stream_t *st = pdfioFileCreatePage(pdf, dict);
+</code></pre>
 <p>Once you have created the stream, use any of the following functions to write to the stream:</p>
 <ul>
 <li><p><a href="#pdfioStreamPrintf"><code>pdfioStreamPrintf</code></a> writes a formatted string to the stream</p>
@ -1371,7 +1388,7 @@ bool pdfioContentFillAndStroke(<a href="#pdfio_stream_t">pdfio_stream_t</a> *st,
 <p class="description"><code>true</code> on success, <code>false</code> on failure</p>
 <h3 class="function"><a id="pdfioContentMatrixConcat">pdfioContentMatrixConcat</a></h3>
 <p class="description">Concatenate a matrix to the current graphics
-state.</p>
+                               state.</p>
 <p class="code">
 bool pdfioContentMatrixConcat(<a href="#pdfio_stream_t">pdfio_stream_t</a> *st, pdfio_matrix_t m);</p>
 <h4 class="parameters">Parameters</h4>
@ -1498,6 +1515,17 @@ bool pdfioContentPathCurve23(<a href="#pdfio_stream_t">pdfio_stream_t</a> *st, d
 </tbody></table>
 <h4 class="returnvalue">Return Value</h4>
 <p class="description"><code>true</code> on success, <code>false</code> on failure</p>
+<h3 class="function"><a id="pdfioContentPathEnd">pdfioContentPathEnd</a></h3>
+<p class="description">Clear the current path.</p>
+<p class="code">
+bool pdfioContentPathEnd(<a href="#pdfio_stream_t">pdfio_stream_t</a> *st);</p>
+<h4 class="parameters">Parameters</h4>
+<table class="list"><tbody>
+<tr><th>st</th>
+<td class="description">Stream</td></tr>
+</tbody></table>
+<h4 class="returnvalue">Return Value</h4>
+<p class="description"><code>true</code> on success, <code>false</code> on failure</p>
 <h3 class="function"><a id="pdfioContentPathLineTo">pdfioContentPathLineTo</a></h3>
 <p class="description">Add a straight line to the current path.</p>
 <p class="code">
@ -1586,6 +1614,9 @@ bool pdfioContentSetDashPattern(<a href="#pdfio_stream_t">pdfio_stream_t</a> *st
 </tbody></table>
 <h4 class="returnvalue">Return Value</h4>
 <p class="description"><code>true</code> on success, <code>false</code> on failure</p>
+<h4 class="discussion">Discussion</h4>
+<p class="discussion">This function sets the stroke pattern when drawing lines.  If &quot;on&quot; and &quot;off&quot;
+are 0, a solid line is drawn.</p>
 <h3 class="function"><a id="pdfioContentSetFillColorDeviceCMYK">pdfioContentSetFillColorDeviceCMYK</a></h3>
 <p class="description">Set device CMYK fill color.</p>
 <p class="code">
@ -2248,6 +2279,34 @@ const char *pdfioDictGetString(<a href="#pdfio_dict_t">pdfio_dict_t</a> *dict, c
 </tbody></table>
 <h4 class="returnvalue">Return Value</h4>
 <p class="description">Value type</p>
+<h3 class="function"><a id="pdfioDictIterateKeys">pdfioDictIterateKeys</a></h3>
+<p class="description">Iterate the keys in a dictionary.</p>
+<p class="code">
+void pdfioDictIterateKeys(<a href="#pdfio_dict_t">pdfio_dict_t</a> *dict, <a href="#pdfio_dict_cb_t">pdfio_dict_cb_t</a> cb, void *cb_data);</p>
+<h4 class="parameters">Parameters</h4>
+<table class="list"><tbody>
+<tr><th>dict</th>
+<td class="description">Dictionary</td></tr>
+<tr><th>cb</th>
+<td class="description">Callback function</td></tr>
+<tr><th>cb_data</th>
+<td class="description">Callback data</td></tr>
+</tbody></table>
+<h4 class="discussion">Discussion</h4>
+<p class="discussion">This function iterates the keys in a dictionary, calling the supplied
+function &quot;cb&quot;:
+
+<pre>
+bool
+my_dict_cb(pdfio_dict_t *dict, const char *key, void *cb_data)
+{
+... &quot;key&quot; contains the dictionary key ...
+... return true to continue or false to stop ...
+}
+</pre>
+
+The iteration continues as long as the callback returns <code>true</code> or all keys
+have been iterated.</p>
 <h3 class="function"><a id="pdfioDictSetArray">pdfioDictSetArray</a></h3>
 <p class="description">Set a key array in a dictionary.</p>
 <p class="code">
@ -2686,9 +2745,10 @@ written:
 ssize_t
 output_cb(void *output_ctx, const void *buffer, size_t bytes)
 {
-// Write buffer to output and return the number of bytes written
+  // Write buffer to output and return the number of bytes written
 }
 </pre>
+
 The &quot;version&quot; argument specifies the PDF version number for the file or
 <code>NULL</code> for the default (&quot;2.0&quot;).<br>
 <br>
@ -2717,6 +2777,35 @@ stored as indirect object references.</blockquote>
 </tbody></table>
 <h4 class="returnvalue">Return Value</h4>
 <p class="description">Contents stream</p>
+<h3 class="function"><a id="pdfioFileCreateTemporary">pdfioFileCreateTemporary</a></h3>
+<p class="description"></p>
+<p class="code">
+<a href="#pdfio_file_t">pdfio_file_t</a> *pdfioFileCreateTemporary(char *buffer, size_t bufsize, const char *version, <a href="#pdfio_rect_t">pdfio_rect_t</a> *media_box, <a href="#pdfio_rect_t">pdfio_rect_t</a> *crop_box, <a href="#pdfio_error_cb_t">pdfio_error_cb_t</a> error_cb, void *error_data);</p>
+<h4 class="parameters">Parameters</h4>
+<table class="list"><tbody>
+<tr><th>buffer</th>
+<td class="description">Filename buffer</td></tr>
+<tr><th>bufsize</th>
+<td class="description">Size of filename buffer</td></tr>
+<tr><th>version</th>
+<td class="description">PDF version number or <code>NULL</code> for default (2.0)</td></tr>
+<tr><th>media_box</th>
+<td class="description">Default MediaBox for pages</td></tr>
+<tr><th>crop_box</th>
+<td class="description">Default CropBox for pages</td></tr>
+<tr><th>error_cb</th>
+<td class="description">Error callback or <code>NULL</code> for default</td></tr>
+<tr><th>error_data</th>
+<td class="description">Error callback data, if any</td></tr>
+</tbody></table>
+<h4 class="returnvalue">Return Value</h4>
+<p class="description">Create a temporary PDF file.</p>
+<p class="discussion">This function creates a PDF file with a unique filename in the current
+temporary directory.  The temporary file is stored in the string &quot;buffer&quot; an
+will have a &quot;.pdf&quot; extension.  Otherwise, this function works the same as
+the <a href="#pdfioFileCreate"><code>pdfioFileCreate</code></a> function.
+
+</p>
 <h3 class="function"><a id="pdfioFileFindObj">pdfioFileFindObj</a></h3>
 <p class="description">Find an object using its object number.</p>
 <p class="code">
@ -3068,7 +3157,7 @@ double pdfioImageGetWidth(<a href="#pdfio_obj_t">pdfio_obj_t</a> *obj);</p>
 <p class="description">Width in columns</p>
 <h3 class="function"><a id="pdfioObjClose">pdfioObjClose</a></h3>
 <p class="description">Close an object, writing any data as needed to the PDF
-file.</p>
+                    file.</p>
 <p class="code">
 bool pdfioObjClose(<a href="#pdfio_obj_t">pdfio_obj_t</a> *obj);</p>
 <h4 class="parameters">Parameters</h4>
@ -3264,6 +3353,32 @@ bool pdfioPageDictAddImage(<a href="#pdfio_dict_t">pdfio_dict_t</a> *dict, const
 </tbody></table>
 <h4 class="returnvalue">Return Value</h4>
 <p class="description"><code>true</code> on success, <code>false</code> on failure</p>
+<h3 class="function"><a id="pdfioPageGetNumStreams">pdfioPageGetNumStreams</a></h3>
+<p class="description">Get the number of content streams for a page object.</p>
+<p class="code">
+size_t pdfioPageGetNumStreams(<a href="#pdfio_obj_t">pdfio_obj_t</a> *page);</p>
+<h4 class="parameters">Parameters</h4>
+<table class="list"><tbody>
+<tr><th>page</th>
+<td class="description">Page object</td></tr>
+</tbody></table>
+<h4 class="returnvalue">Return Value</h4>
+<p class="description">Number of streams</p>
+<h3 class="function"><a id="pdfioPageOpenStream">pdfioPageOpenStream</a></h3>
+<p class="description">Open a content stream for a page.</p>
+<p class="code">
+<a href="#pdfio_stream_t">pdfio_stream_t</a> *pdfioPageOpenStream(<a href="#pdfio_obj_t">pdfio_obj_t</a> *page, size_t n, bool decode);</p>
+<h4 class="parameters">Parameters</h4>
+<table class="list"><tbody>
+<tr><th>page</th>
+<td class="description">Page object</td></tr>
+<tr><th>n</th>
+<td class="description">Stream index (0-based)</td></tr>
+<tr><th>decode</th>
+<td class="description"><code>true</code> to decode/decompress stream</td></tr>
+</tbody></table>
+<h4 class="returnvalue">Return Value</h4>
+<p class="description">Stream</p>
 <h3 class="function"><a id="pdfioStreamClose">pdfioStreamClose</a></h3>
 <p class="description">Close a (data) stream in a PDF file.</p>
 <p class="code">
@ -3303,6 +3418,13 @@ bool pdfioStreamGetToken(<a href="#pdfio_stream_t">pdfio_stream_t</a> *st, char
 </tbody></table>
 <h4 class="returnvalue">Return Value</h4>
 <p class="description"><code>true</code> on success, <code>false</code> on EOF</p>
+<h4 class="discussion">Discussion</h4>
+<p class="discussion">This function reads a single PDF token from a stream.  Operator tokens,
+boolean values, and numbers are returned as-is in the provided string buffer.
+String values start with the opening parenthesis ('(') but have all escaping
+resolved and the terminating parenthesis removed.  Hexadecimal string values
+start with the opening angle bracket ('&lt;') and have all whitespace and the
+terminating angle bracket removed.</p>
 <h3 class="function"><a id="pdfioStreamPeek">pdfioStreamPeek</a></h3>
 <p class="description">Peek at data in a stream.</p>
 <p class="code">
@ -3445,6 +3567,11 @@ typedef struct _pdfio_array_s pdfio_array_t;
 <p class="code">
 typedef enum <a href="#pdfio_cs_e">pdfio_cs_e</a> pdfio_cs_t;
 </p>
+<h3 class="typedef"><a id="pdfio_dict_cb_t">pdfio_dict_cb_t</a></h3>
+<p class="description">Dictionary iterator callback</p>
+<p class="code">
+typedef bool (*pdfio_dict_cb_t)(<a href="#pdfio_dict_t">pdfio_dict_t</a> *dict, const char *key, void *cb_data);
+</p>
 <h3 class="typedef"><a id="pdfio_dict_t">pdfio_dict_t</a></h3>
 <p class="description">Key/value dictionary</p>
 <p class="code">
--- a/doc/pdfio.md
+++ b/doc/pdfio.md
@ -15,8 +15,8 @@ goals of pdfio are:
 PDFio is *not* concerned with rendering or viewing a PDF file, although a PDF
 RIP or viewer could be written using it.

-PDFio is Copyright © 2021 by Michael R Sweet and is licensed under the Apache
-License Version 2.0 with an (optional) exception to allow linking against
+PDFio is Copyright © 2021-2023 by Michael R Sweet and is licensed under the
+Apache License Version 2.0 with an (optional) exception to allow linking against
 GPL2/LGPL2 software.  See the files "LICENSE" and "NOTICE" for more information.


@ -104,7 +104,7 @@ generates a static library that will be installed under "/usr/local" with:

 You can reproduce this with the makefile using:

-    sudo make COMMONFLAGS="-Os -mmacosx-version-min=10.14 -arch x86_64 -arch arm64" install
+    sudo make macos install


 Detecting PDFio
@ -209,7 +209,8 @@ for (i = 0, count = pdfioFileGetNumPages(pdf); i < count; i ++)
 Each page is represented by a "page tree" object (what [`pdfioFileGetPage`](@@)
 returns) that specifies information about the page and one or more "content"
 objects that contain the images, fonts, text, and graphics that appear on the
-page.
+page.  Use the [`pdfioPageGetNumStreams`](@@) and [`pdfioPageOpenStream`](@@)
+functions to access the content streams for each page.

 The [`pdfioFileClose`](@@) function closes a PDF file and frees all memory that
 was used for it:
@ -294,6 +295,15 @@ The first argument is the object pointer.  The second argument is a boolean
 value that specifies whether you want to decode (typically decompress) the
 stream data or return it as-is.

+When reading a page stream you'll use the [`pdfioPageOpenStream`](@@) function
+instead:
+
+```c
+pdfio_file_t *pdf = pdfioFileOpen(...);
+pdfio_obj_t *obj = pdfioFileGetPage(pdf, number);
+pdfio_stream_t *st = pdfioPageOpenStream(obj, 0, true);
+```
+
 Once you have the stream open, you can use one of several functions to read
 from it:

@ -315,14 +325,23 @@ function:

 ```c
 pdfio_file_t *pdf = pdfioFileCreate(...);
-pdfio_obj_t *pdfioFileCreateObj(pdf, ...);
-pdfio_stream_t *pdfioObjCreateStream(obj, PDFIO_FILTER_FLATE);
+pdfio_obj_t *obj = pdfioFileCreateObj(pdf, ...);
+pdfio_stream_t *st = pdfioObjCreateStream(obj, PDFIO_FILTER_FLATE);
 ```

 The first argument is the newly created object.  The second argument is either
 `PDFIO_FILTER_NONE` to specify that any encoding is done by your program or
 `PDFIO_FILTER_FLATE` to specify that PDFio should Flate compress the stream.

