* Switched license to GPLv3.
* Fixed bug when using the --p12-charset without --armor.
* The command --gen-key may now be used instead of the
gpgsm-gencert.sh script.
* Changed key generation to reveal less information about the
machine. Bug fixes for gpg2's card key generation.
For Release 0.66
----------------
- Adaptations to Libgda and Libgnomedb's API changes
- Better handling of transaction status when closing a workspace
- Translations:
- New 'sl' translation (Matic Zgur)
- ca (Josep Puigdemont i Casamaj<C3><B3>)
- dz (Pema Geyleg)
- en_GB (David Lodge)
- vi (Clytie Siddall)
installed. Idea from one of joerg@'s patches.
Fix 'make readme' and make README.html' targets for individual packages.
Issue reported by Gary Thorpe and David Lord in pkgsrc-users@
libgnomedb 3.0.0, 2007-04-23
- fix libgnomedb versioning in gtk-docs (yselkowitz)
- misc. bug fixes (Armin Burgmeier, Vivien Malerba)
- Bug fixes: #422979, #423405 (Vivien Malerba)
- Translations:
- ca (Gil Forcada)
- dz (Pema Geyleg)
- lt (Gintautas Miliauskas)
- ar (Djihed Afifi)
libgnomedb 2.99.6, 2007-03-23
- Corrected linker errors when using C++ (Murray Cumming)
- Spell corrections (Murray Cumming)
- Code cleanups (Murray Cumming)
- Data entry widgets improvements (Vivien Malerba)
- Removed the C# bindings (Murray Cumming)
- Always install the doc (Thomas Klausner)
- Lots of minor corrections (Armin Burgmeier, Vivien Malerba, Murray Cumming)
- Bug fixes: #419063, #390609 (Yselkowitz, Murray Cumming)
- Translations:
- new 'dz' translation (Pema Geyleg)
- ca (Josep Puigdemont i Casamajó)
- en_GB (David Lodge)
- sv (Daniel Nylander)
- vi (Clytie Siddall)
libgnomedb 2.99.5, 2007-02-12
- Updated for libgda 2.99.5.
libgnomedb 2.99.4, 2007-02-12
- The library is no split into separate libgnomedb, libgnomedb-extra, and libgnomedb-graph
shared libraries. You must adjust your #include lines appropriately.
Packagers might want to ensure that previously-installed headers are removed, if possible.
- GnomeDbRawForm:
- Handle the new "reset" signal from the GdaDataProxy.
- Don't show a data entry if gnome_db_data_widget_column_hide() has been called.
(Vivien Malerba)
- GnomeDbBasicForm: Allow the "paramlist" property to be set more than once.
- GnomeDbDataWidget, GnomeDbRawGrid, GnomeDBRawForm: added an "iter_changed" signal.
(Vivien Malerba)
- GnomeDbRawGrid:
- Improved column title for parameters restricted by a data model.
- Hide expander for parameters restricted by a data model.
(Vivien Malerba)
- GnomeDbDataCellRendererCombo: Added a "show_expander" property.
(Vivien Malerba)
- Use enum types instead of guint or gulong for flags, matching the
conventions used in GTK+, and making the API clearer: GdaQuery, GdaObjectRef
(Murray Cumming, Openismus)
- Use specific types for property and signal parameter registrations, and use specific signal marshallers.
This should be more robust, and is needed by language bindings.
(Murray Cumming, Openismus)
- Generate enum GType registrations for libgnomedb-extra.
This is needed for language bindings and use of these enums in GValues.
- Improvements to _new() functions for language bindings, such as C++:
GnomeDbSqlConsole, GnomeDbDbmsUpdateViewer, GnomeDbErrorDialog, GnomeDbEditor,
GnomeDbDataImport, GnomeDbLoginDialog, GnomeDbDsnSpec, GnomeDbGrid,
GnomeDbDataSourceSelector, GnomeDbConnectionProperties, GnomeDBForm.
