- patch-aa is now implemented (as below), that seems the only Change for
this version.
(upstream)
- Update to 0.603 (PKGVERSION is 0.6.3)
-------------------------------------
0.603 2014/09/07
- mark Net::FTP 2.79 as a valid base too, based on patch from
kimmo[AT]suominen[DOT]com, https://github.com/noxxi/p5-net-inet6glue/pull/3
- Add ${PERL5_LICENSE}
(upstream)
- Update 0.05 to 0.07
---------------------
0.07
General:
- Added unit tests
- Changed to proper source structure
Net/Gnats.pm
- Massive amount of refactoring around PBP.
- Removed unecessary 'dead' code
- Changed documentation according to PBP changes
- Began removing old 'data caching' mechanism, metadata needs to be
captured on a per db basis
- fixed/regressed bugs 4, 5, 7, 8, 9, 11 and created unit tests.
Net/Gnats/PR.pm
- reconstructed parse routine
- some PBP refactoring
Net/Gnats/Response.pm
- new module to hold all socket response (raw, code) and eventually
will replace old procedural method of capturing sets of responses.
0.06
Gnats.pm
- Added 'Net::Gnats::updatePR' method so multiple fields can be changed
with one call instead of having to call replaceField multiple times.
- Added 'Net::Gnats::newPR' method.
- Added 'Net::Gnats::filloutPR' method which fills out a new PR with the
field defaults specified in the gnats dbconfig.
- Fixed checking for when gnatsd will return a multiline response.
Previosuly all codes from 300-399 (except 350) where assumed to be
multiline. But the gnatsd spec says only 300-349 will be multiline.
- When PR's are unparsed into a string the MultiText fields are properly
escaped.
- Added ability to turn on debugging with '$Net::Gnats:debugGnatsd = 1'
- Work around a bug in gnatsd where it doesn't properly close file handles,
so gnatsperl closes and reopens the connection after 100 newPRs. I've
submitted a patch to gnats but it's not been incorporated yet.
- Disabled the listClasses method since gnatsd does not support it
anymore.
- Methods listFieldNames, listInitialInputFields, getFieldType,
isValidField, getFieldTypeInfo, getFieldFlags, getFieldDefault, only
query gnatsd once for the information so multiple calls to these methods
are much faster.
- Net::Gnats:lockPR extracts the PR data and returns a Net::Gnats::PR
object.
- Added Net::Gnats::checkPR method. Changed checkNewPR method to call
checkPR.
- Improved checking for "requireChangeReason" checking in replaceField
method.
- submitPR method calls PR::unparse method.
- Split out some code in _doGnatsCmd into new '_readAndPrcessResponse'
method so other methods could use the same code.
PR.pm
- Net::Gnats::PR::new method now requires the Net::Gnats object as the
first argument. This was done to provide better error checking. The
best was to fix existing code is to change any Net::Gnats::PR::new calls
to $g->new(). See documentation.
- The PR number is no longer tracked differently inside PR.pm. It's now
treated like all other fields. Users will not notice this change since
the getNumber method still works like previous versions. However if you
were accessing the $pr->{number} data directly it won't work anymore.
- The setField method takes an optional argument which is the changeReason.
- Currently setField does not require the changeReason to be specified,
but this may change in the next release.
- Major change to methods that parse and unparse PR's. A lot of code was
copied from gnatsweb so gnatsperl now does a much better job of handling
PR data. Also, in the future this will allow us to handle gnatsweb style
attachements.
This module adds an additional target to the Makefile generated by
ExtUtils::MakeMaker. The target, testcover, calls cover, the command-line
script to generate test coverage statistics, to clean up any data from a
previous run. It then runs the tests, as if make test was run, then calls
cover again to generate the coverage statistics.
-------------------
0.72 2015-02-08
- fix update_ref in gitdata (Mike Schilli)
0.71 2015-01-20
- When a query fails, include any additional error messages returned by the API,
in the croak message. (Neil Bowers)
- Make User-Agent header conform to RFCs (Zak Wilcox)
0.70 2014-10-08
add per_page in GET no matter it supports or not
0.69 2014-09-11
pass ua as the args so we can set up proxies for all
-------------------
1.07 Tue Jan 20 19:15:48 CET 2015
- update: copyright notice
- update: Kwalitee
1.06 Wed Apr 9 09:50:52 CEST 2014
- bugfix: https://rt.cpan.org/Ticket/Display.html?id=93409
- bugfix: https://rt.cpan.org/Ticket/Display.html?id=93408
- update: copyright notice
1.05 Wed Nov 14 21:47:44 CET 2012
- bugfix: on length calculation for IPv6::Option
- NEW: IPv6 HopByHop and Destination options extension header and Option
modules
=> contributed by vinsworld.com (Thanks Vince)
- UPDATE: IPv6::Fragment and IPv6::Routing do not export consts at all.