+To create a page content stream call the [`pdfioFileCreatePage`](@@) function:
+
+```c
+pdfio_file_t *pdf = pdfioFileCreate(...);
+pdfio_dict_t *dict = pdfioDictCreate(pdf);
+... set page dictionary keys and values ...
+pdfio_stream_t *st = pdfioFileCreatePage(pdf, dict);
+```
+
 Once you have created the stream, use any of the following functions to write
 to the stream:

--- a/pdfio-common.c
+++ b/pdfio-common.c
@ -1,7 +1,7 @@
 //
 // Common support functions for pdfio.
 //
-// Copyright © 2021 by Michael R Sweet.
+// Copyright © 2021-2023 by Michael R Sweet.
 //
 // Licensed under Apache License v2.0.  See the file "LICENSE" for more
 // information.
@ -38,6 +38,8 @@ _pdfioFileConsume(pdfio_file_t *pdf,	// I - PDF file
  else if (_pdfioFileSeek(pdf, (off_t)bytes, SEEK_CUR) < 0)
    return (false);

+  PDFIO_DEBUG("_pdfioFileConsume: pos=%ld\n", (long)(pdf->bufpos + pdf->bufptr - pdf->buffer));
+
  return (true);
 }

@ -373,9 +375,9 @@ _pdfioFileSeek(pdfio_file_t *pdf,	// I - PDF file
    if (whence != SEEK_END && offset >= pdf->bufpos && offset < (pdf->bufpos + pdf->bufend - pdf->buffer))
    {
      // Yes, seek within existing buffer...
-      pdf->bufptr = pdf->buffer + offset - pdf->bufpos;
+      pdf->bufptr = pdf->buffer + (offset - pdf->bufpos);
      PDFIO_DEBUG("_pdfioFileSeek: Seek within buffer, bufpos=%ld.\n", (long)pdf->bufpos);
-      PDFIO_DEBUG("_pdfioFileSeek: buffer=%p, bufptr=%p, bufend=%p\n", pdf->buffer, pdf->bufptr, pdf->bufend);
+      PDFIO_DEBUG("_pdfioFileSeek: buffer=%p, bufptr=%p(<%02X%02X...>), bufend=%p\n", pdf->buffer, pdf->bufptr, pdf->bufptr[0] & 255, pdf->bufptr[1] & 255, pdf->bufend);
      return (offset);
    }

@ -525,7 +527,7 @@ read_buffer(pdfio_file_t *pdf,		// I - PDF file
  return (rbytes);
 }

-  
+
 //
 // 'write_buffer()' - Write a buffer to a PDF file.
 //
--- a/pdfio-content.c
+++ b/pdfio-content.c
@ -1,7 +1,7 @@
 //
 // Content helper functions for PDFio.
 //
-// Copyright © 2021 by Michael R Sweet.
+// Copyright © 2021-2023 by Michael R Sweet.
 //
 // Licensed under Apache License v2.0.  See the file "LICENSE" for more
 // information.
@ -489,8 +489,11 @@ pdfioContentMatrixRotate(
    pdfio_stream_t *st,			// I - Stream
    double         degrees)		// I - Rotation angle in degrees counter-clockwise
 {
-  double dcos = cos(degrees / M_PI);	// Cosine
-  double dsin = sin(degrees / M_PI);	// Sine
+  double dcos = cos(M_PI * degrees / 180.0);
+					// Cosine
+  double dsin = sin(M_PI * degrees / 180.0);
+					// Sine
+

  return (pdfioStreamPrintf(st, "%g %g %g %g 0 0 cm\n", dcos, -dsin, dsin, dcos));
 }
@ -586,6 +589,17 @@ pdfioContentPathCurve23(
 }


+//
+// 'pdfioContentPathEnd()' - Clear the current path.
+//
+
+bool					// O - `true` on success, `false` on failure
+pdfioContentPathEnd(pdfio_stream_t *st)	// I - Stream
+{
+  return (pdfioStreamPuts(st, "n\n"));
+}
+
+
 //
 // 'pdfioContentPathLineTo()' - Add a straight line to the current path.
 //
@ -656,6 +670,9 @@ pdfioContentSave(pdfio_stream_t *st)	// I - Stream
 //
 // 'pdfioContentSetDashPattern()' - Set the stroke pattern.
 //
+// This function sets the stroke pattern when drawing lines.  If "on" and "off"
+// are 0, a solid line is drawn.
+//

 bool					// O - `true` on success, `false` on failure
 pdfioContentSetDashPattern(
@ -664,7 +681,12 @@ pdfioContentSetDashPattern(
    double         on,			// I - On length
    double         off)			// I - Off length
 {
-  return (pdfioStreamPrintf(st, "[%g %g] %g d\n", on, off, phase));
+  if (on <= 0.0 && off <= 0.0)
+    return (pdfioStreamPrintf(st, "[] %g d\n", phase));
+  else if (fabs(on - off) < 0.001)
+    return (pdfioStreamPrintf(st, "[%g] %g d\n", on, phase));
+  else
+    return (pdfioStreamPrintf(st, "[%g %g] %g d\n", on, off, phase));
 }


@ -1388,6 +1410,7 @@ pdfioFileCreateFontObjFromFile(
    pdfio_dict_t	*cid2gid;	// CIDToGIDMap dictionary
    pdfio_obj_t		*cid2gid_obj;	// CIDToGIDMap object
    size_t		i,		// Looping var
+			start,		// Start character
 			num_cmap;	// Number of CMap entries
    const int		*cmap;		// CMap entries
    unsigned char	*bufptr,	// Pointer into buffer
@ -1396,6 +1419,9 @@ pdfioFileCreateFontObjFromFile(
    pdfio_obj_t		*type2_obj;	// CIDFontType2 font object
    pdfio_array_t	*descendants;	// Decendant font list
    pdfio_dict_t	*sidict;	// CIDSystemInfo dictionary
+    pdfio_array_t	*w_array,	// Width array
+			*temp_array;	// Temporary width sub-array
+    int			w0, w1;		// Widths

    // Create a CIDToGIDMap object for the Unicode font...
    if ((cid2gid = pdfioDictCreate(pdf)) == NULL)
@ -1426,8 +1452,11 @@ pdfioFileCreateFontObjFromFile(

    cmap = ttfGetCMap(font, &num_cmap);

+    PDFIO_DEBUG("pdfioFileCreateFontObjFromFile: num_cmap=%u\n", (unsigned)num_cmap);
+
    for (i = 0, bufptr = buffer, bufend = buffer + sizeof(buffer); i < num_cmap; i ++)
    {
+      PDFIO_DEBUG("pdfioFileCreateFontObjFromFile: cmap[%u]=%d\n", (unsigned)i, cmap[i]);
      if (cmap[i] < 0)
      {
        // Map undefined glyph to .notdef...
@ -1481,6 +1510,54 @@ pdfioFileCreateFontObjFromFile(
      return (NULL);
    }

+    // Width array
+    if ((w_array = pdfioArrayCreate(pdf)) == NULL)
+    {
+      ttfDelete(font);
+      return (NULL);
+    }
+
+    for (start = 0, w0 = ttfGetWidth(font, 0), i = 1; i < 65536; start = i, w0 = w1, i ++)
+    {
+      while (i < 65536 && (w1 = ttfGetWidth(font, i)) == w0)
+        i ++;
+
+      if ((i - start) > 1)
+      {
+        // Encode a repeating sequence...
+        pdfioArrayAppendNumber(w_array, start);
+        pdfioArrayAppendNumber(w_array, i - 1);
+        pdfioArrayAppendNumber(w_array, w0);
+      }
+      else
+      {
+        // Encode a non-repeating sequence...
+        pdfioArrayAppendNumber(w_array, start);
+
+        if ((temp_array = pdfioArrayCreate(pdf)) == NULL)
+        {
+	  ttfDelete(font);
+	  return (NULL);
+        }
+
+        pdfioArrayAppendNumber(temp_array, w0);
+        for (w0 = w1, i ++; i < 65536; w0 = w1, i ++)
+        {
+          if ((w1 = ttfGetWidth(font, i)) == w0 && i < 65535)
+            break;
+
+	  pdfioArrayAppendNumber(temp_array, w0);
+        }
+
+        if (i == 65536)
+	  pdfioArrayAppendNumber(temp_array, w0);
+	else
+	  i --;
+
+        pdfioArrayAppendArray(w_array, temp_array);
+      }
+    }
+
    // CIDSystemInfo mapping to Adobe UCS2 v0 (Unicode)
    pdfioDictSetString(sidict, "Registry", "Adobe");
    pdfioDictSetString(sidict, "Ordering", "Identity");
@ -1493,6 +1570,7 @@ pdfioFileCreateFontObjFromFile(
    pdfioDictSetDict(type2, "CIDSystemInfo", sidict);
    pdfioDictSetObj(type2, "CIDToGIDMap", cid2gid_obj);
    pdfioDictSetObj(type2, "FontDescriptor", desc_obj);
+    pdfioDictSetArray(type2, "W", w_array);

    if ((type2_obj = pdfioFileCreateObj(pdf, type2)) == NULL)
    {
@ -2975,7 +3053,7 @@ write_string(pdfio_stream_t *st,	// I - Stream


  // Start the string...
-  if (!pdfioStreamPuts(st, unicode ? "<" : "("))
+  if (!pdfioStreamPuts(st, unicode ? "<FEFF" : "("))
    return (false);

  // Loop through the string, handling UTF-8 as needed...
--- a/pdfio-content.h
+++ b/pdfio-content.h
@ -1,7 +1,7 @@
 //
 // Public content header file for PDFio.
 //
-// Copyright © 2021 by Michael R Sweet.
+// Copyright © 2021-2023 by Michael R Sweet.
 //
 // Licensed under Apache License v2.0.  See the file "LICENSE" for more
 // information.
@ -9,18 +9,7 @@

 #ifndef PDFIO_CONTENT_H
 #  define PDFIO_CONTENT_H
-
-//
-// Include necessary headers...
-//
-
 #  include "pdfio.h"
-
-
-//
-// C++ magic...
-//
-
 #  ifdef __cplusplus
 extern "C" {
 #  endif // __cplusplus
@ -91,6 +80,7 @@ extern bool		pdfioContentPathClose(pdfio_stream_t *st) _PDFIO_PUBLIC;
 extern bool		pdfioContentPathCurve(pdfio_stream_t *st, double x1, double y1, double x2, double y2, double x3, double y3) _PDFIO_PUBLIC;
 extern bool		pdfioContentPathCurve13(pdfio_stream_t *st, double x1, double y1, double x3, double y3) _PDFIO_PUBLIC;
 extern bool		pdfioContentPathCurve23(pdfio_stream_t *st, double x2, double y2, double x3, double y3) _PDFIO_PUBLIC;
+extern bool		pdfioContentPathEnd(pdfio_stream_t *st) _PDFIO_PUBLIC;
 extern bool		pdfioContentPathLineTo(pdfio_stream_t *st, double x, double y) _PDFIO_PUBLIC;
 extern bool		pdfioContentPathMoveTo(pdfio_stream_t *st, double x, double y) _PDFIO_PUBLIC;
 extern bool		pdfioContentPathRect(pdfio_stream_t *st, double x, double y, double width, double height) _PDFIO_PUBLIC;
@ -150,10 +140,6 @@ extern bool		pdfioPageDictAddFont(pdfio_dict_t *dict, const char *name, pdfio_ob
 extern bool		pdfioPageDictAddImage(pdfio_dict_t *dict, const char *name, pdfio_obj_t *obj) _PDFIO_PUBLIC;


-//
-// C++ magic...
-//
-
 #  ifdef __cplusplus
 }
 #  endif // __cplusplus
--- a/pdfio-crypto.c
+++ b/pdfio-crypto.c
@ -1,7 +1,7 @@
 //
 // Cryptographic support functions for PDFio.
 //
-// Copyright © 2021 by Michael R Sweet.
+// Copyright © 2021-2023 by Michael R Sweet.
 //
 // Licensed under Apache License v2.0.  See the file "LICENSE" for more
 // information.
@ -446,6 +446,7 @@ _pdfio_crypto_cb_t			// O  - Decryption callback or `NULL` for none

        // Initialize the RC4 context using 40 bits of the digest...
 	_pdfioCryptoRC4Init(&ctx->rc4, digest, 5);
+	*ivlen = 0;
 	return ((_pdfio_crypto_cb_t)_pdfioCryptoRC4Crypt);

    case PDFIO_ENCRYPTION_RC4_128 :
@ -663,7 +664,12 @@ _pdfioCryptoUnlock(
        length = 128;
    }
  }
-  // TODO: Implement AES-256 - V6 R6
+  else if (version == 6 && revision == 6)
+  {
+    // TODO: Implement AES-256 - V6 R6
+    pdf->encryption = PDFIO_ENCRYPTION_AES_256;
+    length          = 256;
+  }

  PDFIO_DEBUG("_pdfioCryptoUnlock: encryption=%d, length=%d\n", pdf->encryption, length);

@ -788,6 +794,8 @@ _pdfioCryptoUnlock(
    else
    {
      // TODO: Implement AES-256 security handler
+      _pdfioFileError(pdf, "Unable to unlock AES-256 encrypted file at this time.");
+      return (false);
    }

    // If we get here we need to try another password...
--- a/pdfio-dict.c
+++ b/pdfio-dict.c
@ -1,7 +1,7 @@
 //
 // PDF dictionary functions for PDFio.
 //
-// Copyright © 2021 by Michael R Sweet.
+// Copyright © 2021-2023 by Michael R Sweet.
 //
 // Licensed under Apache License v2.0.  See the file "LICENSE" for more
 // information.
@ -464,6 +464,47 @@ _pdfioDictGetValue(pdfio_dict_t *dict,	// I - Dictionary
 }


+//
+// 'pdfioDictIterateKeys()' - Iterate the keys in a dictionary.
+//
+// This function iterates the keys in a dictionary, calling the supplied
+// function "cb":
+//
+// ```
+// bool
+// my_dict_cb(pdfio_dict_t *dict, const char *key, void *cb_data)
+// {
+// ... "key" contains the dictionary key ...
+// ... return true to continue or false to stop ...
+// }
+// ```
+//
+// The iteration continues as long as the callback returns `true` or all keys
+// have been iterated.
+//
+
+void
+pdfioDictIterateKeys(
+    pdfio_dict_t    *dict,		// I - Dictionary
+    pdfio_dict_cb_t cb,			// I - Callback function
+    void            *cb_data)		// I - Callback data
+{
+  size_t	i;			// Looping var
+  _pdfio_pair_t	*pair;			// Current pair
+
+
+  // Range check input...
+  if (!dict || !cb)
+    return;
+
+  for (i = dict->num_pairs, pair = dict->pairs; i > 0; i --, pair ++)
+  {
+    if (!(cb)(dict, pair->key, cb_data))
+      break;
+  }
+}
+
+
 //
 // '_pdfioDictRead()' - Read a dictionary from a PDF file.
 //
@ -500,8 +541,15 @@ _pdfioDictRead(pdfio_file_t   *pdf,	// I - PDF file
      _pdfioFileError(pdf, "Invalid dictionary contents.");
      break;
    }
+    else if (_pdfioDictGetValue(dict, key + 1))
+    {
+      _pdfioFileError(pdf, "Duplicate dictionary key '%s'.", key + 1);
+      return (NULL);
+    }

    // Then get the next value...
+    PDFIO_DEBUG("_pdfioDictRead: Reading value for '%s'.\n", key + 1);
+
    if (!_pdfioValueRead(pdf, obj, tb, &value, depth))
    {
      _pdfioFileError(pdf, "Missing value for dictionary key.");
@ -707,7 +755,7 @@ pdfioDictSetNull(pdfio_dict_t *dict,	// I - Dictionary
 bool					// O - `true` on success, `false` on failure
 pdfioDictSetNumber(pdfio_dict_t  *dict,	// I - Dictionary
                   const char    *key,	// I - Key
-                   double         value)	// I - Value
+                   double        value)	// I - Value
 {
  _pdfio_value_t temp;			// New value

@ -891,9 +939,9 @@ _pdfioDictSetValue(

 #ifdef DEBUG
  PDFIO_DEBUG("_pdfioDictSetValue(%p): %lu pairs\n", (void *)dict, (unsigned long)dict->num_pairs);
-  PDFIO_DEBUG("_pdfioDictSetValue(%p): ", (void *)dict);
-  PDFIO_DEBUG_DICT(dict);
-  PDFIO_DEBUG("\n");
+//  PDFIO_DEBUG("_pdfioDictSetValue(%p): ", (void *)dict);
+//  PDFIO_DEBUG_DICT(dict);
+//  PDFIO_DEBUG("\n");
 #endif // DEBUG

  return (true);
--- a/pdfio-file.c
+++ b/pdfio-file.c
@ -1,7 +1,7 @@
 //
 // PDF file functions for PDFio.
 //
-// Copyright © 2021 by Michael R Sweet.
+// Copyright © 2021-2023 by Michael R Sweet.
 //
 // Licensed under Apache License v2.0.  See the file "LICENSE" for more
 // information.
@ -24,6 +24,7 @@
 static pdfio_obj_t	*add_obj(pdfio_file_t *pdf, size_t number, unsigned short generation, off_t offset);
 static int		compare_objmaps(_pdfio_objmap_t *a, _pdfio_objmap_t *b);
 static int		compare_objs(pdfio_obj_t **a, pdfio_obj_t **b);
+static const char	*get_info_string(pdfio_file_t *pdf, const char *key);
 static bool		load_obj_stream(pdfio_obj_t *obj);
 static bool		load_pages(pdfio_file_t *pdf, pdfio_obj_t *obj, size_t depth);
 static bool		load_xref(pdfio_file_t *pdf, off_t xref_offset, pdfio_password_cb_t password_cb, void *password_data);
@ -264,46 +265,26 @@ pdfioFileCreate(

  // Write a standard PDF header...
  if (!_pdfioFilePrintf(pdf, "%%PDF-%s\n%%\342\343\317\323\n", version))
-  {
-    pdfioFileClose(pdf);
-    unlink(filename);
-    return (NULL);
-  }
+    goto error;

  // Create the pages object...
  if ((dict = pdfioDictCreate(pdf)) == NULL)
-  {
-    pdfioFileClose(pdf);
-    unlink(filename);
-    return (NULL);
-  }
+    goto error;

  pdfioDictSetName(dict, "Type", "Pages");

  if ((pdf->pages_obj = pdfioFileCreateObj(pdf, dict)) == NULL)
-  {
-    pdfioFileClose(pdf);
-    unlink(filename);
-    return (NULL);
-  }
+    goto error;