(ArminBurgmeir, Openismus)
GnomeDbDataCellRendererBoolean, GnomeDbDataCellRendererCombo,
GnomeDbDataCellRendererInfo, GnomeDbDataCellRendererTextual,
GnomeDbDataEntryCombo
(Johannes Schmid, Openismus)
- Adapt to latest glade3 API changes. (Vincent Geddes, bug #400836)
- Correct all remaining installation location conflicts with libgda 1.2.
(Leonardo Boshell)
- Correct English spellings. (Kraii)
libgnomedb 2.99.3, 2007-01-27
- Build:
- Now builds with GTK+ 2.8 as well as GTK+ 2.10.
(Murray Cumming)
- Now builds with glib 2.8 as well as glib 2.12.
(Vivien Malerba)
- Now builds without gconf, libgnomecanvas, or libgnomeui.
- The graph API is now a separate shared library.
(Vivien Malerba)
- Use glade-3.1/3.2 instead of glade 3.0,
if using glade at all.
(Vivien Malerba)
- Install gnome-db-plugin.h
(Vivien Malerba)
- Do not install private headers.
(Murray Cumming)
- Removed unused utility functions.
(Bug #383533, Min Sik Kim)
- Windows (win32) build improvements.
(Bug #321663, Yselkowitz)
- Do not ignore CFLAGS.
(Bug #354476, Alex Smith)
- Use glib-mkenums to create gnome-db-enum-types.h and
gnome-db-enum-types.c.
(Bug #129035, Armin Burgmeier)
- Fixed compiler warnings.
(Mark Johnson, Bug #366710, Hans de Goede)
- GnomeDbServerOperation.c: Added "opt_header" property.
(Vivien Malerba)
- Renamed gnome_db_basic_form_show_entries_actions() to
gnome_db_basic_form_show_entry_actions().
(Murray Cumming)
- Renamed GnomeDbmsUpdateViewer to GnomeDbDbmsUpdateViewer.
(Bug #389191, Murray Cumming)
- GnomeDbDatabaseLogin improvements.
(Vivien Malerba)
- Picture widgets: Bug fixes.
(Vivien Malerba)
- GnomeDbTransactionStatus: Warn when the transaction is locked.
(Vivien Malerba)
- Added construct properties for language bindings.
(Bug #391514, Johannes Schmid)
- HIG improvements.
(Jeroen Zwartepoorte)
- Corrected some English strings.
(Bug #352123 and bug #315612 from Clytie Siddall)
Based on patch proveided by Makoto Fujiwara in PR 36715.
Pkgsrc Changes:
Added fam option.
from NEWS:
libgda 3.0.1, 2007-05-10
- Bug fixes: #433687, #434464, #432798, #432798
(Armin Burgmeier, Gustavo R. Montesino, Leonardo Boshell, Vivien Malerba)
- Translations:
- ar (Djihed Afifi)
libgda 3.0.0, 2007-04-20
- misc. bug fixes (Armin Burgmeier, Vivien Malerba)
- bug fixes in the SQL delimiter to handle comments, BEGIN and END blocks and PostgreSQL's
dollar quoting (Vivien Malerba)
- more unit tests (Vivien Malerba)
- removed dependecy and test for popt (Baris Cicek)
- fixed SQLite provider's error reporting, and implemented the get_last_insert_id() virtuel
method (Vivien Malerba)
- fixed MySQL's provider error where the same column would be reported several times in
the TABLE_FIELDS schema (Vivien Malerba)
- fixed a misbehaviour in the data model import of XML files (Vivien Malerba)
- docs improvements
- removed ambiguities in the source files: mention LGPL everywhere in the library, and
GPL for the executables (Murray Cumming)
- allow dictionary XML file to specify its own DTD file (Vivien Malerba)
- Bug fixes: #421986, #421986, #420306, #410200, #427605, #407908, #417249 (Carsten Clark, Baris Cicek,
Vivien Malerba
- Translations:
- dz (Pema Geyleg)
- lt (Gintautas Miliauskas)
- ar (Djihed Afifi)
libgda 2.99.6, 2007-03-23
- Merged several libraries into one for efficiency and because of inter-dependencies (yselkowitz,
Gustavo R. Montesino, Christopher Taylor)