- update: added constants for hop-by-hop, destination and mobility
extension headers, and many other protocols
1.04 Thu Feb 23 20:18:58 CET 2012
- UPDATE: renamed IGMPv4 to IGMP
- NEW: IPv6 Fragment and Routing extension headers
=> contributed by vinsworld.com (Thanks Vince)
- update: copyright notice
-------------------
1.06 Fri Jan 23 07:32:52 CET 2015
- new: truncated attribute
- update: copyright notice
- update: Kwalitee
1.05 Wed Jan 25 22:48:18 CET 2012
- UPDATE: %Next variable population removed. Must be done from
developer's own layer 7 modules.
- bugfix: check when capture length is smaller than content length
- update: copyright notice
Features:
* local-zone: example.com inform makes unbound log a message with
client IP for queries in that zone. Eg. for finding infected hosts.
* patch from Stephane Lapie that adds to the python API, that
exposes struct delegpt, and adds the find_delegation function.
* Updated contrib warmup.cmd/sh to support two modes - load from
pre-defined list of domains or (with filename as argument) load from
user-specified list of domains, and updated contrib
unbound_cache.sh/cmd to support loading/save/reload cache to/from
default path or (with secondary argument) arbitrary path/filename
* patch for remote control over local sockets. Use control-interface:
/path/sock and control-use-cert: no.
* unbound-checkconf -f prints chroot with pidfile path.
* infra-cache-min-rtt patch from Florian Riehm, for expected long
uplink roundtrip times.
Bug Fixes:
* config.guess and config.sub update from libtoolize.
* getauxval test for ppc64 linux compatibility.
* make strip works for unbound-host and unbound-anchor.
* print query name when max target count is exceeded.
* patch from Stuart Henderson that fixes DESTDIR in
unbound-control-setup for installs where config is not in the prefix
location.
* [bugzilla: 634 ] Fix#634: fix fail to start on Linux LTS 3.14.X,
ignores missing IP_MTU_DISCOVER OMIT option.
* Patch to contrib/unbound_munin_ that uses type ABSOLUTE.
Allows munin.conf: [idleserver.example.net]
unbound_munin_hits.graph_period minute
* Fix pyunbound ord call, portable for python 2 and 3.
* Fix unintended use of gcc extension for incomplete enum types,
compile with pedantic c99 compliance.
* Fix pyunbound byte string representation for python3.
* Fix 0x20 capsforid fallback to omit gratuitous NS and additional
section changes.
* Fix validation failure in case upstream forwarder (ISC BIND) does
not have the same trust anchors and decides to insert unsigned NS
record in authority section.
* Fix scrubber with harden-glue turned off to reject NS (and other
not-address) records.
* iana portlist update.
* [bugzilla: 643 ] Fix doc/example.conf.in: unnecessary whitespace.
-------------------
1.14 Tue Dec 9 19:12:25 CET 2014
- BUGFIX: Dump::Online2: do not eat 100% and do not block for timeoutOnNext
seconds when it is not REALLY needed. We had to remove the
setnonblock option and to use pcap_dispatch i/o pcap_next_ex.
This should be the final bugfix...
There was also an infinite loop condition.
- new: Dump::Online2: maxRunTime() option to stop looping on next() forever
when we do not want. Default to loop forever.
1.13 Sun Dec 2 15:44:14 CET 2012
- BUGFIX: Dump::Online2: in non-blocking mode, we must handle our internal
received frame ring buffer. Many packets were missed because of
that.
1.12 Tue Nov 13 20:15:19 CET 2012
- UPDATE: Dump::Online2: now uses IO::Select to avoid eating 100% CPU
- update: mailing list link
1.11 Sun Nov 11 15:23:04 CET 2012
- bugfix: Dump::Online2 pod
- new: Dump::Online2: support for capture only mode
- new: examples for Dump::Online2
1.10 Wed Sep 12 20:15:09 CEST 2012
- bugfix: Online: use SUPER::next() instead of next()
- update: error messages scheme changed
- update: copyright notice
- Deleting patch-aa, no related line found (at all).
(upstream)
- Update 0.16 to 0.17
------------------
2012.11.28 - 0.17 - Sebastien Aperghis-Tramoni (SAPER)
- [BUGFIX] CPAN-RT#43308: Net::Pcap no longer enables immediate
(unsafe) signals delivery.
- [BUGFIX] Fix allocated size in pcap_open(), pcap_createsrcstr()
and pcap_parsesrcstr().
- [API] Now make the C-like API available by default. Added the
pcap_perl_settings() function and UNSAFE_SIGNALS pseudo-bloc.
- [CODE] Replaced all occurrences of safemalloc() with Newx().
- [CODE] Silenced some warnings.
- [DOC] Added a long note about Net::Pcap and signals delivery.
