Problems found locating distfiles:
Package acroread7: missing distfile AdobeReader_enu-7.0.9-1.i386.tar.gz
Package acroread8: missing distfile AdobeReader_enu-8.1.7-1.sparc.tar.gz
Package cups-filters: missing distfile cups-filters-1.1.0.tar.xz
Package dvidvi: missing distfile dvidvi-1.0.tar.gz
Package lgrind: missing distfile lgrind.tar.bz2
Otherwise, existing SHA1 digests verified and found to be the same on
the machine holding the existing distfiles (morden). All existing
SHA1 digests retained for now as an audit trail.
Do it for all packages that
* mention perl, or
* have a directory name starting with p5-*, or
* depend on a package starting with p5-
like last time, for 5.18, where this didn't lead to complaints.
Let me know if you have any this time.
a) refer 'perl' in their Makefile, or
b) have a directory name of p5-*, or
c) have any dependency on any p5-* package
Like last time, where this caused no complaints.
They already weren't bl3'd in, so this is to ensure they aren't
accidentally picked up (as happens on Mac OS X).
It remains to optionize them or (better) make split packages.
From Edgar Fuß on tech-pkg@.
Bug-fixes:
* Host font retrieval on Windows was not thread-safe under certain conditions.
* Non-ASCII characters in environment variables could incorrectly trigger
the error message "Invalid UTF-8 sequence..." on Windows.
* PDFlib could crash if the PDFLIBLOGFILE environment variable was specified
and the "logging" parameter was set.
* Implemented a workaround for broken Corel Craw TIFF images with invalid
values for ExtraSamples.
* PDF_create_annotation() in PDF/A mode restricted the "rotate" and "zoom"
options for all annotation types although PDF/A restricts these only for
annotations with type=Text.
* Building the PDFlib Lite binding for Perl on Windows with the old-style
Makefile instead of Makefile.PL didn't produce any usable binary.
* Textflow could wrongly place text with verticalalignment=justify and
lastlinedist=descender if multiple fonts were used on a line and certain
rare geometric conditions were met.
* Invalid PDF output was generated when PDF_begin_template_ext() was called
in page scope and one or more of the following options were used:
"metadata", "opi-1.3", "opi-2.0".
* PHP binding: added a binary for use with Zend Community Server 5.3 which
requires Visual Studio 8 (as opposed to the common PHP build which uses
VS9).
* The pcos_get_string() method in .NET didn't convert non-ASCII return
strings to UTF-16 correctly.
* Implements a parser for legacy Type 1 PFA fonts with a binary data
section.
* Added support for addressing a "temporary work path" via the option
"clippingpathname" and the pseudo path name "Work Path".
* The "transparencygroup" option is no longer allowed in PDF/A mode to
match the forthcoming Technical Corrigendum 2 for ISO 19005-1.
* Textflow ignored space characters in front of mandatory line breaks
(nextline, nextparagraph, nextpage) if alignment=center or =right.
* Invalid RGB TIFF images with four components, but extrasamples=0 were
wrongly processed in passthrough mode which resulted in invalid image data
in the PDF output.
* Fixed a memory violation with malformed XMP where the namespace prefix
was missing from element names.
* The message "Floating point parameter 'x' has bad value (not a number)"
could occur spuriously under the following conditions: Linux x86, PHP,
certain PHP functions called before using PDFlib API functions. The
problem was caused by a bug in the gcc optimizer and solved by disabling
the optimizer for a particular source code module.
* Fixed a potential memory problem in the XMP code.
* The PHP wrapper now honors the "open_basedir" setting in the php.ini file,
see also www.securityfocus.com/bid/36951/info.
* The starter_type3font samples did not set the bounding box for the
glyphs correctly. As a result some undesired clipping could occur
when viewing the generated PDF with Acrobat 9.2 (but not older versions).
This problem affected only the sample code, not the PDFlib implementation.
* textformat=utf16 (i.e. without specifying the byte ordering) didn't work
correctly in Textflow if the text began with a BOM which indicated
non-native byte ordering.
* Large File Support was accidentally disabled in the AIX 64-bit builds.
This is a security update.
changes:
-fix integer overflow which can be used for DOS or system access
(SA35180)
-update bundled libpng to 1.2.36 to fix potential security issues
This changes the buildlink3.mk files to use an include guard for the
recursive include. The use of BUILDLINK_DEPTH, BUILDLINK_DEPENDS,
BUILDLINK_PACKAGES and BUILDLINK_ORDER is handled by a single new
variable BUILDLINK_TREE. Each buildlink3.mk file adds a pair of
enter/exit marker, which can be used to reconstruct the tree and
to determine first level includes. Avoiding := for large variables
(BUILDLINK_ORDER) speeds up parse time as += has linear complexity.
The include guard reduces system time by avoiding reading files over and
over again. For complex packages this reduces both %user and %sys time to
half of the former time.
This is a newer version of what we have as pkgsrc/print/pdflib, but
it comes with much unfriendlier licensing conditions (non-commercial
use only).
I've tested it with gnuplot. There is just some "dash patterns" support
added over pdflib4 afaics. So making it the default choice isn't worth
the hassle, but if you want to try it, here is it.