- More explicit API function names (Murray Cumming)
- Applications can now check installed DBMS providers (yselkowitz)
- C# bindings removed from the libgda's module (Murray Cumming)
- Added regression tests for SQL parsing and delimiting (Vivien Malerba)
- Reworked SQL delimiter and made the SQL parser report non parsed characters (Vivien Malerba)
- Embedded SQLite is now 3.3.13 (Vivien Malerba)
- Added configure tests to work around changes in the MDB API (yselkowitz)
- Improved WIN32 building for Cygwin and MinGW (yselkowitz, Gustavo R. Montesino, Christopher Taylor)
- Bug fixes: #407064, #411811, #314732, #390609, #353404, #411984, #349548, #412122, #411811
(Leonardo Boshell, yselkowitz, Armin Burgmeier, Daniel Espinosa, Murray Cumming, Christopher Taylor,
Vivien Malerba)
- Lots of minor bug fixes
- Translations:
- new dz translation (Pema Geyleg)
- sv (Daniel Nylander)
- en_GB (David Lodge)
- it (Marco Ciampa)
libgda 2.99.5, 2007-02-12
- Correct a linker error when using the (recently renamed to)
gda_query_field_value_get_is_parameter() function.
I'm doing a quick extra release because I imagine that the effects of using
an implicitly-declared function could be odd.
libgda 2.99.4, 2007-02-12
- GdaQuery: Renamed gda_query_is_modif_query() to gda_query_is_modify_query().
(Murray Cumming)
- GdaQueryFieldValue: Added the is_parameter property
(Daniel Espinosa)
- GdaEntityField: Renamed get_data_type() vfunc to get_dict_type().
(Daniel Espinosa)
- GdaQueryField: Renamed get_data_type() vfunc to get_dict_type().
(Daniel Espinosa)
- GdaObject: Added the option for GdaObject sub-classes to copy the
"name" property to the "id" one if the "id" is not set (used by the GdaParameter object)
(Vivien Malerba)
- GdaDataModelQuery:
- Correctly detect when it's not possible to get information about columns.
- Added a "use_transaction" property to allow the object to start a transaction before
running the modification SQL queries and to end it afterwards (so all the modifications are atomic).
(Vivien Malerba)
- GdaDataModel: Added "reset" signal, emitted when the data model has completely changed.
(Vivien Malerba)
- GdaDataProxy:
- Added "re_init_done" signal, emitted when the proxied data model's number of columns changes.
- Corrected a bug which prevented any modification to be applied to the proxied data model.
- Added gda_data_proxy_get_n_modified_rows().
(Vivien Malerba)
- Improvements to _new() functions for language bindings, such as C++:
GdaParameterList, (Johannes Schmid, Openismus)
- Make type checks more flexible, allowing use of derived GTypes by
language bindings: (Armin Burgmeier, Openismus)
- Use enum types instead of guint or gulong for flags, matching the
conventions used in GTK+, and making the API clearer: GdaQuery, GdaObjectRef
(Murray Cumming, Openismus)
- Use specific types for property and signal parameter registrations, and use specific signal marshallers.
This should be more robust, and is needed by language bindings.
(Murray Cumming, Openismus)
- Correct all remaining installation location conflicts with libgda 1.2.
(Leonardo Boshell)
- Correct the check for the gamin version.
(Leonardo Boshell)
- gda-sharp (mono bindings) updated.
(Daniel Espinosa)
- Added tests. (Vivien Malerb)
- Win32 build improvements. (Brecht Sanders)
- MDB provider: Attempt to fix the build for more recent versions of the MDB API.