  // Create the info object...
  if ((info_dict = pdfioDictCreate(pdf)) == NULL)
-  {
-    pdfioFileClose(pdf);
-    unlink(filename);
-    return (NULL);
-  }
+    goto error;

  pdfioDictSetDate(info_dict, "CreationDate", time(NULL));
  pdfioDictSetString(info_dict, "Producer", "pdfio/" PDFIO_VERSION);

  if ((pdf->info_obj = pdfioFileCreateObj(pdf, info_dict)) == NULL)
-  {
-    pdfioFileClose(pdf);
-    unlink(filename);
-    return (NULL);
-  }
+    goto error;

  // Create random file ID values...
  _pdfioCryptoMakeRandom(id_value, sizeof(id_value));
@ -315,6 +296,15 @@ pdfioFileCreate(
  }

  return (pdf);
+
+  // Common error handling code...
+  error:
+
+  pdfioFileClose(pdf);
+
+  unlink(filename);
+
+  return (NULL);
 }


@ -535,41 +525,26 @@ pdfioFileCreateOutput(

  // Write a standard PDF header...
  if (!_pdfioFilePrintf(pdf, "%%PDF-%s\n%%\342\343\317\323\n", version))
-  {
-    pdfioFileClose(pdf);
-    return (NULL);
-  }
+    goto error;

  // Create the pages object...
  if ((dict = pdfioDictCreate(pdf)) == NULL)
-  {
-    pdfioFileClose(pdf);
-    return (NULL);
-  }
+    goto error;

  pdfioDictSetName(dict, "Type", "Pages");

  if ((pdf->pages_obj = pdfioFileCreateObj(pdf, dict)) == NULL)
-  {
-    pdfioFileClose(pdf);
-    return (NULL);
-  }
+    goto error;

  // Create the info object...
  if ((info_dict = pdfioDictCreate(pdf)) == NULL)
-  {
-    pdfioFileClose(pdf);
-    return (NULL);
-  }
+    goto error;

  pdfioDictSetDate(info_dict, "CreationDate", time(NULL));
  pdfioDictSetString(info_dict, "Producer", "pdfio/" PDFIO_VERSION);

  if ((pdf->info_obj = pdfioFileCreateObj(pdf, info_dict)) == NULL)
-  {
-    pdfioFileClose(pdf);
-    return (NULL);
-  }
+    goto error;

  // Create random file ID values...
  _pdfioCryptoMakeRandom(id_value, sizeof(id_value));
@ -581,6 +556,13 @@ pdfioFileCreateOutput(
  }

  return (pdf);
+
+  // Common error handling code...
+  error:
+
+  pdfioFileClose(pdf);
+
+  return (NULL);
 }


@ -657,6 +639,194 @@ pdfioFileCreatePage(pdfio_file_t *pdf,	// I - PDF file
 }


+//
+// 'pdfioFileCreateTemporary()' - Create a temporary PDF file.
+//
+// This function creates a PDF file with a unique filename in the current
+// temporary directory.  The temporary file is stored in the string "buffer" an
+// will have a ".pdf" extension.  Otherwise, this function works the same as
+// the @link pdfioFileCreate@ function.
+//
+// @since PDFio v1.1@
+//
+
+pdfio_file_t *
+pdfioFileCreateTemporary(
+    char             *buffer,		// I - Filename buffer
+    size_t           bufsize,		// I - Size of filename buffer
+    const char       *version,		// I - PDF version number or `NULL` for default (2.0)
+    pdfio_rect_t     *media_box,	// I - Default MediaBox for pages
+    pdfio_rect_t     *crop_box,		// I - Default CropBox for pages
+    pdfio_error_cb_t error_cb,		// I - Error callback or `NULL` for default
+    void             *error_data)	// I - Error callback data, if any
+{
+  pdfio_file_t	*pdf;			// PDF file
+  pdfio_dict_t	*dict;			// Dictionary for pages object
+  pdfio_dict_t	*info_dict;		// Dictionary for information object
+  unsigned char	id_value[16];		// File ID value
+  int		i;			// Looping var
+  const char	*tmpdir;		// Temporary directory
+#if _WIN32 || defined(__APPLE__)
+  char		tmppath[256];		// Temporary directory path
+#endif // _WIN32 || __APPLE__
+  unsigned	tmpnum;			// Temporary filename number
+
+
+  // Range check input...
+  if (!buffer || bufsize < 32)
+  {
+    if (buffer)
+      *buffer = '\0';
+    return (NULL);
+  }
+
+  if (!version)
+    version = "2.0";
+
+  if (!error_cb)
+  {
+    error_cb   = _pdfioFileDefaultError;
+    error_data = NULL;
+  }
+
+  // Allocate a PDF file structure...
+  if ((pdf = (pdfio_file_t *)calloc(1, sizeof(pdfio_file_t))) == NULL)
+  {
+    pdfio_file_t temp;			// Dummy file
+    char	message[8192];		// Message string
+
+    temp.filename = (char *)"temporary.pdf";
+    snprintf(message, sizeof(message), "Unable to allocate memory for PDF file - %s", strerror(errno));
+    (error_cb)(&temp, message, error_data);
+
+    *buffer = '\0';
+
+    return (NULL);
+  }
+
+  // Create the file...
+#if _WIN32
+  if ((tmpdir = getenv("TEMP")) == NULL)
+  {
+    GetTempPathA(sizeof(tmppath), tmppath);
+    tmpdir = tmppath;
+  }
+
+#elif defined(__APPLE__)
+  if ((tmpdir = getenv("TMPDIR")) != NULL && access(tmpdir, W_OK))
+    tmpdir = NULL;
+
+  if (!tmpdir)
+  {
+    // Grab the per-process temporary directory for sandboxed apps...
+#  ifdef _CS_DARWIN_USER_TEMP_DIR
+    if (confstr(_CS_DARWIN_USER_TEMP_DIR, tmppath, sizeof(tmppath)))
+      tmpdir = tmppath;
+    else
+#  endif // _CS_DARWIN_USER_TEMP_DIR
+      tmpdir = "/private/tmp";
+  }
+
+#else
+  if ((tmpdir = getenv("TMPDIR")) == NULL || access(tmpdir, W_OK))
+    tmpdir = "/tmp";
+#endif // _WIN32
+
+  for (i = 0; i < 1000; i ++)
+  {
+    _pdfioCryptoMakeRandom((uint8_t *)&tmpnum, sizeof(tmpnum));
+    snprintf(buffer, bufsize, "%s/%08x.pdf", tmpdir, tmpnum);
+    if ((pdf->fd = open(buffer, O_WRONLY | O_BINARY | O_CREAT | O_TRUNC | O_EXCL, 0666)) >= 0)
+      break;
+  }
+
+  pdf->filename = strdup(buffer);
+
+  if (i >= 1000)
+  {
+    _pdfioFileError(pdf, "Unable to create file - %s", strerror(errno));
+    free(pdf->filename);
+    free(pdf);
+    *buffer = '\0';
+    return (NULL);
+  }
+
+  pdf->version     = strdup(version);
+  pdf->mode        = _PDFIO_MODE_WRITE;
+  pdf->error_cb    = error_cb;
+  pdf->error_data  = error_data;
+  pdf->permissions = PDFIO_PERMISSION_ALL;
+  pdf->bufptr      = pdf->buffer;
+  pdf->bufend      = pdf->buffer + sizeof(pdf->buffer);
+
+  if (media_box)
+  {
+    pdf->media_box = *media_box;
+  }
+  else
+  {
+    // Default to "universal" size (intersection of A4 and US Letter)
+    pdf->media_box.x2 = 210.0 * 72.0f / 25.4f;
+    pdf->media_box.y2 = 11.0f * 72.0f;
+  }
+
+  if (crop_box)
+  {
+    pdf->crop_box = *crop_box;
+  }
+  else
+  {
+    // Default to "universal" size (intersection of A4 and US Letter)
+    pdf->crop_box.x2 = 210.0 * 72.0f / 25.4f;
+    pdf->crop_box.y2 = 11.0f * 72.0f;
+  }
+
+  // Write a standard PDF header...
+  if (!_pdfioFilePrintf(pdf, "%%PDF-%s\n%%\342\343\317\323\n", version))
+    goto error;
+
+  // Create the pages object...
+  if ((dict = pdfioDictCreate(pdf)) == NULL)
+    goto error;
+
+  pdfioDictSetName(dict, "Type", "Pages");
+
+  if ((pdf->pages_obj = pdfioFileCreateObj(pdf, dict)) == NULL)
+    goto error;
+
+  // Create the info object...
+  if ((info_dict = pdfioDictCreate(pdf)) == NULL)
+    goto error;
+
+  pdfioDictSetDate(info_dict, "CreationDate", time(NULL));
+  pdfioDictSetString(info_dict, "Producer", "pdfio/" PDFIO_VERSION);
+
+  if ((pdf->info_obj = pdfioFileCreateObj(pdf, info_dict)) == NULL)
+    goto error;
+
+  // Create random file ID values...
+  _pdfioCryptoMakeRandom(id_value, sizeof(id_value));
+
+  if ((pdf->id_array = pdfioArrayCreate(pdf)) != NULL)
+  {
+    pdfioArrayAppendBinary(pdf->id_array, id_value, sizeof(id_value));
+    pdfioArrayAppendBinary(pdf->id_array, id_value, sizeof(id_value));
+  }
+
+  return (pdf);
+
+  // Common error handling code...
+  error:
+
+  pdfioFileClose(pdf);
+
+  unlink(buffer);
+  *buffer = '\0';
+
+  return (NULL);
+}
+
+
 //
 // '_pdfioFileFindMappedObj()' - Find a mapped object.
 //
@ -723,7 +893,7 @@ pdfioFileFindObj(
 const char *				// O - Author or `NULL` for none
 pdfioFileGetAuthor(pdfio_file_t *pdf)	// I - PDF file
 {
-  return (pdf && pdf->info_obj ? pdfioDictGetString(pdf->info_obj->value.value.dict, "Author") : NULL);
+  return (get_info_string(pdf, "Author"));
 }


@ -735,7 +905,7 @@ time_t					// O - Creation date or `0` for none
 pdfioFileGetCreationDate(
    pdfio_file_t *pdf)			// I - PDF file
 {
-  return (pdf && pdf->info_obj ? pdfioDictGetDate(pdf->info_obj->value.value.dict, "CreationDate") : 0);
+  return (pdf && pdf->info_obj ? pdfioDictGetDate(pdfioObjGetDict(pdf->info_obj), "CreationDate") : 0);
 }


@ -746,7 +916,7 @@ pdfioFileGetCreationDate(
 const char *				// O - Creator string or `NULL` for none
 pdfioFileGetCreator(pdfio_file_t *pdf)	// I - PDF file
 {
-  return (pdf && pdf->info_obj ? pdfioDictGetString(pdf->info_obj->value.value.dict, "Creator") : NULL);
+  return (get_info_string(pdf, "Creator"));
 }


@ -768,7 +938,7 @@ pdfioFileGetID(pdfio_file_t *pdf)	// I - PDF file
 const char *				// O - Keywords string or `NULL` for none
 pdfioFileGetKeywords(pdfio_file_t *pdf)	// I - PDF file
 {
-  return (pdf && pdf->info_obj ? pdfioDictGetString(pdf->info_obj->value.value.dict, "Keywords") : NULL);
+  return (get_info_string(pdf, "Keywords"));
 }


@ -872,7 +1042,7 @@ pdfioFileGetPermissions(
 const char *				// O - Producer string or `NULL` for none
 pdfioFileGetProducer(pdfio_file_t *pdf)	// I - PDF file
 {
-  return (pdf && pdf->info_obj ? pdfioDictGetString(pdf->info_obj->value.value.dict, "Producer") : NULL);
+  return (get_info_string(pdf, "Producer"));
 }


@ -883,7 +1053,7 @@ pdfioFileGetProducer(pdfio_file_t *pdf)	// I - PDF file
 const char *				// O - Subject or `NULL` for none
 pdfioFileGetSubject(pdfio_file_t *pdf)	// I - PDF file
 {
-  return (pdf && pdf->info_obj ? pdfioDictGetString(pdf->info_obj->value.value.dict, "Subject") : NULL);
+  return (get_info_string(pdf, "Subject"));
 }


@ -894,7 +1064,7 @@ pdfioFileGetSubject(pdfio_file_t *pdf)	// I - PDF file
 const char *				// O - Title or `NULL` for none
 pdfioFileGetTitle(pdfio_file_t *pdf)	// I - PDF file
 {
-  return (pdf && pdf->info_obj ? pdfioDictGetString(pdf->info_obj->value.value.dict, "Title") : NULL);
+  return (get_info_string(pdf, "Title"));
 }


@ -1237,6 +1407,51 @@ compare_objs(pdfio_obj_t **a,		// I - First object
 }


+//
+// 'get_info_string()' - Get a string value from the Info dictionary.
+//
+// This function also handles converting binary strings to C strings, which
+// occur in encrypted PDF files.
+//
+
+static const char *			// O - String or `NULL` if not found
+get_info_string(pdfio_file_t *pdf,	// I - PDF file
+                const char   *key)	// I - Dictionary key
+{
+  pdfio_dict_t	*dict;			// Info dictionary
+  _pdfio_value_t *value;		// Value
+
+  // Range check input...
+  if (!pdf || !pdf->info_obj || (dict = pdfioObjGetDict(pdf->info_obj)) == NULL || (value = _pdfioDictGetValue(dict, key)) == NULL)
+    return (NULL);
+
+  // If we already have a value, return it...
+  if (value->type == PDFIO_VALTYPE_NAME || value->type == PDFIO_VALTYPE_STRING)
+  {
+    return (value->value.string);
+  }
+  else if (value->type == PDFIO_VALTYPE_BINARY && value->value.binary.datalen < 4096)
+  {
+    // Convert binary string to regular string...
+    char	temp[4096];		// Temporary string
+
+    memcpy(temp, value->value.binary.data, value->value.binary.datalen);
+    temp[value->value.binary.datalen] = '\0';
+
+    free(value->value.binary.data);
+    value->type         = PDFIO_VALTYPE_STRING;
+    value->value.string = pdfioStringCreate(pdf, temp);
+
+    return (value->value.string);
+  }
+  else
+  {
+    // Something else that is not a string...
+    return (NULL);
+  }
+}
+
+
 //
 // 'load_obj_stream()' - Load an object stream.
 //
@ -1299,6 +1514,7 @@ load_obj_stream(pdfio_obj_t *obj)	// I - Object to load

    // Skip offset
    _pdfioTokenGet(&tb, buffer, sizeof(buffer));
+    PDFIO_DEBUG("load_obj_stream: %ld at offset %s\n", (long)number, buffer);
  }

  if (!buffer[0])
@ -1522,6 +1738,10 @@ load_xref(
        return (false);
      }

+      PDFIO_DEBUG("load_xref: tb.bufptr=%p, tb.bufend=%p, tb.bufptr[0]=0x%02x, tb.bufptr[0]=0x%02x\n", tb.bufptr, tb.bufend, tb.bufptr[0], tb.bufptr[1]);
+      if (tb.bufptr && tb.bufptr < tb.bufend && (tb.bufptr[0] == 0x0d || tb.bufptr[0] == 0x0a))
+	tb.bufptr ++;			// Skip trailing CR or LF after token
+
      _pdfioTokenFlush(&tb);

      obj->stream_offset = _pdfioFileTell(pdf);
@ -1544,9 +1764,9 @@ load_xref(
      w_2     = w[0];
      w_3     = w[0] + w[1];

-      if (w[1] == 0 || w[2] > 2 || w[0] > sizeof(buffer) || w[1] > sizeof(buffer) || w[2] > sizeof(buffer) || w_total > sizeof(buffer))
+      if (w[1] == 0 || w[2] > 4 || w[0] > sizeof(buffer) || w[1] > sizeof(buffer) || w[2] > sizeof(buffer) || w_total > sizeof(buffer))
      {
-	_pdfioFileError(pdf, "Cross-reference stream has invalid W key.");
+	_pdfioFileError(pdf, "Cross-reference stream has invalid W key [%u %u %u].", (unsigned)w[0], (unsigned)w[1], (unsigned)w[2]);
 	return (false);
      }

