Based on PR 30397 by Cameron Patrick, I modify to adapt to recent pkgsrc
tree and update from 1.85 to 1.86.
GNU Units can convert quantities expressed in one unit to an equivalent in a
difference scale. Compared to traditional BSD implementations, it offers more
units, non-linear conversions (e.g. between degrees Celsius and Fahrenheit), a
more lenient input parser, and tab completion. It can also be used to perform
calculations taking physical units into account, and provide the definitions of
units (e.g. "joule = N m = 1 kg m^2 / s^2").
* 1.10:
+ loading CHM files by topic.
+ now keeping the scroll position between HTML window
resizes.
+ WxWidgets 2.7 friendly code.
+ F9 now works to toggle the contents, just like Firefox.
+ added Swedish translation.
+ added Portuguese translation (in addition to the old - now
pt_BR - translation).
+ fixed the Portuguese translations (they were using the
wrong charset).
+ wxGTK 2.8.x friendly.
+ workaround to be able to print relative path CHM images.
============
Version 0.12
============
o improves category validation code to not catch false positives (Vincent Fretin, Ville Skyttä, Ray Strode, Vincent Untz)
o make category validation code non-fatal (Ray)
o fix mem leaks and double frees (Pascal Terjan)
Version 3.3.8 adds support for full-text search using the FTS1
module. There are also minor bug fixes. Upgrade only if you want
to try out the new full-text search capabilities or if you are
having problems with 3.3.7.
Version 3.3.7 includes support for loadable extensions and virtual
tables. But both features are still considered "beta" and their
APIs are subject to change in a future release. This release is
mostly to make available the minor bug fixes that have accumulated
since 3.3.6. Upgrading is not necessary. Do so only if you encounter
one of the obscure bugs that have been fixed or if you want to try
out the new features.
2006-Jun-6 - Version 3.3.6
Changes include improved tolerance for windows virus scanners and
faster :memory: databases. There are also fixes for several obscure
bugs. Upgrade if you are having problems.
2006-Apr-5 - Version 3.3.5
This release fixes many minor bugs and documentation typos and
provides some minor new features and performance enhancements.
Upgrade only if you are having problems or need one of the new
features.
CHANGES:
0.8.6 - 0.8.7
- Revert the alignment fix, in order to mimick the way WP for Mac renders
the documents (Fridrich)
- Support for WP Mac 1.x documents (Fridrich)
- Fix memory leak in standalone converters (Andrew Ziem)
- Miscellaneous code cleanups (Fridrich)
0.8.5 -> 0.8.6
- Initial support of WP Mac 2.x documents (Fridrich)
- Fix an alignment bug where alignment was propagated from multicolumn
sections into following single column sections (Fridrich)
- Initial conversion of headers/footers in WP42 parser (Fridrich)
- Fix some parser errors in WP42 parser; now we seek to the right place
and parse correctly all functions (Fridrich)
- Fixed crash in wpd2foo if invoked with non-existing filename (Jody Goldberg)
- Assure that the table we send to the listener implementation is a valid
table independent of how messy the table is in the WP document (Fridrich)
- Fix situations where there is no tab position information in WPUs inside
the tabulator function; the fallback works now even if the position is 0
like in WP6 for DOS documents.
- Try not to crash under any circumstance by sanitizing the code and throwing
exception in illogical, but in WP documents very real, situations (sum1,
Fridrich)
- Make every implementation of WPXInputStream that we use return !0 if a seek
fails; be more paranoiac with the result of read functions (sum1, Fridrich)
- Ignore corrupted prefix packets in WP5 documents (Fridrich)
- Try to convert as much information as possible even in corrupted WP3, WP5 and
WP6 documents (Fridrich) - Libwpd now converts many documents that make crash
WordPerfect itself.
- Fine-tune the WP42 document detection. Detect as many WP42 files as possible
while still discarding plain text documents (Fridrich, Marc)
- Let the WP6 parser not die if the prefix packet it refers to is corrupted or
absent (Fridrich)
- Fix hangs while parsing a corrupted WP6ExtendedDocumentSummaryPacket (sum1,
Fridrich)
- General improvement of libwpd's conversion-success-rate and extensive QA with
more than 46'000 real life documents (Mr. SUM1)
0.8.4 -> 0.8.5
- Conversion of font face, size and colour in WP5 parser, including the
default font information (Fridrich)
- Conversion of foot/endnotes in WP5.x format (Fridrich)
- Conversion of headers/footers in WP3 and WP5 parsers (Fridrich)
- Prevent negative paragraph margins due to page margin change.