(Murray Cumming)
libgda 2.99.3, 2007-01-23
- Build fixes:
- With gamin/fam. (Murray Cumming)
- removed unnecessary libxslt dependency. (Vivien Malerba)
- gda-sharp binding fixes. (Vivien Malerba)
- Fixed many compiler warnings. (Murray Cumming)
- Replace use of deprecated sys_errlist() with strerror().
(Joseph Sacco, Bug #353729)
- Win32 build fixes. (yselkowitz, Bug #314732)
- Darwin (MacOS X) build fixes. (Min Sik Kim, Bug #382933)
- GdaConfig: Handle the GDA_DSN_LIST_IN_MEMORY environment variable.
When defined, initial DSN list will be loaded from the filename specified
in that environment variable, and the DSN list modifications will not be
written to a file. (Vivien Malerba)
- GdaQueryFieldValue: can now render parameters in SQL with the :param_name
syntax, as used by Oracle. (Vivien Malerba)
- GdaBlobOp: New object, for improved blob handling. (Vivien Malerba)
- GdaQuarkList: gda_quark_list_add_from_string():
Prevent a crash when the connection string has a final ; character.
(Andrea Zagli, Bug #317831)
- GdaQuery:
- Allow use of derived types, for language bindings. (Armin Burgmeier)
- Avoid memory corruption. (Chris Shoemaker, Bug #341061)
- GdaQueryFieldField: The field_name property is now readable. (Vivien Malerba)
- GdaServerProvider:
- Added the escape_string() and unescape_string() virtual methods. (Vivien Malerba)
- Removed unimplemented gda_server_provider_create_blob() and
gda_server_provider_fetch_blob_by_id() functions. (Murray Cumming)
- Removed unused "prop" properties. (Murray Cumming)
- Spelling and grammar corrections. (Clytie Siddall, bug #352122)
- Providers:
- FreeTDS: Build fixes. (Leonardo Boshell)
- MySQL:
- Slight speed optimization. (Esteban Sanchez)
- Fixed a bug with FK reporting in the CONSTRAINTS schema report.
Added the "bit" data type as a synonym to "tinyint"
removed GDA_TYPE_BLOB reference as MySQL does not support blobs.
(Vivien Malerba)
- Oracle: Major improvments, including blob handling. (Vivien Malerba)
- Postgres
- Fix failure of current transaction when trying to delete a
nonexisting blob Oid
- Improved GdaBinary handling.
(Vivien Malerba)
- Sqlite: Update Sqlite sources to 3.3.10.
- BDB and ODBC: Correct a cast (Hans de Goede, bug #366694)
- Update configuration template, add documentation for Prelude
generic TCP options.
- Implement modified patch from Pierre Chifflier <chifflier@inl.fr>
to fix the example log path (fix#224).
- Move IDMEF message normalization in the scheduler, rather than
doing it upon reception. This remove some load from the server
and allow Prelude-Manager own IDMEF messages to go through the
normalizer path.
- Implement heartbeat->analyzer normalization.
- Improve IPv4 / IPv6 address normalization.
IPv4 mapped IPv6 addresses are now mapped back to IPv4.
Additionally, the Normalize plugin now provide two additionals option:
ipv6-only: Map any incoming IPv4 address to IPv6.
keep-ipv4-mapped-ipv6: do not map IPv4 mapped IPv6 addresses back to IPv4.
- Make a difference between exceptional report plugin failure (example:
a single message couldn't be processed) and "global" plugin failure
(example: database server is down). We use a different failover for
'exceptional' failure, so that we don't try to reinsert a bogus message
(fix#247).
- Start of a Prelude-Manager manpages (#236).
- Various bug fixes.
- Ability to use regular expressions in plugins.rules to define
monitored sources, this can be very useful when combined to file
globing.
- [SPEEDUP] When the "*" keyword is used, the data is passed to the
upper layer without trying to match anything.
- Fix NULL pointer dereference when a rule reference an existing,
but empty context (fix#226).