@ -1586,9 +1806,11 @@ load_xref(
 	    }
 	  }

+          // Offset
 	  for (i = 1, offset = buffer[w_2]; i < w[1]; i ++)
 	    offset = (offset << 8) | buffer[w_2 + i];

+          // Generation number
 	  switch (w[2])
 	  {
 	    default :
@ -1600,6 +1822,19 @@ load_xref(
 	    case 2 :
 		generation = (buffer[w_3] << 8) | buffer[w_3 + 1];
 		break;
+	    case 3 :
+	        // Issue #46: Stupid Microsoft PDF generator using 3 bytes to
+	        // encode 16-bit generation numbers == 0 (probably a lazy coder
+	        // stuffing things into an array of 64-bit unsigned integers)
+		generation = (buffer[w_3] << 16) | (buffer[w_3 + 1] << 8) | buffer[w_3 + 2];
+		if (generation > 65535)
+		  generation = 65535;
+		break;
+	    case 4 : // Even stupider :)
+		generation = (buffer[w_3] << 24) | (buffer[w_3 + 1] << 16) | (buffer[w_3 + 2] << 8) | buffer[w_3 + 3];
+		if (generation > 65535)
+		  generation = 65535;
+		break;
 	  }

 	  // Create a placeholder for the object in memory...
@ -1631,8 +1866,18 @@ load_xref(
 		break;
 	    }

-	    if (i >= num_sobjs && num_sobjs < (sizeof(sobjs) / sizeof(sobjs[0])))
-	      sobjs[num_sobjs ++] = (size_t)offset;
+	    if (i >= num_sobjs)
+	    {
+	      if (num_sobjs < (sizeof(sobjs) / sizeof(sobjs[0])))
+	      {
+		sobjs[num_sobjs ++] = (size_t)offset;
+	      }
+	      else
+	      {
+		_pdfioFileError(pdf, "Too many object streams.");
+	        return (false);
+	      }
+	    }
 	  }
 	  else if (!current)
 	  {
@ -1680,12 +1925,12 @@ load_xref(
 	}
      }
    }
-    else if (!strcmp(line, "xref"))
+    else if (!strncmp(line, "xref", 4) && !line[4] || isspace(line[4] & 255))
    {
      // Read the xref tables
      while (_pdfioFileGets(pdf, line, sizeof(line)))
      {
-	if (!strcmp(line, "trailer"))
+	if (!strncmp(line, "trailer", 7) && !line[7] || isspace(line[7] & 255))
 	  break;
 	else if (!line[0])
 	  continue;
@ -1750,7 +1995,7 @@ load_xref(
 	}
      }

-      if (strcmp(line, "trailer"))
+      if (strncmp(line, "trailer", 7))
      {
 	_pdfioFileError(pdf, "Missing trailer.");
 	return (false);
@ -1769,6 +2014,8 @@ load_xref(
 	return (false);
      }

+      PDFIO_DEBUG("load_xref: Got trailer dict.\n");
+
      _pdfioTokenFlush(&tb);

      if (!pdf->trailer_dict)
--- a/pdfio-object.c
+++ b/pdfio-object.c
@ -1,7 +1,7 @@
 //
 // PDF object functions for PDFio.
 //
-// Copyright © 2021 by Michael R Sweet.
+// Copyright © 2021-2023 by Michael R Sweet.
 //
 // Licensed under Apache License v2.0.  See the file "LICENSE" for more
 // information.
@ -33,8 +33,14 @@ pdfioObjClose(pdfio_obj_t *obj)		// I - Object
  if (!obj)
    return (false);

+  // Clear the current object pointer...
+  obj->pdf->current_obj = NULL;
+
  if (obj->pdf->mode != _PDFIO_MODE_WRITE)
-    return (true);			// Nothing to do when reading
+  {
+    // Nothing to do when reading
+    return (true);
+  }

  // Write what remains for the object...
  if (!obj->offset)
@ -165,6 +171,12 @@ pdfioObjCreateStream(
    return (NULL);
  }

+  if (obj->pdf->current_obj)
+  {
+    _pdfioFileError(obj->pdf, "Another object (%u) is already open.", (unsigned)obj->pdf->current_obj->number);
+    return (NULL);
+  }
+
  // Write the header...
  if (!_pdfioDictGetValue(obj->value.value.dict, "Length"))
  {
@ -193,7 +205,8 @@ pdfioObjCreateStream(
  if (!_pdfioFilePuts(obj->pdf, "stream\n"))
    return (NULL);

-  obj->stream_offset = _pdfioFileTell(obj->pdf);
+  obj->stream_offset    = _pdfioFileTell(obj->pdf);
+  obj->pdf->current_obj = obj;

  // Return the new stream...
  return (_pdfioStreamCreate(obj, length_obj, filter));
@ -407,6 +420,9 @@ _pdfioObjLoad(pdfio_obj_t *obj)		// I - Object
  }

  ptr += 3;
+  while (*ptr && isspace(*ptr & 255))
+    ptr ++;
+
  _pdfioFileConsume(obj->pdf, (size_t)(ptr - line));

  // Then grab the object value...
@ -425,11 +441,15 @@ _pdfioObjLoad(pdfio_obj_t *obj)		// I - Object
    return (false);
  }

+  PDFIO_DEBUG("_pdfioObjLoad: tb.bufptr=%p, tb.bufend=%p, tb.bufptr[0]=0x%02x, tb.bufptr[0]=0x%02x\n", tb.bufptr, tb.bufend, tb.bufptr[0], tb.bufptr[1]);
+  if (tb.bufptr && tb.bufptr < tb.bufend && (tb.bufptr[0] == 0x0d || tb.bufptr[0] == 0x0a))
+    tb.bufptr ++;			// Skip trailing CR or LF after token
+
  _pdfioTokenFlush(&tb);

  if (!strcmp(line, "stream"))
  {
-    // Yes, save its location...
+    // Yes, this is an embedded stream so save its location...
    obj->stream_offset = _pdfioFileTell(obj->pdf);
    PDFIO_DEBUG("_pdfioObjLoad: stream_offset=%lu.\n", (unsigned long)obj->stream_offset);
  }
@ -454,6 +474,12 @@ pdfioObjOpenStream(pdfio_obj_t *obj,	// I - Object
  if (!obj)
    return (NULL);

+  if (obj->pdf->current_obj)
+  {
+    _pdfioFileError(obj->pdf, "Another object (%u) is already open.", (unsigned)obj->pdf->current_obj->number);
+    return (NULL);
+  }
+
  // Make sure we've loaded the object dictionary...
  if (!obj->value.type)
  {
@ -466,6 +492,8 @@ pdfioObjOpenStream(pdfio_obj_t *obj,	// I - Object
    return (NULL);

  // Open the stream...
+  obj->pdf->current_obj = obj;
+
  return (_pdfioStreamOpen(obj, decode));
 }

--- a/pdfio-page.c
+++ b/pdfio-page.c
@ -1,7 +1,7 @@
 //
 // PDF page functions for PDFio.
 //
-// Copyright © 2021 by Michael R Sweet.
+// Copyright © 2021-2022 by Michael R Sweet.
 //
 // Licensed under Apache License v2.0.  See the file "LICENSE" for more
 // information.
@ -14,6 +14,13 @@
 #include "pdfio-private.h"


+//
+// Local functions...
+//
+
+static _pdfio_value_t	*get_contents(pdfio_obj_t *page);
+
+
 //
 // 'pdfioPageCopy()' - Copy a page to a PDF file.
 //
@ -47,3 +54,74 @@ pdfioPageCopy(pdfio_file_t *pdf,	// I - PDF file
  else
    return (_pdfioFileAddPage(pdf, dstpage));
 }
+
+
+//
+// 'pdfioPageGetNumStreams()' - Get the number of content streams for a page object.
+//
+
+size_t					// O - Number of streams
+pdfioPageGetNumStreams(
+    pdfio_obj_t *page)			// I - Page object
+{
+  _pdfio_value_t *contents = get_contents(page);
+					// Contents value
+
+
+  if (!contents)
+    return (0);
+  else if (contents->type == PDFIO_VALTYPE_ARRAY)
+    return (pdfioArrayGetSize(contents->value.array));
+  else
+    return (1);
+}
+
+
+//
+// 'pdfioPageOpenStream()' - Open a content stream for a page.
+//
+
+pdfio_stream_t *			// O - Stream
+pdfioPageOpenStream(
+    pdfio_obj_t *page,			// I - Page object
+    size_t      n,			// I - Stream index (0-based)
+    bool        decode)			// I - `true` to decode/decompress stream
+{
+  _pdfio_value_t *contents = get_contents(page);
+					// Contents value
+
+
+  if (!contents)
+    return (NULL);
+  else if (contents->type == PDFIO_VALTYPE_ARRAY && n < pdfioArrayGetSize(contents->value.array))
+    return (pdfioObjOpenStream(pdfioArrayGetObj(contents->value.array, n), decode));
+  else if (n)
+    return (NULL);
+  else
+    return (pdfioObjOpenStream(pdfioFileFindObj(page->pdf, contents->value.indirect.number), decode));
+}
+
+
+//
+// 'get_contents()' - Get a page's Contents value.
+//
+
+static _pdfio_value_t *			// O - Value or NULL on error
+get_contents(pdfio_obj_t *page)		// I - Page object
+{
+  // Range check input...
+  if (!page)
+    return (NULL);
+
+  // Load the page object as needed...
+  if (page->value.type == PDFIO_VALTYPE_NONE)
+  {
+    if (!_pdfioObjLoad(page))
+      return (NULL);
+  }
+
+  if (page->value.type != PDFIO_VALTYPE_DICT)
+    return (NULL);
+
+  return (_pdfioDictGetValue(page->value.value.dict, "Contents"));
+}
--- a/pdfio-private.h
+++ b/pdfio-private.h
@ -1,7 +1,7 @@
 //
 // Private header file for PDFio.
 //
-// Copyright © 2021 by Michael R Sweet.
+// Copyright © 2021-2023 by Michael R Sweet.
 //
 // Licensed under Apache License v2.0.  See the file "LICENSE" for more
 // information.
@ -9,41 +9,22 @@

 #ifndef PDFIO_PRIVATE_H
 #  define PDFIO_PRIVATE_H
-
-//
-// Include necessary headers...
-//
-
 #  ifdef _WIN32
-/*
- * Disable bogus VS warnings/errors...
- */
-
-#    define _CRT_SECURE_NO_WARNINGS
+#    define _CRT_SECURE_NO_WARNINGS	// Disable bogus VS warnings/errors...
 #  endif // _WIN32
-
 #  include "pdfio.h"
 #  include <stdarg.h>
 #  include <stdint.h>
 #  include <string.h>
+#  include <ctype.h>
 #  include <errno.h>
 #  include <inttypes.h>
 #  include <fcntl.h>
 #  ifdef _WIN32
 #    include <io.h>
 #    include <direct.h>
-
-/*
- * Microsoft renames the POSIX functions to _name, and introduces
- * a broken compatibility layer using the original names.  As a result,
- * random crashes can occur when, for example, strdup() allocates memory
- * from a different heap than used by malloc() and free().
- *
- * To avoid moronic problems like this, we #define the POSIX function
- * names to the corresponding non-standard Microsoft names.
- */
-
-#    define access	_access
+#    include <windows.h> // GetTempPathA
+#    define access	_access		// Map standard POSIX/C99 names
 #    define close	_close
 #    define fileno	_fileno
 #    define lseek	_lseek
@ -56,25 +37,18 @@
 #    define unlink	_unlink
 #    define vsnprintf	_vsnprintf
 #    define write	_write
-
-/*
- * Map various parameters for POSIX...
- */
-
-#    define F_OK	00
+#    define F_OK	00		// POSIX parameters/flags
 #    define W_OK	02
 #    define R_OK	04
-#    define O_RDONLY	_O_RDONLY
+#    define O_RDONLY	_O_RDONLY	// Map standard POSIX open flags
 #    define O_WRONLY	_O_WRONLY
 #    define O_CREAT	_O_CREAT
 #    define O_TRUNC	_O_TRUNC
 #    define O_BINARY	_O_BINARY
 #  else // !_WIN32
 #    include <unistd.h>
-#    define O_BINARY	0
+#    define O_BINARY	0		// Map Windows-specific open flag
 #  endif // _WIN32
-#  include <string.h>
-#  include <ctype.h>
 #  include <zlib.h>


@ -289,7 +263,8 @@ struct _pdfio_file_s			// PDF file structure
  pdfio_dict_t	**dicts;		// Dictionaries
  size_t	num_objs,		// Number of objects
 		alloc_objs;		// Allocated objects
-  pdfio_obj_t	**objs;			// Objects
+  pdfio_obj_t	**objs,			// Objects
+		*current_obj;		// Current object being written/read
  size_t	num_objmaps,		// Number of object maps
 		alloc_objmaps;		// Allocated object maps
  _pdfio_objmap_t *objmaps;		// Object maps
@ -410,4 +385,5 @@ extern void		_pdfioValueDelete(_pdfio_value_t *v) _PDFIO_INTERNAL;
 extern _pdfio_value_t	*_pdfioValueRead(pdfio_file_t *pdf, pdfio_obj_t *obj, _pdfio_token_t *ts, _pdfio_value_t *v, size_t depth) _PDFIO_INTERNAL;
 extern bool		_pdfioValueWrite(pdfio_file_t *pdf, pdfio_obj_t *obj, _pdfio_value_t *v, off_t *length) _PDFIO_INTERNAL;

+
 #endif // !PDFIO_PRIVATE_H
--- a/pdfio-sha256.c
+++ b/pdfio-sha256.c
@ -1,7 +1,7 @@
 //
 // SHA-256 functions for PDFio.
 //
-// Copyright © 2021 by Michael R Sweet.
+// Copyright © 2021-2023 by Michael R Sweet.
 // Copyright © 2011 IETF Trust and the persons identified as authors of the
 // code.  All rights reserved.
 //
@ -217,7 +217,7 @@ _pdfioCryptoSHA256Append(_pdfio_sha256_t *context, const uint8_t *message_array,
 */
 void
 _pdfioCryptoSHA256Finish(_pdfio_sha256_t *context,
-                 uint8_t Message_Digest[SHA256HashSize])
+                 uint8_t *Message_Digest)
 {
  SHA224_256ResultN(context, Message_Digest, SHA256HashSize);
 }
--- a/pdfio-stream.c
+++ b/pdfio-stream.c
@ -1,7 +1,7 @@
 //
 // PDF stream functions for PDFio.
 //
-// Copyright © 2021 by Michael R Sweet.
+// Copyright © 2021-2023 by Michael R Sweet.
 //
 // Licensed under Apache License v2.0.  See the file "LICENSE" for more
 // information.
@ -174,6 +174,8 @@ pdfioStreamClose(pdfio_stream_t *st)	// I - Stream

  done:

+  st->pdf->current_obj = NULL;
+
  free(st->prbuffer);
  free(st->psbuffer);
  free(st);
@ -364,6 +366,13 @@ pdfioStreamConsume(pdfio_stream_t *st,	// I - Stream
 //
 // 'pdfioStreamGetToken()' - Read a single PDF token from a stream.
 //
+// This function reads a single PDF token from a stream.  Operator tokens,
+// boolean values, and numbers are returned as-is in the provided string buffer.
+// String values start with the opening parenthesis ('(') but have all escaping
+// resolved and the terminating parenthesis removed.  Hexadecimal string values
+// start with the opening angle bracket ('<') and have all whitespace and the
+// terminating angle bracket removed.
+//

 bool					// O - `true` on success, `false` on EOF
 pdfioStreamGetToken(
@ -372,6 +381,7 @@ pdfioStreamGetToken(
    size_t         bufsize)		// I - Size of string buffer
 {
  _pdfio_token_t	tb;		// Token buffer/stack
+  bool			ret;		// Return value


  // Range check input...
@ -381,7 +391,10 @@ pdfioStreamGetToken(
  // Read using the token engine...
  _pdfioTokenInit(&tb, st->pdf, (_pdfio_tconsume_cb_t)pdfioStreamConsume, (_pdfio_tpeek_cb_t)pdfioStreamPeek, st);

-  return (_pdfioTokenRead(&tb, buffer, bufsize));
+  ret = _pdfioTokenRead(&tb, buffer, bufsize);
+  _pdfioTokenFlush(&tb);
+
+  return (ret);
 }