Removes the ugly text-border lines running across the text in
OpenOffice.org (Fridrich)
- Make page margins constant between two hard page breaks (Fridrich)
- Convert page margin changes into section margins in multi-column sections
- Move absolute position values in multicolumn sections from whatever column
they are in into the first one. Fixes the off-page position of the second
table in 05mechanicalservice.wpd (Fridrich)
- Defer page span change to the end of the current paragraph if it is
opened in order to prevent a paragraph break where it is not there in the
original document (Fridrich)
- Fix http://bugzilla.abisource.com/show_bug.cgi?id=10105, an incorrect
conversion of table alignment in WP3.x file format (Fridrich)
- Fix small issue with incorrect number of pages in page-spans (Fridrich)
- Fix an issue with "==" operator for WPXPageSpan classes. The result is now
the same independent on the order of the operands (Fridrich)
- Fix http://bugzilla.abisource.com/show_bug.cgi?id=10279, a crash if a table
from the middle of the tableList is in footnote and/or endnote (Marc, Fridrich)
- Add an option "--info" to wpd2text; called with this option, wpd2text dumps the
information of the document instead of converting it. This could be useful for
beagle (Fridrich)
- Refactoring of the listener structure and split of WPXListener into WPXContentListener
and WPXStylesListener (Fridrich, Marc and Cyrille Moureaux as a guest star)
(phpldapadmin works fine without). This removes a number of second-order
dependencies for this package. Bump PKGREVISION.
Ok with adrianp (who originally added the dependency).
* 1.4.3: [25 August 2006]
** Fixing 'unicode string filename' bug [Bug #1546546]. Regressive bug:
code to handle unicode strings added in v1.3.3 was lost in v1.4 during
code re-arranging/refactoring
Changes since 1.1.1:
xvidcore library
* Fixed typo in xvid.c
* Updated bitstream.h to fix the previous faulty non-aligned
bitstream buffers fix
Changes since 1.1.0:
xvidcore library
* Fixed bug when frame-drop (N-VOP) feature is used in combination
with packed B-frames
* Fix for premature EOF in xvid_decraw example
* Fixed potential crash on Intel EMT64 architecture
* Several fixes for IA64 platform (patch by Thomas Koeckerbauer)
* Fix for visual_object_verid vs. video_object_layer_verid problem
* Ensure intervening bytes are preserved in BitstreamInit()
VFW frontend
* Prevent segfault when encoding application calls compress_end
with NULL codec context (PerfectDark at yandex dot ru)
* Profile definitions updates
DShow frontend
* Add missing debug.c
=== RELEASE 2.1pre24 ===
Mon Nov 6 02:13:48 cet 2006 mikulas:
Fixed incorrect page displayed when search string ended with space and
the word was last on page
Thu Oct 26 16:10:01 MET 2006 user:
More strict parsing of ftp list, fixes ftp.su.se
Sun Oct 15 22:55:28 MET DST 2006 mikulas:
Allow cookie path to point to file, not only directory
Wed Oct 11 00:17:27 MET DST 2006 mikulas:
Fixed quirk with displaying of links in text mode
Mon Oct 9 23:33:03 MET 2006 Jakub Bogusz <qboosh@pld-linux.org>:
Updated Polish translation
Mon Oct 9 21:55:24 MET 2006 Jindrich Makovicka <makovick@gmail.com>:
Fixed memory corruption when redirecting URLs with user:password
Mon Sep 11 02:43:32 MET DST 2006 mikulas:
Test for socklen_t
Work without PF_INET (for Minix 3)
Cleaned up use of AF_ and PF_ constants
Sat Sep 9 00:05:16 MET 2006 user:
Fixed double-to-string conversion for config files
Fri Aug 25 00:48:05 MET 2006 user:
Parsing ftp directories starts to be difficult because fixing one bug
introduces other. Previous fix sometimes mistook size for year :-/
Sun Aug 20 05:06:57 cet 2006 mikulas:
Correctly translate character set of "label" attribute
Thu Aug 17 19:57:34 MET 2006 user:
Fixed incremental loading of XBM images
Mon Aug 7 05:11:47 CEST 2006 Matthew Fischer <futhark@mchsi.com>:
Fixed Japanese fonts
Mon Jul 31 00:32:58 MET 2006 user:
More cygwin fixes
bash 3.1.x gets shell quoting wrong, causing configure to fail on platforms
that use this shell
New in this version according to
http://www.modpython.org/live/mod_python-3.2.10/doc-html/node98.html:
- Added support for Apache 2.2.