- Remove deprecated use of prelude_client_print_setup_error(),
directly handled via prelude_perror().
- Make the log parser more robust.
- Implement an Auto-Refresh system (fix#231). (including code from
Paul Robert Marino <prmarino1@gmail.com>).
- Ability to filter on missing/offline/online/unknown agents. Make more easier
to read each agent status in collapsed mode.
- Fix filter load/save/delete issue with translation.
- New 'My account' tabs, under the Settings section (fix#241).
- New messageid and analyzerid parameters, allowing link to a Prewikka alert
from an external tool (previously required a database query in order to
retrieve the database event id).
- Don't redirect to user listing once an user preference are recorded. Fix
changing of another user language by an user with PERM_USER_MANAGEMENT.
Display target user language rather than current user language.
- Improve the timeline control table layout.
- Fix translation of string possibly using plural.
* channels/chan_iax2.c: Don't create the Asterisk channel until we
are starting the PBX on it. (ASA-2007-018)
* channels/chan_agent.c: (closes issue #5866) Reported by: tyler Do
not force channel format changes when a generator is present. The
generator may have changed the formats itself and changing them
back would cause issues.
* channels/chan_sip.c: (closes issue #10236) Reported by: homesick
Patches: rpid_1.4_75840.patch uploaded by homesick (license 91)
Accept Remote Party ID on guest calls.
* include/asterisk/app.h: We should not use C++ reserved words in
API headers (closes issue #10266)
* channels/chan_sip.c: Backport a fix for a memory leak that was
fixed in trunk in reivision 76221 by rizzo. The memory used for
the localaddr list was not freed during a configuration reload.
* channels/chan_sip.c: (closes issue #10247) Reported by:
fkasumovic Patches: chan_sip.patch uploaded by fkasumovic
(license #101) Drop any peer realm authentication entries when
reloading so multiple entries do not get added to the peer.
* channels/chan_iax2.c: When processing full frames, take sequence
number wraparound into account when deciding whether or not we
need to request retransmissions by sending a VNAK. This code
could cause VNAKs to be sent erroneously in some cases, and to
not be sent in other cases when it should have been. (closes
issue #10237, reported and patched by mihai)
* channels/chan_iax2.c: When traversing the queue of frames for
possible retransmission after receiving a VNAK, handle sequence
number wraparound so that all frames that should be retransmitted
actually do get retransmitted. (issue #10227, reported and
patched by mihai)
* apps/app_voicemail.c: Store prior to copy (closes issue #10193)
* apps/app_queue.c: removed the word 'pissed' from ast_log(...)
Changes
- The help command is now able to show documentation on subcommands
(such as 'attr set').
- The help command now shows a brief abstract of each command,
instead of only listing their names.
- The command `list changed` now outputs the new path of any
renamed item making it easier to copy and paste these paths
for external program usage.
- `automate attributes` has been renamed to `automate get_attributes`,
also a bug has been fixed there so resurrected attributes are now
properly outputted as "new" and not "changed".
New features
- Two new commands to set and drop attributes over automate:
`automate set_attribute` and `automate drop_attribute`
- There is a new function available to the lua hooks,
'server_request_sync(what, address, include, exclude)', which will
initate a netsync connection to the server at "address", with the
given include and exclude patterns, and will sync, push, or pull,
as given in the "what" argument. If called from a monotone instance
which is not acting as a server, this function will do nothing.
- There is a new hook available,
'get_netsync_key(server, include, exclude)', which is called to
determine which key to use for netsync operations. Note that the
server calls this once at startup with the address it is listening
on, "*", and "" as arguments, rather than for each connection.
Other
- Giving the --confdir argument will automatically set the key store
directory to keys/ under that directory, unless --keydir is also
given. This is a bugfix.
- Fixed a regression in 0.35 that resulted in some databases
becoming significantly larger when storing new revisions. Existing
databases with this problem can be fixed by pulling into a fresh
database using 0.36.