@ -439,6 +452,7 @@ _pdfioStreamOpen(pdfio_obj_t *obj,	// I - Object
      return (NULL);
    }

+    PDFIO_DEBUG("_pdfioStreamOpen: ivlen=%d\n", (int)ivlen);
    if (ivlen > 0)
      _pdfioFileConsume(st->pdf, ivlen);

@ -451,11 +465,19 @@ _pdfioStreamOpen(pdfio_obj_t *obj,	// I - Object
    // Try to decode/decompress the contents of this object...
    const char	*filter = pdfioDictGetName(dict, "Filter");
 					// Filter value
+    pdfio_array_t *fa = pdfioDictGetArray(dict, "Filter");
+					// Filter array
+
+    if (!filter && fa && pdfioArrayGetSize(fa) == 1)
+    {
+      // Support single-valued arrays...
+      filter = pdfioArrayGetName(fa, 0);
+    }

    if (!filter)
    {
      // No single filter name, do we have a compound filter?
-      if (pdfioDictGetArray(dict, "Filter"))
+      if (fa)
      {
 	// TODO: Implement compound filters...
 	_pdfioFileError(st->pdf, "Unsupported compound stream filter.");
@ -546,6 +568,7 @@ _pdfioStreamOpen(pdfio_obj_t *obj,	// I - Object
      else
        st->predictor = _PDFIO_PREDICTOR_NONE;

+      PDFIO_DEBUG("_pdfioStreamOpen: pos=%ld\n", (long)_pdfioFileTell(st->pdf));
      if (sizeof(st->cbuffer) > st->remaining)
 	rbytes = _pdfioFileRead(st->pdf, st->cbuffer, st->remaining);
      else
@ -566,12 +589,6 @@ _pdfioStreamOpen(pdfio_obj_t *obj,	// I - Object
      st->flate.next_in  = (Bytef *)st->cbuffer;
      st->flate.avail_in = (uInt)rbytes;

-      if (st->cbuffer[0] == 0x0a)
-      {
-        st->flate.next_in ++;		// Skip newline
-        st->flate.avail_in --;
-      }
-
      PDFIO_DEBUG("_pdfioStreamOpen: avail_in=%u, cbuffer=<%02X%02X%02X%02X%02X%02X%02X%02X...>\n", st->flate.avail_in, st->cbuffer[0], st->cbuffer[1], st->cbuffer[2], st->cbuffer[3], st->cbuffer[4], st->cbuffer[5], st->cbuffer[6], st->cbuffer[7]);

      if ((status = inflateInit(&(st->flate))) != Z_OK)
@ -995,6 +1012,7 @@ stream_read(pdfio_stream_t *st,		// I - Stream
            size_t         bytes)	// I - Number of bytes to read
 {
  ssize_t	rbytes;			// Bytes read
+  uInt		avail_in, avail_out;	// Previous flate values


  if (st->filter == PDFIO_FILTER_NONE)
@ -1047,11 +1065,19 @@ stream_read(pdfio_stream_t *st,		// I - Stream
      st->flate.next_out  = (Bytef *)buffer;
      st->flate.avail_out = (uInt)bytes;

+      avail_in  = st->flate.avail_in;
+      avail_out = st->flate.avail_out;
+
      if ((status = inflate(&(st->flate), Z_NO_FLUSH)) < Z_OK)
      {
 	_pdfioFileError(st->pdf, "Unable to decompress stream data: %s", zstrerror(status));
 	return (-1);
      }
+      else if (avail_in == st->flate.avail_in && avail_out == st->flate.avail_out)
+      {
+	_pdfioFileError(st->pdf, "Corrupt stream data.");
+	return (-1);
+      }

      return (st->flate.next_out - (Bytef *)buffer);
    }
@ -1100,12 +1126,15 @@ stream_read(pdfio_stream_t *st,		// I - Stream
 	  st->flate.avail_in = (uInt)rbytes;
 	}

+        avail_in  = st->flate.avail_in;
+        avail_out = st->flate.avail_out;
+
 	if ((status = inflate(&(st->flate), Z_NO_FLUSH)) < Z_OK)
 	{
 	  _pdfioFileError(st->pdf, "Unable to decompress stream data: %s", zstrerror(status));
 	  return (-1);
 	}
-	else if (status == Z_STREAM_END)
+	else if (status == Z_STREAM_END || (avail_in == st->flate.avail_in && avail_out == st->flate.avail_out))
 	  break;
      }

@ -1167,12 +1196,15 @@ stream_read(pdfio_stream_t *st,		// I - Stream
 	  st->flate.avail_in = (uInt)rbytes;
 	}

+        avail_in  = st->flate.avail_in;
+        avail_out = st->flate.avail_out;
+
 	if ((status = inflate(&(st->flate), Z_NO_FLUSH)) < Z_OK)
 	{
 	  _pdfioFileError(st->pdf, "Unable to decompress stream data: %s", zstrerror(status));
 	  return (-1);
 	}
-	else if (status == Z_STREAM_END)
+	else if (status == Z_STREAM_END || (avail_in == st->flate.avail_in && avail_out == st->flate.avail_out))
 	  break;
      }

--- a/pdfio-token.c
+++ b/pdfio-token.c
@ -1,7 +1,7 @@
 //
 // PDF token parsing functions for PDFio.
 //
-// Copyright © 2021 by Michael R Sweet.
+// Copyright © 2021-2023 by Michael R Sweet.
 //
 // Licensed under Apache License v2.0.  See the file "LICENSE" for more
 // information.
@ -129,9 +129,20 @@ _pdfioTokenGet(_pdfio_token_t *tb,	// I - Token buffer/stack
  if (tb->num_tokens > 0)
  {
    // Yes, return it...
+    size_t len;				// Length of token
+
    tb->num_tokens --;
-    strncpy(buffer, tb->tokens[tb->num_tokens], bufsize - 1);
-    buffer[bufsize - 1] = '\0';
+
+    if ((len = strlen(tb->tokens[tb->num_tokens])) > (bufsize - 1))
+    {
+      // Value too large...
+      PDFIO_DEBUG("_pdfioTokenGet(tb=%p, buffer=%p, bufsize=%u): Token '%s' from stack too large.\n", tb, buffer, (unsigned)bufsize, tb->tokens[tb->num_tokens]);
+      *buffer = '\0';
+      return (false);
+    }
+
+    memcpy(buffer, tb->tokens[tb->num_tokens], len);
+    buffer[len] = '\0';

    PDFIO_DEBUG("_pdfioTokenGet(tb=%p, buffer=%p, bufsize=%u): Popping '%s' from stack.\n", tb, buffer, (unsigned)bufsize, buffer);

@ -484,6 +495,13 @@ _pdfioTokenRead(_pdfio_token_t *tb,	// I - Token buffer/stack
 	  *bufptr++ = (char)ch;
 	  break;
 	}
+	else if (ch == '>')
+	{
+	  // Issue #46: Empty hex string from Microsoft PDF generator; treat as
+	  // empty literal string...
+	  *buffer = '(';
+	  break;
+	}
 	else if (!isspace(ch & 255) && !isxdigit(ch & 255))
 	{
 	  _pdfioFileError(tb->pdf, "Syntax error: '<%c'", ch);
@ -536,7 +554,7 @@ _pdfioTokenRead(_pdfio_token_t *tb,	// I - Token buffer/stack

  *bufptr = '\0';

-  PDFIO_DEBUG("_pdfioTokenRead: Read '%s'.\n", buffer);
+//  PDFIO_DEBUG("_pdfioTokenRead: Read '%s'.\n", buffer);

  return (bufptr > buffer);
 }
@ -573,6 +591,7 @@ get_char(_pdfio_token_t *tb)		// I - Token buffer
    tb->bufptr = tb->buffer;
    tb->bufend = tb->buffer + bytes;

+#if 0
 #ifdef DEBUG
    unsigned char *ptr;			// Pointer into buffer

@ -586,6 +605,7 @@ get_char(_pdfio_token_t *tb)		// I - Token buffer
    }
    PDFIO_DEBUG("'\n");
 #endif // DEBUG
+#endif // 0
  }

  // Return the next character...
--- a/pdfio-value.c
+++ b/pdfio-value.c
@ -1,7 +1,7 @@
 //
 // PDF value functions for PDFio.
 //
-// Copyright © 2021 by Michael R Sweet.
+// Copyright © 2021-2023 by Michael R Sweet.
 //
 // Licensed under Apache License v2.0.  See the file "LICENSE" for more
 // information.
@ -219,7 +219,6 @@ _pdfioValueRead(pdfio_file_t   *pdf,	// I - PDF file


  PDFIO_DEBUG("_pdfioValueRead(pdf=%p, obj=%p, v=%p)\n", pdf, obj, v);
-  (void)obj; // TODO: Implement decryption

  if (!_pdfioTokenGet(tb, token, sizeof(token)))
    return (NULL);
@ -284,6 +283,7 @@ _pdfioValueRead(pdfio_file_t   *pdf,	// I - PDF file
 	}
      }
    }
+
    if (token[i])
    {
      // Just a string...
@ -367,6 +367,33 @@ _pdfioValueRead(pdfio_file_t   *pdf,	// I - PDF file

      *dataptr++ = (unsigned char)d;
    }
+
+    if (obj && pdf->encryption)
+    {
+      // Decrypt the string...
+      _pdfio_crypto_ctx_t ctx;		// Decryption context
+      _pdfio_crypto_cb_t cb;		// Decryption callback
+      size_t	ivlen;			// Number of initialization vector bytes
+      uint8_t	temp[32768];		// Temporary buffer for decryption
+      size_t	templen;		// Number of actual data bytes
+
+      if (v->value.binary.datalen > (sizeof(temp) - 32))
+      {
+	_pdfioFileError(pdf, "Unable to read encrypted binary string - too long.");
+	return (false);
+      }
+
+      cb      = _pdfioCryptoMakeReader(pdf, obj, &ctx, v->value.binary.data, &ivlen);
+      templen = (cb)(&ctx, temp, v->value.binary.data + ivlen, v->value.binary.datalen - ivlen);
+
+      // Copy the decrypted string back to the value and adjust the length...
+      memcpy(v->value.binary.data, temp, templen);
+
+      if (pdf->encryption >= PDFIO_ENCRYPTION_AES_128)
+        v->value.binary.datalen = templen - temp[templen - 1];
+      else
+	v->value.binary.datalen = templen;
+    }
  }
  else if (strchr("0123456789-+.", token[0]) != NULL)
  {
--- a/pdfio.h
+++ b/pdfio.h
@ -1,7 +1,7 @@
 //
 // Public header file for PDFio.
 //
-// Copyright © 2021 by Michael R Sweet.
+// Copyright © 2021-2023 by Michael R Sweet.
 //
 // Licensed under Apache License v2.0.  See the file "LICENSE" for more
 // information.
@ -9,27 +9,23 @@

 #ifndef PDFIO_H
 #  define PDFIO_H
-
-//
-// Include necessary headers...
-//
-
 #  include <stdio.h>
 #  include <stdlib.h>
 #  include <stdbool.h>
 #  include <sys/types.h>
 #  include <time.h>
-
-
-//
-// C++ magic...
-//
-
 #  ifdef __cplusplus
 extern "C" {
 #  endif // __cplusplus


+//
+// Version number...
+//
+
+#  define PDFIO_VERSION		"1.1.3"
+
+
 //
 // Visibility and other annotations...
 //
@ -55,6 +51,8 @@ typedef struct _pdfio_array_s pdfio_array_t;
 					// Array of PDF values
 typedef struct _pdfio_dict_s pdfio_dict_t;
 					// Key/value dictionary
+typedef bool (*pdfio_dict_cb_t)(pdfio_dict_t *dict, const char *key, void *cb_data);
+					// Dictionary iterator callback
 typedef struct _pdfio_file_s pdfio_file_t;
 					// PDF file
 typedef bool (*pdfio_error_cb_t)(pdfio_file_t *pdf, const char *message, void *data);
@ -165,6 +163,7 @@ extern pdfio_obj_t	*pdfioDictGetObj(pdfio_dict_t *dict, const char *key) _PDFIO_
 extern pdfio_rect_t	*pdfioDictGetRect(pdfio_dict_t *dict, const char *key, pdfio_rect_t *rect) _PDFIO_PUBLIC;
 extern const char	*pdfioDictGetString(pdfio_dict_t *dict, const char *key) _PDFIO_PUBLIC;
 extern pdfio_valtype_t	pdfioDictGetType(pdfio_dict_t *dict, const char *key) _PDFIO_PUBLIC;
+extern void		pdfioDictIterateKeys(pdfio_dict_t *dict, pdfio_dict_cb_t cb, void *cb_data) _PDFIO_PUBLIC;
 extern bool		pdfioDictSetArray(pdfio_dict_t *dict, const char *key, pdfio_array_t *value) _PDFIO_PUBLIC;
 extern bool		pdfioDictSetBinary(pdfio_dict_t *dict, const char *key, const unsigned char *value, size_t valuelen) _PDFIO_PUBLIC;
 extern bool		pdfioDictSetBoolean(pdfio_dict_t *dict, const char *key, bool value) _PDFIO_PUBLIC;
@ -185,6 +184,7 @@ extern pdfio_obj_t	*pdfioFileCreateObj(pdfio_file_t *pdf, pdfio_dict_t *dict) _P
 extern pdfio_file_t	*pdfioFileCreateOutput(pdfio_output_cb_t output_cb, void *output_ctx, const char *version, pdfio_rect_t *media_box, pdfio_rect_t *crop_box, pdfio_error_cb_t error_cb, void *error_data) _PDFIO_PUBLIC;
 // TODO: Add number, array, string, etc. versions of pdfioFileCreateObject?
 extern pdfio_stream_t	*pdfioFileCreatePage(pdfio_file_t *pdf, pdfio_dict_t *dict) _PDFIO_PUBLIC;
+extern pdfio_file_t	*pdfioFileCreateTemporary(char *buffer, size_t bufsize, const char *version, pdfio_rect_t *media_box, pdfio_rect_t *crop_box, pdfio_error_cb_t error_cb, void *error_data) _PDFIO_PUBLIC;
 extern pdfio_obj_t	*pdfioFileFindObj(pdfio_file_t *pdf, size_t number) _PDFIO_PUBLIC;
 extern const char	*pdfioFileGetAuthor(pdfio_file_t *pdf) _PDFIO_PUBLIC;
 extern time_t		pdfioFileGetCreationDate(pdfio_file_t *pdf) _PDFIO_PUBLIC;
@ -240,10 +240,6 @@ extern char		*pdfioStringCreate(pdfio_file_t *pdf, const char *s)  _PDFIO_PUBLIC
 extern char		*pdfioStringCreatef(pdfio_file_t *pdf, const char *format, ...) _PDFIO_FORMAT(2,3) _PDFIO_PUBLIC;