- New req.is_https() and req.ssl_var_lookup() methods. These communicate
direct with the Apache mod_ssl module, allowing it to be determined if
the connection is using SSL/TLS and what the values of internal ssl
variables are.
- New req.server.get_options() method. This returns the subset of Python
options set at global scope within the Apache configuration. That is,
outside of the context of any VirtualHost, Location, Directory or Files
directives.
- The directory used for mutex locks can now be specified at at compile
time using ./configure --with-mutex-dir value or at run time with
PythonOption mod_python.mutex_directory value.
- The number of mutex locks can now be specified at run time with
PythonOption mod_python.mutex_locks value.
- Fixed three memory leaks that were found in _apachemodule.parse_qsl,
req.readlines and util.cfgtree_walk.
- Third party C modules that use the simplified API for the Global Interpreter
Lock (GIL), as described in PEP 311, can now be used. The only requirement
is that such modules can only be used in the context of the "main_interpreter".
- DbmSession unit test no longer uses the default directory for the dbm file,
so the test will not interfer with the user's current apache instance.
- Added additional debugging and logging output for where mod_python cannot
initialise itself properly due to Python or mod_python version mismatches or
missing Python module code files.
- Fixed configure problem when using bash 3.1.x.
- Fixed DbmSession to create db file with mode 0640.
- Fixed request.sendfile() bug for symlinked files on Win32.
version 1.2.13beta1 [October 2, 2006]
Removed AC_FUNC_MALLOC from configure.ac
Work around Intel-Mac compiler bug by setting PNG_NO_MMX_CODE in pngconf.h
Change "logical" to "bitwise" throughout documentation.
Detect and fix attempt to write wrong iCCP profile length.
version 1.0.21, 1.2.13 [November 15, 2006]
Fix potential buffer overflow in sPLT chunk handler.
Fix Makefile.am to not try to link to noexistent files.
Check all exported functions for NULL png_ptr.
* Bug fixes
* Added the --log-file=FILE and --log-file-format=FORMAT options.
* Added the --remove-source-files option as a replacement for the (now
deprecated) --remove-sent-files option.
* Added the option --no-motd to suppress the message-of-the-day output
from a daemon when doing a copy.
* Added a new environment variable to the pre-/post-xfer exec commands (in
the daemon's config file): RSYNC_PID.
Mention docbook and LaTeX in DESCRiption.
Add sourceforge to MASTER_SITES.
Add more directories to INSTALLATION_DIRS (maybe some are redundant).
Also install a2x.
Install man pages.
Install various configs and icons.
Add do-test target.
Following is from the AsciiDoc ChangeLog:
1. Version 8.1.0 (2006-10-22)
* AsciiDoc generated XHTML documents now display as table of
contents if the toc attribute is defined (JavaScript needs to be
enabled for this to work). Thanks to Troy Hanson who contributed
this feature based on a JavaScript by Mihai Bazon. I've simplified
things somewhat to match Docbook XSL Stylesheets style, see Troy's
[1]tpl User Guide for a fancier layout. Use the -a toc -a numbered
command-line options to produce a number table of contents.
* A [2]music filter is included in the distribution ./filters/
directory. It translates music in [3]LilyPond or [4]ABC notation
to standard classical notation in the form of a trimmed PNG image
which is inserted into the AsciiDoc output document.
* Incorporated Paul Melis's Win32 filter patch. This workaround
allows AsciiDoc to run filters under Windows.
* Added uninstall.sh script.
* Rather than proliferate a confusing number of filter block
delimiters the following convention has been adopted: delimiters
belonging to DelimitedBlock filters distributed with AsciiDoc will
consist of a word (normally a noun identifying the block content)
followed by four or more tilde characters. This has necessitated
changing existing filter delimiters (the old delimiters still work
but may be deprecated in future versions):
+ The example code filter block delimiter is now the word code
followed by four or more tilde characters.
+ The source highlight filter block delimiter is now the word
source followed by four or more tilde characters.
* Conditionally redefined subscript and superscripting so they use
the old replacements mechanism when asciidoc7compatible is defined
rather than the asciidoc 8 default unconstrained quoting (patch
for affected files attached).
* Moved the source highlight filter from ./examples/ to ./filter/.
* Added {verbose} intrinsic attribute (useful for passing verbose
flag to filters).
* Added {outdir} intrinsic attribute.
* Renamed {docdir} intrinsic attribute to unambiguous`{indir}
({docdir}` still works but may be removed in future release).