- contrib/lua-mode.el, a Lua mode for GNU emacs.
- contrib/monotone-buildbot-notification.lua, a netsync hook to have a
server notify a buildbot when new changes have arrived. Useful for
anyone who uses a buildbot with monotone as source.
- contrib/monotone-cluster-push.lua, a netsync hook script to have
arriving changes be forwarded to other servers automatically. It
uses the new internal lua function 'server_request_sync'.
- contrib/mtn_makepermissions, a simple script to create
read-permissions and write-permissions from files in the directories
read-permissions.d and write-permissions.d, Debian style.
- contrib/Monotone.pm, a first attempt to write a Perl module to
interface with 'monotone automate stdio'.
- contrib/monotone-import.pl has been removed since monotone now has
an internal import command.
Internal
- Commands are now defined as a tree of commands instead of a
plain list, which allows the help system to look up information
of a command at an level in the tree.
- The command class, the automate class and all the associated
macros have been cleaned up.
- All C++ files now depend on base.hh, which includes the few things
that are used virtually everywhere. 'make distcheck' will check for
the presence of base.hh in all source files and will protest if
it's not there. This is explained further in HACKING.
- Update the internal SQLite to version 3.4.0.
- Updated Visual C building system, which now also builds the test
programs. The script visualc/runtests.bat can be used to run the
tests.
- Monotone can now be built successfully with Boost 1.34. Older
versions of monotone would sometimes seem to work depending on
the compiler used, but would have bugs in path normalization.
- Monotone now requires Boost 1.33 or later.
- The Boost filesystem library is no longer required.
- The Boost unit test system is no longer required.
Updated: 14 June 2007
imap-2006j is a maintenance release, consisting primarily of bugfixes to
problems discovered in the release that affected a small number of users.
Updated: 5 June 2007
imap-2006i is a maintenance release, consisting primarily of bugfixes to
problems discovered in the release that affected a small number of users.
imapd now supports the CHILDREN and ESEARCH extensions.
imapd's attempt to return COPYUID/APPENDUID information for a traditional
UNIX (and MMDF) format mailbox when the mailbox is open by another process
has been declared to be a failure and is now revoked. It was subject to a
timing race, loss of which involved an expensive reset of the mailbox's UID
regime. Any imapd COPY or APPEND to a traditional UNIX or MMDF format that
is open by some other process will now no longer return COPYUID/APPEND.
Although this is technically in violation of RFC 4315, there is a loophole
in that document and the timing race/performance problem is worse.
Updated: 4 April 2007
imap-2006h is a maintenance release, consisting primarily of bugfixes to
problems discovered in the release that affected a small number of users.
Updated: 30 March 2007
imap-2006g is a maintenance release, consisting primarily of bugfixes to
problems discovered in the release that affected a small number of users.
Updated: 30 January 2007
imap-2006f is a maintenance release, consisting primarily of bugfixes to
problems discovered in the release that affected a small number of users.
For the benefit of multi-threaded applications, use of strtok() has been
abolished in the c-client library. imapd and ipop3d stuff use it though.
The TOPS-20 and VAX/VMS ports still use strtok() since they don't use UNIX
threads.
This version has been test-built on Linux, Mac OS X, NeXT, Windows XP,
TOPS-20, and VAX/VMS. This will probably be the last test-build on VAX/VMS
since the system I use for that purpose is being shut down. I have no way
to test-build on DOS, legacy Mac OS (OS 9 and earlier), OS/2, or Windows CE;
and the builds on those systems are probably broken.
Updated: 26 January 2007
imap-2006e is a maintenance release, consisting primarily of bugfixes to
problems discovered in the release that affected a small number of users.
Updated: 6 December 2006
imap-2006d is a maintenance release, consisting primarily of bugfixes to
problems discovered in the release that affected a small number of users.
The decomposition mapping, title-case mapping, and character widths tables
have been updated to comply with the Unicode 5.0 standard.