-//
-// C++ magic...
-//
-
 #  ifdef __cplusplus
 }
 #  endif // __cplusplus
--- a/pdfio.vcxproj
+++ b/pdfio.vcxproj
@ -87,7 +87,7 @@
    <ClCompile>
      <WarningLevel>Level3</WarningLevel>
      <SDLCheck>true</SDLCheck>
-      <PreprocessorDefinitions>PDFIO_VERSION="1.0b2";WIN32;_DEBUG;_CONSOLE;%(PreprocessorDefinitions)</PreprocessorDefinitions>
+      <PreprocessorDefinitions>WIN32;_DEBUG;_CONSOLE;%(PreprocessorDefinitions)</PreprocessorDefinitions>
      <ConformanceMode>true</ConformanceMode>
    </ClCompile>
    <Link>
@ -101,7 +101,7 @@
      <FunctionLevelLinking>true</FunctionLevelLinking>
      <IntrinsicFunctions>true</IntrinsicFunctions>
      <SDLCheck>true</SDLCheck>
-      <PreprocessorDefinitions>PDFIO_VERSION="1.0b2";WIN32;NDEBUG;_CONSOLE;%(PreprocessorDefinitions)</PreprocessorDefinitions>
+      <PreprocessorDefinitions>WIN32;NDEBUG;_CONSOLE;%(PreprocessorDefinitions)</PreprocessorDefinitions>
      <ConformanceMode>true</ConformanceMode>
    </ClCompile>
    <Link>
@ -115,7 +115,7 @@
    <ClCompile>
      <WarningLevel>Level3</WarningLevel>
      <SDLCheck>true</SDLCheck>
-      <PreprocessorDefinitions>PDFIO_VERSION="1.0b2";_DEBUG;_CONSOLE;%(PreprocessorDefinitions)</PreprocessorDefinitions>
+      <PreprocessorDefinitions>_DEBUG;_CONSOLE;%(PreprocessorDefinitions)</PreprocessorDefinitions>
      <ConformanceMode>true</ConformanceMode>
    </ClCompile>
    <Link>
@ -130,7 +130,7 @@
      <FunctionLevelLinking>true</FunctionLevelLinking>
      <IntrinsicFunctions>true</IntrinsicFunctions>
      <SDLCheck>true</SDLCheck>
-      <PreprocessorDefinitions>PDFIO_VERSION="1.0b2";NDEBUG;_CONSOLE;%(PreprocessorDefinitions)</PreprocessorDefinitions>
+      <PreprocessorDefinitions>NDEBUG;_CONSOLE;%(PreprocessorDefinitions)</PreprocessorDefinitions>
      <ConformanceMode>true</ConformanceMode>
    </ClCompile>
    <Link>
--- a/pdfio.xcodeproj/project.pbxproj
+++ b/pdfio.xcodeproj/project.pbxproj
@ -372,7 +372,7 @@
 				CLANG_WARN__DUPLICATE_METHOD_MATCH = YES;
 				CODE_SIGN_IDENTITY = "Apple Development";
 				COPY_PHASE_STRIP = NO;
-				CURRENT_PROJECT_VERSION = 1.0.0;
+				CURRENT_PROJECT_VERSION = 1.1.2;
 				DEBUG_INFORMATION_FORMAT = dwarf;
 				ENABLE_STRICT_OBJC_MSGSEND = YES;
 				ENABLE_TESTABILITY = YES;
@ -381,7 +381,6 @@
 				GCC_PREPROCESSOR_DEFINITIONS = (
 					"$(inherited)",
 					"DEBUG=1",
-					"'PDFIO_VERSION=\"$(CURRENT_PROJECT_VERSION)\"'",
 				);
 				GCC_TREAT_IMPLICIT_FUNCTION_DECLARATIONS_AS_ERRORS = YES;
 				GCC_WARN_64_TO_32_BIT_CONVERSION = YES;
@ -397,7 +396,7 @@
 				GCC_WARN_UNUSED_LABEL = YES;
 				GCC_WARN_UNUSED_PARAMETER = YES;
 				GCC_WARN_UNUSED_VARIABLE = YES;
-				MACOSX_DEPLOYMENT_TARGET = 10.14;
+				MACOSX_DEPLOYMENT_TARGET = 11.0;
 				MTL_ENABLE_DEBUG_INFO = INCLUDE_SOURCE;
 				MTL_FAST_MATH = YES;
 				ONLY_ACTIVE_ARCH = YES;
@ -450,7 +449,7 @@
 				CLANG_WARN__DUPLICATE_METHOD_MATCH = YES;
 				CODE_SIGN_IDENTITY = "Apple Development";
 				COPY_PHASE_STRIP = NO;
-				CURRENT_PROJECT_VERSION = 1.0.0;
+				CURRENT_PROJECT_VERSION = 1.1.2;
 				DEBUG_INFORMATION_FORMAT = "dwarf-with-dsym";
 				ENABLE_HARDENED_RUNTIME = YES;
 				ENABLE_NS_ASSERTIONS = NO;
@ -459,7 +458,6 @@
 				GCC_NO_COMMON_BLOCKS = YES;
 				GCC_PREPROCESSOR_DEFINITIONS = (
 					"$(inherited)",
-					"'PDFIO_VERSION=\"$(CURRENT_PROJECT_VERSION)\"'",
 				);
 				GCC_TREAT_IMPLICIT_FUNCTION_DECLARATIONS_AS_ERRORS = YES;
 				GCC_WARN_64_TO_32_BIT_CONVERSION = YES;
@ -475,7 +473,7 @@
 				GCC_WARN_UNUSED_LABEL = YES;
 				GCC_WARN_UNUSED_PARAMETER = YES;
 				GCC_WARN_UNUSED_VARIABLE = YES;
-				MACOSX_DEPLOYMENT_TARGET = 10.14;
+				MACOSX_DEPLOYMENT_TARGET = 11.0;
 				MTL_ENABLE_DEBUG_INFO = NO;
 				MTL_FAST_MATH = YES;
 				RUN_CLANG_STATIC_ANALYZER = YES;
@ -515,7 +513,7 @@
 				ENABLE_HARDENED_RUNTIME = YES;
 				GCC_DYNAMIC_NO_PIC = NO;
 				GCC_OPTIMIZATION_LEVEL = 0;
-				MACOSX_DEPLOYMENT_TARGET = 10.14;
+				MACOSX_DEPLOYMENT_TARGET = 11.0;
 				PRODUCT_BUNDLE_IDENTIFIER = org.msweet.testpdfio;
 				PRODUCT_NAME = "$(TARGET_NAME)";
 				PROVISIONING_PROFILE_SPECIFIER = "";
@ -529,7 +527,7 @@
 				CODE_SIGN_STYLE = Automatic;
 				DEVELOPMENT_TEAM = "";
 				ENABLE_HARDENED_RUNTIME = YES;
-				MACOSX_DEPLOYMENT_TARGET = 10.14;
+				MACOSX_DEPLOYMENT_TARGET = 11.0;
 				PRODUCT_BUNDLE_IDENTIFIER = org.msweet.testpdfio;
 				PRODUCT_NAME = "$(TARGET_NAME)";
 				PROVISIONING_PROFILE_SPECIFIER = "";
--- a/pdfio1.def
+++ b/pdfio1.def
@ -101,6 +101,7 @@ pdfioContentPathClose
 pdfioContentPathCurve
 pdfioContentPathCurve13
 pdfioContentPathCurve23
+pdfioContentPathEnd
 pdfioContentPathLineTo
 pdfioContentPathMoveTo
 pdfioContentPathRect
@ -154,6 +155,7 @@ pdfioDictGetObj
 pdfioDictGetRect
 pdfioDictGetString
 pdfioDictGetType
+pdfioDictIterateKeys
 pdfioDictSetArray
 pdfioDictSetBinary
 pdfioDictSetBoolean
@ -177,6 +179,7 @@ pdfioFileCreateImageObjFromFile
 pdfioFileCreateObj
 pdfioFileCreateOutput
 pdfioFileCreatePage
+pdfioFileCreateTemporary
 pdfioFileFindObj
 pdfioFileGetAuthor
 pdfioFileGetCreationDate
@ -219,6 +222,8 @@ pdfioPageCopy
 pdfioPageDictAddColorSpace
 pdfioPageDictAddFont
 pdfioPageDictAddImage
+pdfioPageGetNumStreams
+pdfioPageOpenStream
 pdfioStreamClose
 pdfioStreamConsume
 pdfioStreamGetToken
--- a/pdfio_native.nuspec
+++ b/pdfio_native.nuspec
@ -3,7 +3,7 @@
    <metadata>
        <id>pdfio_native</id>
        <title>PDFio Library for VS2019+</title>
-        <version>1.0.0-b7</version>
+        <version>1.1.3</version>
        <authors>Michael R Sweet</authors>
        <owners>michaelrsweet</owners>
        <projectUrl>https://github.com/michaelrsweet/pappl</projectUrl>
@ -12,11 +12,11 @@
        <readme>build/native/README.md</readme>
        <requireLicenseAcceptance>false</requireLicenseAcceptance>
        <description>PDFio Library for VS2019+</description>
-        <summary>PDFio is a simple C library for reading and writing PDF files. PDFio is licensed under the Apache License Version 2.0 with an exception to allow linking against GNU GPL2-only software.</summary>
-        <copyright>Copyright © 2019-2021 by Michael R Sweet</copyright>
+        <summary>PDFio is a simple C library for reading and writing PDF files. PDFio is licensed under the Apache License Version 2.0 with an (optional) exception to allow linking against GNU GPL2-only software.</summary>
+        <copyright>Copyright © 2019-2023 by Michael R Sweet</copyright>
        <tags>pdf file native</tags>
 	<dependencies>
-	    <dependency id="pdfio_native.redist" version="1.0.0-b7" />
+	    <dependency id="pdfio_native.redist" version="1.1.3" />
 	    <dependency id="zlib_native.redist" version="1.2.11" />
 	</dependencies>
    </metadata>
--- a/pdfio_native.redist.nuspec
+++ b/pdfio_native.redist.nuspec
@ -3,7 +3,7 @@
    <metadata>
        <id>pdfio_native.redist</id>
        <title>PDFio Library for VS2019+</title>
-        <version>1.0.0-b7</version>
+        <version>1.1.3</version>
        <authors>Michael R Sweet</authors>
        <owners>michaelrsweet</owners>
        <projectUrl>https://github.com/michaelrsweet/pappl</projectUrl>
@ -12,9 +12,12 @@
        <readme>build/native/README.md</readme>
        <requireLicenseAcceptance>false</requireLicenseAcceptance>
        <description>PDFio Library for VS2019+</description>
-        <summary>PDFio is a simple C library for reading and writing PDF files. This package provides the redistributable content for the PDFio library. PDFio is licensed under the Apache License Version 2.0 with an exception to allow linking against GNU GPL2-only software.</summary>
-        <copyright>Copyright © 2019-2021 by Michael R Sweet</copyright>
+        <summary>PDFio is a simple C library for reading and writing PDF files. This package provides the redistributable content for the PDFio library. PDFio is licensed under the Apache License Version 2.0 with an (optional) exception to allow linking against GNU GPL2-only software.</summary>
+        <copyright>Copyright © 2019-2023 by Michael R Sweet</copyright>
        <tags>pdf file native</tags>
+	<dependencies>
+	    <dependency id="zlib_native.redist" version="1.2.11" />
+	</dependencies>
    </metadata>
    <files>
        <file src="doc\pdfio-128.png" target="build\native" />
--- a/pdfiototext.c
+++ b/pdfiototext.c
@ -0,0 +1,95 @@
+//
+// PDF to text program for PDFio.
+//
+// Copyright © 2022 by Michael R Sweet.
+//
+// Licensed under Apache License v2.0.  See the file "LICENSE" for more
+// information.
+//
+// Usage:
+//
+//   ./pdfiototext FILENAME.pdf > FILENAME.txt
+//
+
+#include "pdfio.h"
+#include <string.h>
+
+
+//
+// 'main()' - Main entry.
+//
+
+int					// O - Exit status
+main(int  argc,				// I - Number of command-line arguments
+     char *argv[])			// I - Command-line arguments
+{
+  pdfio_file_t	*file;			// PDF file
+  size_t	i, j,			// Looping vars
+		num_pages,		// Number of pages
+		num_streams;		// Number of streams for page
+  pdfio_obj_t	*obj;			// Current page object
+  pdfio_stream_t *st;			// Current page content stream
+  char		buffer[1024];		// String buffer
+  bool		first;			// First string token?
+
+
+  // Verify command-line arguments...
+  if (argc != 2)
+  {
+    puts("Usage: pdfiototext FILENAME.pdf > FILENAME.txt");
+    return (1);
+  }
+
+  // Open the PDF file...
+  if ((file = pdfioFileOpen(argv[1], NULL, NULL, NULL, NULL)) == NULL)
+    return (1);
+
+//  printf("%s: %u pages\n", argv[1], (unsigned)pdfioFileGetNumPages(file));
+
+  // Try grabbing content from all of the pages...
+  for (i = 0, num_pages = pdfioFileGetNumPages(file); i < num_pages; i ++)
+  {
+    if ((obj = pdfioFileGetPage(file, i)) == NULL)
+      continue;
+
+    num_streams = pdfioPageGetNumStreams(obj);
+
+//    printf("%s: page%u=%p, num_streams=%u\n", argv[1], (unsigned)i, obj, (unsigned)num_streams);
+
+    for (j = 0; j < num_streams; j ++)
+    {
+      if ((st = pdfioPageOpenStream(obj, j, true)) == NULL)
+	continue;
+
+//      printf("%s: page%u st%u=%p\n", argv[1], (unsigned)i, (unsigned)j, st);
+
+      first = true;
+      while (pdfioStreamGetToken(st, buffer, sizeof(buffer)))
+      {
+	if (buffer[0] == '(')
+	{
+	  if (first)
+	    first = false;
+	  else
+	    putchar(' ');
+
+	  fputs(buffer + 1, stdout);
+	}
+	else if (!strcmp(buffer, "Td") || !strcmp(buffer, "TD") || !strcmp(buffer, "T*") || !strcmp(buffer, "\'") || !strcmp(buffer, "\""))
+	{
+	  putchar('\n');
+	  first = true;
+	}
+      }
+
+      if (!first)
+	putchar('\n');
+
+      pdfioStreamClose(st);
+    }
+  }
+
+  pdfioFileClose(file);
+
+  return (0);
+}
--- a/testpdfio.c
+++ b/testpdfio.c
@ -34,6 +34,7 @@ static int	do_test_file(const char *filename, int objnum, bool verbose);
 static int	do_unit_tests(void);
 static int	draw_image(pdfio_stream_t *st, const char *name, double x, double y, double w, double h, const char *label);
 static bool	error_cb(pdfio_file_t *pdf, const char *message, bool *error);
+static bool	iterate_cb(pdfio_dict_t *dict, const char *key, void *cb_data);
 static ssize_t	output_cb(int *fd, const void *buffer, size_t bytes);
 static const char *password_cb(void *data, const char *filename);
 static int	read_unit_file(const char *filename, size_t num_pages, size_t first_image, bool is_output);
@ -405,7 +406,7 @@ do_test_file(const char *filename,	// I - PDF filename

      filter = pdfioDictGetName(dict, "Filter");

-      if ((st = pdfioObjOpenStream(obj, (filter && !strcmp(filter, "FlateDecode")) ? PDFIO_FILTER_FLATE : PDFIO_FILTER_NONE)) == NULL)
+      if ((st = pdfioObjOpenStream(obj, filter && !strcmp(filter, "FlateDecode"))) == NULL)
      {
        _pdfioValueDebug(&obj->value, stdout);
 	putchar('\n');
@ -454,7 +455,7 @@ do_test_file(const char *filename,	// I - PDF filename
 	      }
 	    }

-	    printf("    Page #%d is %gx%g.\n", (int)n + 1, media_box.x2, media_box.y2);
+	    printf("    Page #%d (obj %d) is %gx%g.\n", (int)n + 1, (int)pdfioObjGetNumber(obj), media_box.x2, media_box.y2);
 	  }
 	}

@ -506,6 +507,9 @@ do_unit_tests(void)
  _pdfio_value_t	value;		// Value
  size_t		first_image,	// First image object
 			num_pages;	// Number of pages written
+  char			temppdf[1024];	// Temporary PDF file
+  pdfio_dict_t		*dict;		// Test dictionary
+  int			count = 0;	// Number of key/value pairs
  static const char	*complex_dict =	// Complex dictionary value
    "<</Annots 5457 0 R/Contents 5469 0 R/CropBox[0 0 595.4 842]/Group 725 0 R"
    "/MediaBox[0 0 595.4 842]/Parent 23513 0 R/Resources<</ColorSpace<<"
@ -967,6 +971,41 @@ do_unit_tests(void)

  // TODO: Test for known values in this test file.

