* channels/chan_sip.c: Copy the From header into a variable so that
pedantic SIP handling does not try to mess with a NULL pointer.
(AST-2008-008)
* channels/chan_iax2.c: When we receive a full frame that is
supposed to contain our call number, ensure that it has the
correct one. (closes issue #10078) (AST-2008-006)
soname version bump.
2008-03-08 -- 0.8.3
* Fixed: No more compile errors from GCC 4.3 (#1815520)
Thanks to Cristian Rodriguez for the patch!
* Added: Version guards for Autoconf/Automake
Thanks to Martin Michlmayr for reporting!
* Fixed: Compile error on Windows caused by out-of-date
file lists
* Fixed: Bundled Win32 uriparser updated to 0.6.4
* Soname: 1:3:0
2008-04-27 -- 0.7.1
* Fixed: Bogus syntax error when parsing URIs with port-like
passwords, e.g. "http://user:21@host/" (#1948038)
Thanks to Friedrich Delgado Friedrichs for reporting!
* Fixed: Parser did not handle trailing slashes correctly in some cases,
which also made the structures produced from parsing "http://e.com/"
and "http://e.com" indistinguishable. (#1950126)
Thanks to Edward Z. Yang for reporting!
2008-04-04 -- 0.7.0
* Added: Dissection and composition of query strings
* Added: Documentation improvements
(in|out|inout indicators, addition of \since and \see)
* Changed: Code::Blocks project files updated from file format
version 1.4 to 1.6, which is produced by Code::Blocks 8.02
* Added: Code::Blocks workspace file
* Changed: Soname set to 1:5:0
2008-02-25 -- 0.6.4
* Added: Syntax-based normalization can now handle relative URIs,
e.g. "../../a/b/.././c" is normalized to "../../a/c"
* Fixed: Normalization code could free foreign memory
* Fixed: Normalization processed the path segment even when asked not to
* Added: MinGW Makefile and related readme
Thanks to Michael Anthony Puls II!
* Fixed: Documentation bug not requiring enough memory for the output
buffer when converting a relative file URI back to a filename
* Changed: Soname set to 1:4:0
2008-02-11 -- 0.6.3
* Fixed: Two major crash bugs in normalization code
Thanks to Adrian Manrique for the patch!
* Added: Brief usage tutorial
* Changed: Soname set to 1:3:0
2008-02-08 -- 0.6.2
* Fixed: Freeing a normalized URI like "http://test?"
caused a crash. Thanks to Adrian Manrique for reporting!
* Fixed: Filename <--> URI string conversion helpers can
now handle relative URIs and filenames
* Changed: Soname set to 1:2:0
2007-12-23 -- 0.6.1
* Fixed: Percent-encodings in hostnames were not repaired during
normalization.
Thanks to Adrian Manrique for reporting!
* Fixed: Percent-encodings were fixed after dot removal not before during
normalization.
Thanks to Adrian Manrique for reporting!
* Fixed: Include path order bug
Thanks to Ed Schouten for reporting this!
* Fixed: Shadow builds now possible
Thanks to Adeodato Simó for the patch!
* Added: Version guards for Autoconf/Automake
Thanks to Martin Michlmayr for reporting!
* Changed: Soname set to 1:1:0
some sensible default values for the server configuration file via
the SUBST framework. Use REPLACE_PYTHON. New variable CODEVILLE_DATADIR,
which defaults to ${VARBASE}/db/cdv.
New in version 0.8.0:
* Remember commit message and autofill it on next attempt if commit
fails
* fixed: catch requests for non-existent diffs
* fixed: error handling cleanup in networking code
* fixed: Windows-specific networking fixes
* fixed: binary file corruption on Windows with 'construct'
* fixed: glob expansion with 'describe -d'
* fixed: renaming corner case involving directories, renaming and
deleting
* fixed: files renamed out of a deleted directory on another branch were
marked as a name conflict
* fixed: use case-insensitive regexps for post-commit hooks because the
config parser lowercases everything
* fixed: filesystem timestamp precision problems
* fixed: Python 2.5 fixes
These fixes have been collecting for a while. Now you all get to enjoy
them too.
The version number has been changed to reflect that the currect code is
stable, yet is still missing some features which would be necessary for a
1.0 release
Three pkgsrc patches integrated, plus changes since 1.1:
----- November 2007 - Version 1.6 released -----
1. Check for gpg2 first, and prefer it if available.
----- November 2007 - Version 1.5 released -----
3. Add a --- Status --- line for the beginning and end of the PGP
stuff which includes a timestamp to help avoid a socially engineered
message that includes the "good signature" output in the same format
as what we print.
2. Make sure that all --- Status --- messages are 76 characters long.
1. When displaying MIME messages after verification, substitute a space
for the MIME'ified =20 character at the end of a line, and substitute
an = sign for the MIME'ified =3D. This greatly improves readability
of the verified message.