Prototypes for the utf8aux.c functions have been moved to a new utf8aux.h.
The general c-client modules now include c-client.h instead of the individual
files. Use of c-client.h instead of individual include files insulates
against future shuffling of include files.
Updated: 23 October 2006
imap-2006c is a maintenance release, consisting primarily of bugfixes to
problems discovered in the release that affected a small number of users.
By popular request, if a user has a mix (or other dual-use) format INBOX,
it will no longer be listed as \NoInferiors. It's a bad idea to depend
upon this due to the case ambiguity issue, but it's there.
Updated: 26 September 2006
imap-2006b is a maintenance release, consisting entirely of bugfixes to
problems discovered in the release that affected a small number of users.
Updated: 15 September 2006
imap-2006a is a maintenance release, consisting entirely of bugfixes to
problems discovered in the release that affected a small number of users.
If it is necessary to build IPv4-only on one of the ports that has IPv6
preconfigured (ldb, lfd, lmd, lrh, lsu, osx, oxp), this can be done by
using IP6=4. You can't do IP=4 in the build command directly since these
ports set IP themselves; however, now instead of setting IP=6 they now set
IP=$(IP6).
Updated: 30 August 2006
imap-2006 is a major release. Programs written for imap-2004g should
build with this version with minor or no modification. imap-2005 was not
released except as development snapshots.
imap-2006 contains major extensions to its Unicode support. Searching and
sorting are now done with strings canonicalized to titlecase and decomposed
form. Among other things, this means that Latin letters with diacriticals
will now sort with the basic Latin letter, and case-independent searching of
such letters (e.g., German umlauts) now works. Previously, sorting was done
strictly by Unicode codepoint, and case-independence only worked with ASCII.
imapd now supports the UIDPLUS extension for mailboxes in unix, mmdf, mbx, mx,
and mix formats. UID EXPUNGE is fully implemented. Note that UIDPLUS is not
supported in the little-used drivers (mh, mtx, tenex) in which meaningful
APPENDUID/COPYUID data can not be returned. Refer to bugs.txt for more
details.
The new mix format is a dual-use mailbox format designed for performance and
reliability with large mailboxes. mix is documented in file mixfmt.txt.
SSL/TLS certificate validation on UNIX now checks the alternative names in the
certificate if the CN does not match.
The new /tls-sslv23 flag in a mailbox name causes a TLS session to use the
(incorrect) SSLv23 client method instead of the TLSv1 client method. Some
broken servers use the SSLv23 server method, and this flag works around that
problem. WARNING: use of this flag will cause TLS negotiation to fail with
a server which uses the proper TLSv1 server method. Additionally, there are
known security risks in SSLv2; so users should be suspicious if this switch
suddenly becomes necesary.
The silly mailbox flag combination /ssl/tls is now rejected as an invalid
remote specification. Previous versions tried to negotiate TLS over an SSL
session; even if the server permitted such a thing it couldn't work.
The memory management of several drivers has been redesigned to consume less
memory and hopefully be faster.
The private.data member of the MESSAGECACHE (elt) has been replaced with
a union that contains private.spare.data and private.spare.ptr, the latter
being a pointer.
A new FT_RETURNSTRINGSTRUCT flag has been added for mail_fetch_body() and
mail_fetch_text() calls. If this flag is set, *and* if the function returns
NIL, then the requested string data is available on a stringstruct on
stream->private.string. This is a special hack for the IMAP and POP servers
and is subject to incompatible change. The result is a major performance
improvement in the servers with the mbx driver, particularly with large
messages.
- for pkg_create, split-off the package creation from the plist
processing. package creation is disabled for --enable-bootstrap
- change package creation to use libarchive
- add two options (-u and -g) to override file system ownership
- pkg_create now creates files in the tarball with correct owner/group
based on -u/-g or @owner/@group
- add a function to compute the +CONTENTS file in memory
Bump version to 20070802.
Review of a slightly older version by jlam@