+  // Test dictionary APIs
+  fputs("pdfioDictCreate: ", stdout);
+  if ((dict = pdfioDictCreate(inpdf)) != NULL)
+  {
+    puts("PASS");
+
+    fputs("pdfioDictSet*: ", stdout);
+    if (pdfioDictSetBoolean(dict, "Boolean", true) && pdfioDictSetName(dict, "Name", "Name") && pdfioDictSetNumber(dict, "Number", 42.0) && pdfioDictSetString(dict, "String", "String"))
+    {
+      puts("PASS");
+    }
+    else
+    {
+      puts("FAIL");
+      return (1);
+    }
+
+    fputs("pdfioDictIterateKeys: ", stdout);
+    pdfioDictIterateKeys(dict, iterate_cb, &count);
+    if (count == 4)
+    {
+      puts("PASS");
+    }
+    else
+    {
+      printf("FAIL (got %d, expected 4)\n", count);
+      return (1);
+    }
+  }
+  else
+  {
+    puts("FAIL");
+    return (1);
+  }
+
  // Test the value parsers for edge cases...
  fputs("_pdfioValueRead(complex_dict): ", stdout);
  s = complex_dict;
@ -1107,6 +1146,18 @@ do_unit_tests(void)
  if (read_unit_file("testpdfio-aesp.pdf", num_pages, first_image, false))
    return (1);

+  fputs("pdfioFileCreateTemporary: ", stdout);
+  if ((outpdf = pdfioFileCreateTemporary(temppdf, sizeof(temppdf), NULL, NULL, NULL, (pdfio_error_cb_t)error_cb, &error)) != NULL)
+    printf("PASS (%s)\n", temppdf);
+  else
+    return (1);
+
+  if (write_unit_file(inpdf, outpdf, &num_pages, &first_image))
+    return (1);
+
+  if (read_unit_file(temppdf, num_pages, first_image, false))
+    return (1);
+
  pdfioFileClose(inpdf);

  return (0);
@ -1199,6 +1250,27 @@ error_cb(pdfio_file_t *pdf,		// I  - PDF file
 }


+//
+// 'iterate_cb()' - Test pdfioDictIterateKeys function.
+//
+
+static bool				// O - `true` to continue, `false` to stop
+iterate_cb(pdfio_dict_t *dict,		// I - Dictionary
+           const char   *key,		// I - Key
+           void         *cb_data)	// I - Callback data
+{
+  int	*count = (int *)cb_data;	// Pointer to counter
+
+
+  if (!dict || !key || !cb_data)
+    return (false);
+
+  (*count)++;
+
+  return (true);
+}
+
+
 //
 // 'output_cb()' - Write output to a file.
 //
@ -1238,6 +1310,7 @@ read_unit_file(const char *filename,	// I - File to read
 {
  pdfio_file_t	*pdf;			// PDF file
  size_t	i;			// Looping var
+  const char	*s;			// String
  bool		error = false;		// Error callback data


@ -1248,6 +1321,87 @@ read_unit_file(const char *filename,	// I - File to read
  else
    return (1);

+  // Verify metadata...
+  fputs("pdfioFileGetAuthor: ", stdout);
+  if ((s = pdfioFileGetAuthor(pdf)) != NULL && !strcmp(s, "Michael R Sweet"))
+  {
+    puts("PASS");
+  }
+  else if (s)
+  {
+    printf("FAIL (got '%s', expected 'Michael R Sweet')\n", s);
+    return (1);
+  }
+  else
+  {
+    puts("FAIL (got NULL, expected 'Michael R Sweet')");
+    return (1);
+  }
+
+  fputs("pdfioFileGetCreator: ", stdout);
+  if ((s = pdfioFileGetCreator(pdf)) != NULL && !strcmp(s, "testpdfio"))
+  {
+    puts("PASS");
+  }
+  else if (s)
+  {
+    printf("FAIL (got '%s', expected 'testpdfio')\n", s);
+    return (1);
+  }
+  else
+  {
+    puts("FAIL (got NULL, expected 'testpdfio')");
+    return (1);
+  }
+
+  fputs("pdfioFileGetKeywords: ", stdout);
+  if ((s = pdfioFileGetKeywords(pdf)) != NULL && !strcmp(s, "one fish,two fish,red fish,blue fish"))
+  {
+    puts("PASS");
+  }
+  else if (s)
+  {
+    printf("FAIL (got '%s', expected 'one fish,two fish,red fish,blue fish')\n", s);
+    return (1);
+  }
+  else
+  {
+    puts("FAIL (got NULL, expected 'one fish,two fish,red fish,blue fish')");
+    return (1);
+  }
+
+  fputs("pdfioFileGetSubject: ", stdout);
+  if ((s = pdfioFileGetSubject(pdf)) != NULL && !strcmp(s, "Unit test document"))
+  {
+    puts("PASS");
+  }
+  else if (s)
+  {
+    printf("FAIL (got '%s', expected 'Unit test document')\n", s);
+    return (1);
+  }
+  else
+  {
+    puts("FAIL (got NULL, expected 'Unit test document')");
+    return (1);
+  }
+
+  fputs("pdfioFileGetTitle: ", stdout);
+  if ((s = pdfioFileGetTitle(pdf)) != NULL && !strcmp(s, "Test Document"))
+  {
+    puts("PASS");
+  }
+  else if (s)
+  {
+    printf("FAIL (got '%s', expected 'Test Document')\n", s);
+    return (1);
+  }
+  else
+  {
+    puts("FAIL (got NULL, expected 'Test Document')");
+    return (1);
+  }
+
  // Verify the number of pages is the same...
  fputs("pdfioFileGetNumPages: ", stdout);
  if (num_pages == pdfioFileGetNumPages(pdf))
@ -2237,24 +2391,11 @@ write_font_test(pdfio_file_t *pdf,	// I - PDF file
  };


-#if 0
-  if (unicode)
-  {
-    fputs("pdfioFileCreateFontObjFromFile(NotoSansJP-Regular.otf): ", stdout);
-    if ((opensans = pdfioFileCreateFontObjFromFile(pdf, "testfiles/NotoSansJP-Regular.otf", true)) != NULL)
-      puts("PASS");
-    else
-      return (1);
-  }
+  fputs("pdfioFileCreateFontObjFromFile(OpenSans-Regular.ttf): ", stdout);
+  if ((opensans = pdfioFileCreateFontObjFromFile(pdf, "testfiles/OpenSans-Regular.ttf", unicode)) != NULL)
+    puts("PASS");
  else
-#endif // 0
-  {
-    fputs("pdfioFileCreateFontObjFromFile(OpenSans-Regular.ttf): ", stdout);
-    if ((opensans = pdfioFileCreateFontObjFromFile(pdf, "testfiles/OpenSans-Regular.ttf", unicode)) != NULL)
-      puts("PASS");
-    else
-      return (1);
-  }
+    return (1);

  fputs("pdfioDictCreate: ", stdout);
  if ((dict = pdfioDictCreate(pdf)) != NULL)
@ -3167,13 +3308,13 @@ write_unit_file(
  // Create some image objects...
  fputs("pdfioFileCreateImageObjFromFile(\"testfiles/color.jpg\"): ", stdout);
  if ((color_jpg = pdfioFileCreateImageObjFromFile(outpdf, "testfiles/color.jpg", true)) != NULL)
-    puts("PASS");
+    printf("PASS (%u)\n", (unsigned)pdfioObjGetNumber(color_jpg));
  else
    return (1);

  fputs("pdfioFileCreateImageObjFromFile(\"testfiles/gray.jpg\"): ", stdout);
  if ((gray_jpg = pdfioFileCreateImageObjFromFile(outpdf, "testfiles/gray.jpg", true)) != NULL)
-    puts("PASS");
+    printf("PASS (%u)\n", (unsigned)pdfioObjGetNumber(gray_jpg));
  else
    return (1);

--- a/testttf.c
+++ b/testttf.c
@ -0,0 +1,305 @@
+//
+// Unit test program for TTF library
+//
+//     https://github.com/michaelrsweet/ttf
+//
+// Copyright © 2018-2023 by Michael R Sweet.
+//
+// Licensed under Apache License v2.0.  See the file "LICENSE" for more
+// information.
+//
+// Usage:
+//
+//   ./testttf [FILENAME]
+//
+
+#include <stdio.h>
+#include "ttf.h"
+
+
+//
+// Local functions...
+//
+
+static void	error_cb(void *data, const char *message);
+static int	test_font(const char *filename);
+
+
+//
+// 'main()' - Main entry for unit tests.
+//
+
+int					// O - Exit status
+main(int  argc,				// I - Number of command-line arguments
+     char *argv[])			// I - Command-line arguments
+{
+  int		i;			// Looping var
+  int		errors = 0;		// Number of errors
+
+
+  if (argc > 1)
+  {
+    for (i = 1; i < argc; i ++)
+      errors += test_font(argv[i]);
+  }
+  else
+  {
+    // Test with the bundled TrueType files...
+    errors += test_font("testfiles/OpenSans-Bold.ttf");
+    errors += test_font("testfiles/OpenSans-Regular.ttf");
+    errors += test_font("testfiles/NotoSansJP-Regular.otf");
+  }
+
+  if (!errors)
+    puts("\nALL TESTS PASSED");
+  else
+    printf("\n%d TEST(S) FAILED\n", errors);
+
+  return (errors);
+}
+
+
+//
+// 'error_cb()' - Error callback.
+//
+
+static void
+error_cb(void       *data,		// I - User data (not used)
+         const char *message)		// I - Message string
+{
+  fprintf(stderr, "FAIL (%s)\n", message);
+}
+
+
+//
+// 'test_font()' - Test a font file.
+//
+
+static int				// O - Number of errors
+test_font(const char *filename)		// I - Font filename
+{
+  int		i,			// Looping var
+		errors = 0;		// Number of errors
+  ttf_t		*font;			// Font
+  const char	*value;			// Font (string) value
+  int		intvalue;		// Font (integer) value
+  float		realvalue;		// Font (real) value
+  ttf_rect_t	bounds;			// Bounds
+  ttf_rect_t	extents;		// Extents
+  size_t	num_fonts;		// Number of fonts
+  ttf_style_t	style;			// Font style
+  ttf_weight_t	weight;			// Font weight
+  static const char * const stretches[] =
+  {					// Font stretch strings
+    "TTF_STRETCH_NORMAL",		// normal
+    "TTF_STRETCH_ULTRA_CONDENSED",	// ultra-condensed
+    "TTF_STRETCH_EXTRA_CONDENSED",	// extra-condensed
+    "TTF_STRETCH_CONDENSED",		// condensed
+    "TTF_STRETCH_SEMI_CONDENSED",	// semi-condensed
+    "TTF_STRETCH_SEMI_EXPANDED",	// semi-expanded
+    "TTF_STRETCH_EXPANDED",		// expanded
+    "TTF_STRETCH_EXTRA_EXPANDED",	// extra-expanded
+    "TTF_STRETCH_ULTRA_EXPANDED"	// ultra-expanded
+  };
+  static const char * const strings[] =	// Test strings
+  {
+    "Hello, World!",			// English
+    "مرحبا بالعالم!",			// Arabic
+    "Bonjour le monde!",		// French
+    "Γειά σου Κόσμε!",			// Greek
+    "שלום עולם!",			// Hebrew
+    "Привет мир!",			// Russian
+    "こんにちは世界！"			// Japanese
+  };
+  static const char * const styles[] =	// Font style names
+  {
+    "TTF_STYLE_NORMAL",
+    "TTF_STYLE_ITALIC",
+    "TTF_STYLE_OBLIQUE"
+  };
+
+
+  printf("ttfCreate(\"%s\"): ", filename);
+  if ((font = ttfCreate(filename, 0, error_cb, NULL)) != NULL)
+    puts("PASS");
+  else
+    errors ++;
+
+  fputs("ttfGetAscent: ", stdout);
+  if ((intvalue = ttfGetAscent(font)) > 0)
+  {
+    printf("PASS (%d)\n", intvalue);
+  }
+  else
+  {
+    printf("FAIL (%d)\n", intvalue);
+    errors ++;
+  }
+
+  fputs("ttfGetBounds: ", stdout);
+  if (ttfGetBounds(font, &bounds))
+  {
+    printf("PASS (%g %g %g %g)\n", bounds.left, bounds.bottom, bounds.right, bounds.top);
+  }
+  else
+  {
+    puts("FAIL");
+    errors ++;
+  }
+
+  fputs("ttfGetCapHeight: ", stdout);
+  if ((intvalue = ttfGetCapHeight(font)) > 0)
+  {
+    printf("PASS (%d)\n", intvalue);
+  }
+  else
+  {
+    printf("FAIL (%d)\n", intvalue);
+    errors ++;
+  }
+
+  fputs("ttfGetCopyright: ", stdout);
+  if ((value = ttfGetCopyright(font)) != NULL)
+  {
+    printf("PASS (%s)\n", value);
+  }
+  else
+  {
+    puts("FAIL");
+    errors ++;
+  }
+
+  for (i = 0; i < (int)(sizeof(strings) / sizeof(strings[0])); i ++)
+  {
+    printf("ttfGetExtents(\"%s\"): ", strings[i]);
+    if (ttfGetExtents(font, 12.0f, strings[i], &extents))
+    {
+      printf("PASS (%.1f %.1f %.1f %.1f)\n", extents.left, extents.bottom, extents.right, extents.top);
+    }
+    else
+    {
+      puts("FAIL");
+      errors ++;
+    }
+  }
+
+  fputs("ttfGetFamily: ", stdout);
+  if ((value = ttfGetFamily(font)) != NULL)
+  {
+    printf("PASS (%s)\n", value);
+  }
+  else
+  {
+    puts("FAIL");
+    errors ++;
+  }
+
+  fputs("ttfGetItalicAngle: ", stdout);
+  if ((realvalue = ttfGetItalicAngle(font)) >= 0.0)
+  {
+    printf("PASS (%g)\n", realvalue);
+  }
+  else
+  {
+    printf("FAIL (%g)\n", realvalue);
+    errors ++;
+  }
+
+  fputs("ttfGetNumFonts: ", stdout);
+  if ((num_fonts = ttfGetNumFonts(font)) > 0)
+  {
+    printf("PASS (%u)\n", (unsigned)num_fonts);
+  }
+  else
+  {
+    puts("FAIL");
+    errors ++;
+  }
+
+  fputs("ttfGetPostScriptName: ", stdout);
+  if ((value = ttfGetPostScriptName(font)) != NULL)
+  {
+    printf("PASS (%s)\n", value);
+  }
+  else
+  {
+    puts("FAIL");
+    errors ++;
+  }
+
+  fputs("ttfGetStretch: ", stdout);
+  if ((intvalue = (int)ttfGetStretch(font)) >= TTF_STRETCH_NORMAL && intvalue <= TTF_STRETCH_ULTRA_EXPANDED)
+  {
+    printf("PASS (%s)\n", stretches[intvalue]);
+  }
+  else
+  {
+    printf("FAIL (%d)\n", intvalue);
+    errors ++;
+  }
+
+  fputs("ttfGetStyle: ", stdout);
+  if ((style = ttfGetStyle(font)) >= TTF_STYLE_NORMAL && style <= TTF_STYLE_ITALIC)
+  {
+    printf("PASS (%s)\n", styles[style]);
+  }
+  else
+  {
+    puts("FAIL");
+    errors ++;
+  }
+
+  fputs("ttfGetVersion: ", stdout);
+  if ((value = ttfGetVersion(font)) != NULL)
+  {
+    printf("PASS (%s)\n", value);
+  }
+  else
+  {
+    puts("FAIL");
+    errors ++;
+  }
+
+  fputs("ttfGetWeight: ", stdout);
+  if ((weight = ttfGetWeight(font)) >= 0)
+  {
+    printf("PASS (%u)\n", (unsigned)weight);
+  }
+  else
+  {
+    puts("FAIL");
+    errors ++;
+  }
+
+  fputs("ttfGetWidth(' '): ", stdout);
+  if ((intvalue = ttfGetWidth(font, ' ')) > 0)
+  {
+    printf("PASS (%d)\n", intvalue);
+  }
+  else
+  {
+    printf("FAIL (%d)\n", intvalue);
+    errors ++;
+  }
+
+  fputs("ttfGetXHeight: ", stdout);
+  if ((intvalue = ttfGetXHeight(font)) > 0)
+  {
+    printf("PASS (%d)\n", intvalue);
+  }
+  else
+  {
+    printf("FAIL (%d)\n", intvalue);
+    errors ++;
+  }
+
+  fputs("ttfIsFixedPitch: ", stdout);
+  if (ttfIsFixedPitch(font))
+    puts("PASS (true)");
+  else
+    puts("PASS (false)");
+
+  ttfDelete(font);
+
+  return (errors);
+}
--- a/ttf.c
+++ b/ttf.c
@ -3,16 +3,12 @@
 //
 //     https://github.com/michaelrsweet/ttf
 //
-// Copyright © 2018-2021 by Michael R Sweet.
+// Copyright © 2018-2023 by Michael R Sweet.
 //
 // Licensed under Apache License v2.0.  See the file "LICENSE" for more
 // information.
 //

-//
-// Include necessary headers...
-//
-
 #ifdef _WIN32
 #  define _CRT_SECURE_NO_WARNINGS
 #endif // _WIN32
@ -75,7 +71,7 @@ typedef __int64 ssize_t;		// POSIX type not present on Windows...