----- July - October 2007 - Version 1.4 developed, not released -----
NB: Support for older versions of pgp, and PGP from PGP Inc. has
been dropped in this version. GnuPG is now a requirement. This is
in line with the requirements for other PGP addons such as enigmail
for Thunderbird.
Benjamin R. Haskell, pine-pgp ~ benizi.com was instrumental in helping
to suggest methods for and test the results of changes 1-5 in this
version. Thanks!
6. Fix error handling in ppf_mime_decrypt, and make use of the $cfile
variable instead of guessing which file has the message.
5. Look for gpg in PREFIX/bin, /usr/bin, and /bin in that order.
4. Move "build" of the scripts to an all target in the Makefile.
3. Change to using csplit since command line options for both
GNU and BSD versions are compatible, unlike split.
2. Improve portability for users of GNU make and mktemp.
1. Add error handling for failed TDIR creation in ppf_decrypt,
ppf_verify, and ppf_mime*.
----- October 2006 - Version 1.3 released -----
6. Include more mail headers in the grep'ed output of the MIME scripts,
and try to strip more MIME headers from the message.
5. Use local for function variables.
4. Remove commented out code for really old versions of pgp.
3. Do a better job of finding the crucial bits of a MIME message, and
fail more gracefully if we do not.
2. Fix bug that prevented ppf_mime* from working with messages sent by
Evolution, KMail, and any other MIME client that interprets RFC 3156
the same way. (Thanks to Bjoern A. Zeeb <bzeeb@zabbadoz.not> for the tip!)
1. Update e-mail and web addresses.
----- April 2005 - Version 1.2 released -----
7. Update copyright for this year, and use my full given name.
6. For sign and encrypt, scan the headers to determine the right user.
This requires adding _INCLUDEALLHDRS_ to the filter line.
5. Add a message to tell users why their password is needed for ppf_decrypt.
4. Create symlinks in PREFIX/bin for the ppf_mime* scripts.
3. Add User-Agent to the list of headers to display in mime filters.
2. Add a first pass at a ppf_mime_decrypt script. More work is
needed here, see BUGS.
1. In some past version of Pine, the display of filter output changed,
so adjust the line length of some of the messages to accommodate.
from <URL:http://cr.yp.to/distributors.html>:
What are the distribution terms for dot-forward?
2008.06.01: I hereby place the dot-forward package (in particular,
dot-forward-0.71.tar.gz, with MD5 checksum
1fefd9760e4706491fb31c7511d69bed) into the public domain. The
package is no longer copyrighted.
What are the distribution terms for ezmlm?
2008.06.01: I hereby place the ezmlm package (in particular,
ezmlm-0.53.tar.gz, with MD5 checksum 108c632caaa8cdbfd3041e6c449191b2)
into the public domain. The package is no longer copyrighted.
What are the distribution terms for fastforward?
2008.06.01: I hereby place the fastforward package (in particular,
fastforward-0.51.tar.gz, with MD5 checksum
6dc619180ba9726380dc1047e45a1d8d) into the public domain. The
package is no longer copyrighted.
Set DJB_RESTRICTED=no in these packages (and in ezmlm-idx, by
extension). Add commented-out LICENSE=public-domain. As usual,
pkgsrc will strive to keep modifications to a tasteful minimum.
- For direct build failures, make the package location in the first
column a hyper link to the last (supposedly failing) phase.
- For indirect failures, list the first 10 failing dependencies.
readline emulation, the GNU readline library which this module depends
on would not be picked up, causing "perl -d -e '1;'" to fail to start
the perl debugger because it could not find the rl_initialize function.
Instead, point includes and libs into ${PREFIX}, so that we have a chance
of picking up the proper GNU readline library.
Provides an adapter layer between core routines for converting to and from
UTF8 and other encodings. In essence, a way to give multiple existing
Unicode modules a single common interface so you don't have to know
the underlaying implementations to do simple UTF8 to-from other character
set encoding conversions. As such, it wraps the Unicode::String,
Unicode::Map8, Unicode::Map and Jcode modules in a standardized and
simple API.
This also provides general character set conversion operation based
on UTF8 - it is possible to convert between any two compatible and
supported character sets via a simple two step chaining of conversions.
This module converts strings from and to 2-byte Unicode UCS2 format. All
mappings happen via 2 byte UTF16 encodings, not via 1 byte UTF8 encoding.
To transform these use Unicode::String.
This module can't deal directly with utf8. Use Unicode::String to convert
utf8 to utf16 and vice versa.
Character mapping is according to the data of binary mapfiles in
Unicode::Map hierarchy. Binary mapfiles can also be created with this module,
enabling you to install own specific character sets.