- [DOC] Improved README.
- [DOC] Mention Net::Pcap::Easy.
- [TESTS] Fixed t/17-lib_version.t to handle two digits versions.
- [TESTS] CPAN-RT#44448: Fixed t/17-lib_version.t to handle
versions from the CVS, thanks to Craig Davison.
- [TESTS] Refactored t/01-api.t and added checks for the new API.
New in 0.9.5:
* The data tables and line breaking algorithm have been updated to Unicode
version 7.0.0.
* In the include file uniname.h, the function unicode_name_character
has been extended to look for name aliases.
Noteworthy changes in version 2.0.27 (2015-02-18)
-------------------------------------------------
* gpg: Detect faulty use of --verify on detached signatures.
* gpg: New import option "keep-ownertrust".
* gpg: Uses SHA-256 for all signature types also on RSA keys.
* gpg: Added support for algo names when generating keys using the
--command-fd method.
* gpg: Unless --allow-weak-digest-algos is used the insecure MD5
based fingerprints are shown as all zeroe
* gpg: Fixed DoS based on bogus and overlong key packets.
* gpg: Better error reporting for keyserver problems.
* Fixed several bugs related to bogus keyrings and improved some
other code.
Version 2.3.0 (feb.18th 2015)
New features:
* Grinder can now find the CA[] property (ChAracter set, e.g. "CA[big5]") in
the sgf and decode the file appropriate. However, some decodings are not
supported by Java, e.g. HZ / CA[hz-gb-2312] / ~{NbGeT49mJV~}
* When reading a file with no "CA[]" property set, we use the default of the
jvm to read the sgf (this was and is the default behaviour). You can set
a different character set as preferred default in the settings window.
The selection of a default decoding (settings) helps decoding files, if
MyGoGrinder is used in portable mode on different OSs: Linux and Mac
usually use UTF-8, while Windows uses the local decoding/encoding.
* If comments of sgf files are not displayed correctly, you can set the
decoding manually: you can switch the decoding ("View" menu -> "Change
text decoding") and you can also set this for all files of the current
folder - without subdirectories (Grinder will remember this).
* The settings file (mygrinder.ini) is now saved and read always with UTF-8
encoding, so that it doesn't matter which default decoding the machine
uses.
* Log files are now also always written with UTF-8 encoding, so that
messages with Russian, Chinese, Hebrew and Inuktitut characters can be
read side by side with Latin (ASCII) characters. The old logs and
settings are backed up.
* When importing tags from a file, you can select the decoding.
* You can let Grinder automatically move defect and unusable files to the
folders "defect" resp. "notsupported"; activate this manually in the
settings file (mygrinder.ini)! This will deactivate editing of defect
files.
* Changed display order of directories in the selection window: Java's file
lists were not sorted, when got from some file system: FAT32 (thumb
drive), ext4 (Linux). NTFS file systems gave an ordered list, so
possibly you didn't ever observe this. You can switch sorted display
off, if you liked the old unsorted list.
* If you want to use the wgf editor, you need to activate it's main window's
menue entry manually in the settings file: WGFFRAME.ENABLED=true
I took the check box out of the settings window.
Tweaks, bug fixes:
* I reduced the count of warning messages, if starting a new profile.
* "\" and "]" are now also handled correctly in the comment area of the main
window
* If there is an error in the sgf code or the file was not found (moved or
deleted), instead of "Exit" we use now a default code, so that the
program can continue normally.
* Refresh the directories view in the selection window without closing the
program
* While working on the code, I remembered a behaviour in Linux: if you make
a soft link to an external problems collection in your problems folder,
these problems are used as if they were inside your problems folder. So
stats files are created (but defect files cannot be moved). This doesn't
work in Windows!
* Sometimes, when deleting a tag in the selection window, the tags list was
suddenly empty (shock). This was just a refresh-display-problem -
seemingly solved
* I threw the MYGG_DEBUG... system variables out.
* ... and some more small tweaks and fixes, which I don't remember
Enable odbc, mysql and pgsql support unconditionally (no dependency cost).
Changes in 15.02:
- Add Elixir support, allows to write plugins in Elixir (not enabled in
this pkgsrc package)
- New command to reload configuration without restart
- Support old style erlang expressions in YAML configuration
- Improved captcha listener parsing when protocol not specified
- Fix upgrade of old unbinarized pubsub table from 2.1
- Minor updates in the documentation
- Other bugfixes
These flags are only used for creating static GHCi libraries
(HS*.o). Setting them to ${LDFLAGS} does more harm than good because
our ${LDFLAGS} contains -Wl,* flags. It's true that
../../mk/wrapper/cmd-sink-ld transforms them but those flags will also
be baked into the compiler (see ${WRKSRC}/compiler/ghc.mk) so they
cause problems when used outside the buildlink.