 //
-// DEBUG is typically defined for debug builds.  TTF_DEBUG maps to printf when
+// DEBUG is typically defined for debug builds.  TTF_DEBUG maps to fprintf when
 // DEBUG is defined and is a no-op otherwise...
 //

@ -420,7 +416,7 @@ ttfCreate(const char   *filename,	// I - Filename
  if (read_os_2(font, &os_2))
  {
    // Copy key values from OS/2 table...
-    static const ttf_stretch_t widths[] =
+    static const ttf_stretch_t stretches[] =
    {
      TTF_STRETCH_ULTRA_CONDENSED,	// ultra-condensed
      TTF_STRETCH_EXTRA_CONDENSED,	// extra-condensed
@ -433,8 +429,8 @@ ttfCreate(const char   *filename,	// I - Filename
      TTF_STRETCH_ULTRA_EXPANDED	// ultra-expanded
    };

-    if (os_2.usWidthClass >= 1 && os_2.usWidthClass <= (int)(sizeof(widths) / sizeof(widths[0])))
-      font->stretch = widths[os_2.usWidthClass - 1];
+    if (os_2.usWidthClass >= 1 && os_2.usWidthClass <= (int)(sizeof(stretches) / sizeof(stretches[0])))
+      font->stretch = stretches[os_2.usWidthClass - 1];

    font->weight     = (short)os_2.usWeightClass;
    font->cap_height = os_2.sCapHeight;
@ -452,7 +448,7 @@ ttfCreate(const char   *filename,	// I - Filename
    font->cap_height = font->ascent;

  if (font->x_height == 0)
-    font->x_height   = 3 * font->ascent / 5;
+    font->x_height = 3 * font->ascent / 5;

  // Build a sparse glyph widths table...
  font->min_char = -1;
@ -481,6 +477,11 @@ ttfCreate(const char   *filename,	// I - Filename
      else
 	font->widths[bin][i & 255] = widths[glyph];
    }
+
+#ifdef DEBUG
+    if (i >= ' ' && i < 127)
+      TTF_DEBUG("ttfCreate: width['%c']=%d(%d)\n", (char)i, font->widths[0][i].width, font->widths[0][i].left_bearing);
+#endif // DEBUG
  }

  // Cleanup and return the font...
@ -737,17 +738,6 @@ ttfGetFamily(ttf_t *font)		// I - Font
 }


-//
-// 'ttfIsFixedPitch()' - Determine whether a font is fixedpitch.
-//
-
-bool					// O - `true` if fixed pitch, `false` otherwise
-ttfIsFixedPitch(ttf_t *font)		// I - Font
-{
-  return (font ? font->is_fixed : false);
-}
-
-
 //
 // 'ttfGetItalicAngle()' - Get the italic angle.
 //
@ -856,13 +846,14 @@ int					// O - Width in 1000ths
 ttfGetWidth(ttf_t *font,		// I - Font
            int   ch)			// I - Unicode character
 {
-  int	bin = ch >> 8;			// Bin in widths array
+  int	bin =  ch >> 8;			// Bin in widths array


  // Range check input...
  if (!font || ch < ' ' || ch == 0x7f)
    return (0);
-  else if (font->widths[bin])
+
+  if (font->widths[bin])
    return ((int)(1000.0f * font->widths[bin][ch & 255].width / font->units));
  else if (font->widths[0])		// .notdef
    return ((int)(1000.0f * font->widths[0][0].width / font->units));
@ -882,6 +873,17 @@ ttfGetXHeight(ttf_t *font)		// I - Font
 }


+//
+// 'ttfIsFixedPitch()' - Determine whether a font is fixedpitch.
+//
+
+bool					// O - `true` if fixed pitch, `false` otherwise
+ttfIsFixedPitch(ttf_t *font)		// I - Font
+{
+  return (font ? font->is_fixed : false);
+}
+
+
 //
 // 'copy_name()' - Copy a name string from a font.
 //
@ -1305,17 +1307,17 @@ read_cmap(ttf_t *font)			// I - Font
              {
                // Use an "obscure indexing trick" (words from the spec, not
                // mine) to look up the glyph index...
-                temp = segment->idRangeOffset / 2 + ch - segment->startCode + seg - segCount - 1;
+                temp = segment->idRangeOffset / 2 + ch - segment->startCode + seg - segCount;

-                if (temp < 0 || temp >= numGlyphIdArray || !glyphIdArray[temp])
+                if (temp < 0 || temp >= numGlyphIdArray)
                  glyph = -1;
 		else
-		  glyph = ((glyphIdArray[temp] + segment->idDelta) & 65535) - 1;
+		  glyph = (glyphIdArray[temp] + segment->idDelta) & 65535;
              }
              else
              {
                // Just use idDelta to compute a glyph index...
-                glyph = ((ch + segment->idDelta) & 65535) - 1;
+                glyph = (ch + segment->idDelta) & 65535;
 	      }

 	      cmapptr[ch] = glyph;
@ -1582,6 +1584,8 @@ read_hmtx(ttf_t           *font,	// I - Font
  {
    widths[i].width        = (short)read_ushort(font);
    widths[i].left_bearing = (short)read_short(font);
+
+    TTF_DEBUG("read_hmtx: widths[%d].width=%d, .left_bearing=%d\n", i, widths[i].width, widths[i].left_bearing);
  }

  return (widths);
--- a/ttf.h
+++ b/ttf.h
@ -3,7 +3,7 @@
 //
 //     https://github.com/michaelrsweet/ttf
 //
-// Copyright © 2018-2021 by Michael R Sweet.
+// Copyright © 2018-2023 by Michael R Sweet.
 //
 // Licensed under Apache License v2.0.  See the file "LICENSE" for more
 // information.
@ -11,17 +11,11 @@

 #ifndef TTF_H
 #  define TTF_H
-
-//
-// Include necessary headers...
-//
-
 #  include <stdbool.h>
 #  include <sys/types.h>
-
 #  ifdef __cplusplus
 extern "C" {
-#  endif //
+#  endif // __cplusplus


 //
@ -89,8 +83,8 @@ extern ttf_t		*ttfCreate(const char *filename, size_t idx, ttf_err_cb_t err_cb,
 extern void		ttfDelete(ttf_t *font);
 extern int		ttfGetAscent(ttf_t *font);
 extern ttf_rect_t	*ttfGetBounds(ttf_t *font, ttf_rect_t *bounds);
-extern int		ttfGetCapHeight(ttf_t *font);
 extern const int	*ttfGetCMap(ttf_t *font, size_t *num_cmap);
+extern int		ttfGetCapHeight(ttf_t *font);
 extern const char	*ttfGetCopyright(ttf_t *font);
 extern int		ttfGetDescent(ttf_t *font);
 extern ttf_rect_t	*ttfGetExtents(ttf_t *font, float size, const char *s, ttf_rect_t *extents);
@ -112,5 +106,4 @@ extern bool		ttfIsFixedPitch(ttf_t *font);
 #  ifdef __cplusplus
 }
 #  endif // __cplusplus
-
 #endif // !TTF_H
Author	SHA1	Message	Date
Michael R Sweet	9fec2195d0	Update changelog.	2023-11-15 10:14:02 -05:00
Michael R Sweet	8ccbdaed94	Update documentation.	2023-11-15 09:28:57 -05:00
Michael R Sweet	4804db38a5	Fix missing ivlen initializer for 40-bit RC4 (Issue #51 )	2023-11-15 08:43:07 -05:00
Michael R Sweet	ddd984215a	Save work (debug printfs, etc.)	2023-11-15 08:38:47 -05:00
Michael R Sweet	efe7c01015	Fix typo and 'make debug'.	2023-11-14 18:38:26 -05:00
Michael R Sweet	600fa4ce59	Fix Unicode font handling (Issue #16 )	2023-11-14 18:19:34 -05:00
Michael R Sweet	688810f143	Save work.	2023-11-13 16:18:02 -05:00
Michael R Sweet	858cc101b6	Save work.	2023-11-13 13:39:06 -05:00
Michael R Sweet	43114f43bf	Bump version.	2023-10-10 10:22:27 -04:00
Michael R Sweet	c4abceef79	Make Visual Studio compiler happy.	2023-10-10 07:24:27 -04:00
Michael R Sweet	2e91e05d7b	Allow "compound" filters that consist of a single named filter (Issue #47 )	2023-10-10 07:14:12 -04:00
Michael R Sweet	7e3db6b639	Merge pull request #48 from crystalidea/master added windows.h header for GetTempPathA	2023-10-10 07:07:46 -04:00
kleuter	acb6b66bdb	added windows.h header for GetTempPathA	2023-10-10 09:12:03 +02:00
Michael R Sweet	b0a66eef78	Fix reading of PDF files from Crystal Reports (Issue #45 )	2023-10-09 10:04:20 -04:00
Michael R Sweet	ed88322496	Debug logging, work in progress for Unicode text support.	2023-10-07 12:05:18 -04:00
Michael R Sweet	59959bf0e5	Merge TTF changes to fix off-by-one error.	2023-10-06 16:44:20 -04:00
Michael R Sweet	19c45871fa	Update pdfioContentSetDashPattern to support setting solid line styles (Issue #41 )	2023-10-06 15:47:27 -04:00
Michael R Sweet	b0e4646f9d	Rework CR/LF skip code to be more consistent.	2023-10-06 14:41:55 -04:00
Michael R Sweet	12ef2fe2c3	Remove LGTM badges.	2023-10-06 14:40:40 -04:00
Michael R Sweet	4630060ee7	Update security reporting and contribution text.	2023-10-06 14:40:28 -04:00
Michael R Sweet	74a6fb1860	Get rid of superfluous comments.	2023-10-06 14:40:08 -04:00
Michael R Sweet	a3ea0a99ff	Cleanup spacing and comments.	2023-10-06 14:39:42 -04:00
Michael R Sweet	fdfa700442	Update ignored files.	2023-10-06 14:39:10 -04:00
Michael R Sweet	d759baf11e	Bump version and put PDFIO_VERSION definition in the pdfio.h header.	2023-10-06 14:38:38 -04:00
Michael R Sweet	7f6ffcda22	Fix a couple issues with parsing PDF files produced by Microsoft Reporting Services (Issue #46) - Odd cross-reference stream containing 3-byte generation number field for this 16-bit value - Odd empty hex strings	2023-10-06 10:46:30 -04:00
Michael R Sweet	87ca4db73f	Clean up private header.	2023-10-02 05:27:40 -04:00
Michael R Sweet	a83f7f50ff	Allow extra whitespace/newlines after stream tokens (Issue #40 )	2023-10-02 05:06:33 -04:00
Michael R Sweet	6a4ce57d09	Bump versions for Mac/Windows project files.	2023-03-20 10:40:25 -04:00
Michael R Sweet	d4c594cec4	Bump copyright in readme.	2023-03-20 10:22:19 -04:00
Michael R Sweet	35c674b633	Fix another build issue.	2023-03-20 10:11:05 -04:00
Michael R Sweet	97d4955666	Fix potential denial-of-service in flate stream code.	2023-03-20 09:27:19 -04:00
Michael R Sweet	e138232a93	Fix build error due to mismatched function declarations.	2023-03-20 08:19:31 -04:00
Michael R Sweet	8d8225f4a1	Fix release date.	2023-02-07 17:10:48 -05:00
Michael R Sweet	7045d9dad9	Bump Windows version numbers and update exports file.	2023-02-06 17:36:54 -05:00
Michael R Sweet	4f10021e7e	Fix denial-of-service attack when reading corrupt PDF files.	2023-02-03 20:39:04 -05:00
Michael R Sweet	57d5894f33	Update pdfioStreamGetToken documentation (Issue #37 )	2023-01-11 17:13:58 -05:00
Michael R Sweet	2b8a1c8481	Fix CodeQL config file syntax.	2022-12-09 11:31:56 -05:00
Michael R Sweet	948ee16b06	Fix the one "value" complaint from CodeQL and suppress all useless queries.	2022-12-09 11:22:10 -05:00
Michael R Sweet	c7101ae9dd	Add CodeQL scanning.	2022-12-09 11:09:34 -05:00
Michael R Sweet	599640eda1	Update makefile to be silent with basic progress reporting.	2022-08-02 09:41:13 -04:00
Michael R Sweet	a3f3bbfe11	Fix pdfioFileGetAuthor, etc. APIs (Issue #33 )	2022-07-12 18:36:08 -04:00
Michael R Sweet	26d485cfc5	Update Windows DLL exports file.	2022-07-06 15:25:45 -04:00
Michael R Sweet	64d306a322	Cleanup.	2022-07-06 08:47:52 -04:00
Michael R Sweet	067683cbcd	Add some protection against opening multiple streams in the same file at the same time.	2022-07-04 13:03:11 -04:00
Michael R Sweet	50f27974cf	Update documentation.	2022-07-03 10:01:20 -04:00
Michael R Sweet	ae9a91719b	Add pdfioContentPathEnd function.	2022-07-03 10:01:10 -04:00
Michael R Sweet	1a17933635	Fix pdfioContentMatrixRotate function.	2022-07-01 20:30:40 -04:00
Michael R Sweet	acea6fdbed	Changelog.	2022-06-27 17:17:44 -04:00
Michael R Sweet	66fa12f928	Update Windows DLL exports file.	2022-06-27 10:17:21 -04:00
Michael R Sweet	f4b8983c61	Implement pdfioDictIterateKeys API (Issue #31 )	2022-06-27 10:17:00 -04:00
Michael R Sweet	ed4e2fc38a	Merge pull request #32 from ire4ever1190/patch-1 Fix `install-shared` Make task	2022-06-09 09:46:23 -04:00
Jake Leahy	1ed7f0089c	Update Makefile	2022-06-09 14:33:53 +10:00
Michael R Sweet	563d53edd4	Update Windows DLL exports file.	2022-05-24 19:16:20 -04:00
Michael R Sweet	316b0ad559	Add pdfioFileCreateTemporary function (Issue #29 )	2022-05-15 22:52:53 -04:00
Michael R Sweet	f8b471acfd	Update README and NOTICE files...	2022-03-02 09:50:14 -05:00
Michael R Sweet	cedd7d104f	Changelog update.	2022-03-02 09:47:14 -05:00
Michael R Sweet	6378047026	Update VC project.	2022-03-02 09:31:33 -05:00
Michael R Sweet	54578144a0	Update documentation and prep for 1.0.1 release.	2022-03-02 09:30:01 -05:00
Michael R Sweet	f7f2969e3a	Fix pdfioStreamGetToken implementation (wasn't flushing input), update pdfiototext code to better handle different text operators that affect the location of the text.	2022-03-01 09:18:56 -05:00
Michael R Sweet	93a3fcea6c	Add missing pdfioPageGetNumStreams and pdfioPageOpenStream functions. Add initial version of pdfiototext text extraction utility.	2022-02-28 15:00:25 -05:00
Michael R Sweet	fa20982e5d	Coverity certs are fixed.	2021-12-15 18:20:54 -05:00
Michael R Sweet	44d20eba1b	Add stub code for AES-256 to force Coverity to re-analyze...	2021-12-15 07:35:55 -05:00
Michael R Sweet	c0b7925cdf	Fix typo.	2021-12-15 07:28:17 -05:00
Michael R Sweet	68dcf021b2	Download Entrust root cert for validation.	2021-12-15 07:25:44 -05:00
Michael R Sweet	b0a8e60968	Also allow posts to coverity.com while we wait for Ubuntu to pick up the new Entrust root certificate.	2021-12-15 07:10:13 -05:00
Michael R Sweet	9d47745e43	Prep for 1.0rc1.	2021-12-15 06:53:09 -05:00
Michael R Sweet	b0bf2e04b9	Coverity's certificate has expired.	2021-12-14 16:26:57 -05:00
Michael R Sweet	f030112372	See what is happening when downloading Coverity build tool (drop quiet option).	2021-12-14 16:21:49 -05:00
Michael R Sweet	79c4b6f8a8	See what is happening when downloading Coverity build tool.	2021-12-14 16:20:34 -05:00
Michael R Sweet	bd2f9d44d4	Prep for 1.0.0 release.	2021-12-14 12:36:33 -05:00