* If asciidoc(1) outputs to stdout then instrinsic attribute
{docname} is extracted from the input file name.
_________________________________________________________________
2. Version 8.0.0 (2006-08-27)
This is a major release because changes to quoting and index entry
handling may break existing documents (see Additions and changes below
and Appendix A: Migration Notes in the AsciiDoc User Guide).
Please report any problems you encounter.
[5]Stuart Rackham
Additions and changes
* Quoting can can occur within words (based on patch submitted by
Benjamin Klum). See the Unconstrained Quotes sub-section in the
User Guide.
* The underline and plus characters can be used as alternatives to
the existing apostrophe and backtick quote characters. They are
arguably better choices than the apostrophe and backtick as they
are not confused with punctuation.
* The syntax for index entry macros have have been deprecated from
+...+ and +...+ to ((...)) and (((...))) respectively. Rationale:
+ Bracketing is consistent other with [[...]] and <<...>>
reference macros.
+ To easily confused with triple plus passthroughs.
+ To make way for the new monospace quoting.
* Superscripts and subscripts are implemented as constrained quotes
so they can now be escaped with a leading backslash and prefixed
with with an attribute list.
* An experimental LaTeX backend has been written by Benjamin Klum (a
number additions in this release are to accommodate the LaTeX
backend).
* include macro file names now expand environment variables and
tilde expansions.
* A configuration file [quotes] entry can be undefined by setting to
a blank value.
* Added callto inline macro for Skype callto links.
* Added colnumber attribute for table data markup.
* A leading comment block or comment lines are now skipped
(previously a document had to start with either attribute entries
or a document Title).
* Experimental rows attribute (number of source lines in table)
available in table markup templates (used by experimental LaTeX
backend).
* Included install shell script written by [6]Jacob Mandelson for
installing the tarball distribution.
* Added INSTALL documentation file.
* Added replacements2 substitution options -- a second replacements
section.
* Added the ability to redefine normal and verbatim substitutions
with subsnormal and subsverbatim entries in configuration file
[miscellaneous] section.
* By default AttributeEntry values are substituted for
specialcharacters and attributes, if you want a different
AttributeEntry substitution set the attributeentry-subs attribute.
* The name in name=value configuration file entries can now end with
a backslash, just escape the trailing backslash with a backslash.
For example:
abc\\=xyz
Results in name=abc\ and value=xyz -- previously this would have
escaped the = character.
* A blank configuration file section deletes any preceding section
with the same name (applies to non-markup template sections).
* A command-line attribute value with a @ suffix does not override
existing document and configuration file attributes (normally
command-line attributes have precedence over document and
configuration file attributes).
* localtime attribute is now encoded from the native system encoding
to the output encoding. Patch submitted by [7]FKtPp -- here's his
description of the problem:
"I am a Chinese user of AsciiDoc and I find that when I use UTF-8
(the default encoding) to write asciidoc documents in Windows
platform the resulting html footer line will get screwed. It was
caused by a localized tzname that was always encoded in the
windows native encoding, which in my case is cp936."
* a2x(1) can generate Open Document Text files using [8]docbook2odf.
Currently docbook2odf(1) only processes a subset of DocBook,
unimplemented elements are skipped.
* The a2x(1) format option defaults to xhtml (previously a format
had to be specified explicitly).
* The -d, --doctype=DOCTYPE option has been added to a2x(1) which is
a shortcut for --asciidoc-options="--doctype=DOCTYPE".
* Replaced a2x(1) --no-icons and --no-copy options with their
negated equivalents: --icons and --copy respectively. The default
behavior has also changed: copying and use of icons is disabled by
default. Rationale:
+ To make the default behavior more consistent since use of
icons and CSS stylesheets does not apply to all formats.
+ To make the default behavior less surprising (the creation of
icon and stylesheet output files must now be explicit).
* a2x(1) has been bumped from version 0.1.1 to version 1.0.0.
Bug fixes
* Removed duplicate ./doc/a2x.1.txt from distribution tarball.
* Documentation errata.
* Attribute replacement is no longer performed twice in Titles and
AttributeEntrys.
* a2x(1) skipped asciidoc(1) execution when rerun with different
--asciidoc-options options, it now always executes asciidoc(1).
The problem was that previously asciidoc(1) was executed only if
the output file was missing or older than the source file.
'c' isn't in USE_LANGUAGES by default, and this cause the configure
script to fail (I still don't know why it works for me).
Pointed out by Martti Kuparinen in private mail, and Robert Elz in
PR pkg/35064.
