Log message:
Bump PKGREVISION for libXft changes for NetBSD native X support on
NetBSD 6, requested by tron.

Log message:
Try to fix the fallout caused by the fix for PR pkg/47882. Part 3:

Recursively bump package revisions again after the "freetype2" and
"fontconfig" handling was fixed.

Log message:
Bump freetype2 and fontconfig dependencies to current pkgsrc versions,
to address issues with NetBSD-6(and earlier)'s fontconfig not being
new enough for pango.

While doing that, also bump freetype2 dependency to current pkgsrc
version.

Suggested by tron in PR 47882

Log message:
Bump all packages for perl-5.18, that
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.

Log message:
Be explicit and consistent in the visiblity of globalMemTrace.

Log message:
PKGREVISION bumps for net/libpcap update.

Log message:
Update to version 2.35

ns-2.35 Released on Nov 4, 2011

* Provide fixes to the Tmix traffic generator, mainly to the one-way
  TCP implementation, but some changes to the Full-TCP version as well.
  Tmix with one-way TCP now performs comparably to Tmix with Full-TCP (albeit
  with longer running times and higher memory consumption). Includes
  contributions from David Hayes and DongXia Xu at Swinburne and relevant
  updates to the documentation and tests.
* Add Marcello Caleffi's Multi-path Dynamic Address RouTing (M-DART)
  protocol.
* Add Sidney Doria's Datagram Congestion Control Protocol (DCCP)
  implementation.
* Added the SYN_immediate_ack flag so that a DelAckSink agent will
  immediately ACK the first (SYN) packet sent by a TcpAgent. The patch contains
  changes to tcp-sink.{cc,h}, ns-default.tcl, and affected validation tests.
  All validation tests that use the syn_ in TcpAgent were set to
  SYN_immediate_ack=false except for the delayed test in test-suite-simple.tcl.
* During long simulations (30-60 minutes), ns would sometimes abort with the
  message "TcpAgent: negative RTO!". The problem was that \ 
TcpAgent::t_backoff_
  was overflowing and becoming negative, since it was doubled each time a
  backoff was triggered but maxrto_ prevented the time between doublings from
  growing beyond 60 seconds.
* This patch adds support for using one-way TcpAgents with the Tmix traffic
  generator, which previously only worked with FullTcp agents.
* Protocol for Unified Multicasting Through Announcements (PUMA)
  has been added.  PUMA is a distributed, receiver initiated, mesh based
  multicast routing protocol.
* A new SCTP patch from Nasif Ekiz and the University of Delaware team.
* added support for Non-Renegable Selective Acknowledgements (NR-SACK).
  To enable NR-SACKs, set useNonRenegSacks_ tcl bindable variable to 1 at
  both the data sender and the data receiver.
* Concurrent Multipath Transfer with Potentially Failed(CMT-PF) extension
  is default now. To use CMT, set useCmtPF tcl bindable variable to 0
  at the data sender.
* added new feature to support a limited size send buffer at the data
  receiver. Send buffer size can be specified by setting initialSwnd_ tcl
  bindable variable to desired value. By default, the variable is set to
  0 which means there is no constraint on the send buffer.
* congestion window (cwnd) updates are now done as specified at Section 7.2.1,
  7.2.2 and 7.2.3 of RFC4960. Also the number of missing reports to trigger
  fast retransmits is reduced to three as specified in RFC4960.
* bug fix. A bug discovered by Eduardo Ribeiro and Igor Gavriloff regarding
  to memory deallocation is fixed.
* new validation tests for NR-SACKs and limited send buffer size are added to
  the test suite.

ns-2.34 Released on June 17, 2009

* bugfix for 802.11Ext model (from Felix-Schmidt-Eisenlohr and team)
* mac/wirelessPhyExt:
* the power monitor class is responsible to continously track the sum of all
  reception power values of all frames arriving in parallel and of the noise
  floor. In the earlier implementation this cumulated power was only decreased
  in situations where no packet arrived. Otherwise, powers were always further
  summed up, leading to a wrong setting of this variable, and in consequence,
  a wrong behavior of physical packet reception calculations. Furthermore, the
  busy/idle signalling was not correct.
* Add AOMDV (On-demand Multipath Distance Vector Routing in Ad Hoc Networks)
  implementation from Marcello Caleffi
* Collection of bug fixes for 802.11Ext model (from Felix-Schmidt-Eisenlohr
  and team)
* mac/mac-802.11Ext:
* feedback to higher layers on transmission failure: xmit_failure_data_ is
  delivered when a unicast packet was finally not sent successfully- corrected
  NAV update: in case a NAV is active and a new NAV is advertised an updated
  only occurs in case the new NAV takes longer than the one already running
  - bugfix: corrected backoff handling when 0 backoff slots are selected:-
  bugfix: resolved memory leak on broadcast packet transmission- duplicate
  suppression support: reactivate the feature of duplicate detection on MAC
  layer as it was done in earlier versions
* code cleanup: removed TxTimer_t class
* bugfix: ACK frames: explicitly set basic modulation scheme
* mac/wirelessPhyExt:
* update of modulation table: new threshold values are taken for the different
  modulation schemes due to newer measurements
* bugfix: resched() instead of sched() in PowerMonitor::expire()
* code cleanup: removed friend class PowerTimer
* code cleanup: updated comment in recordPowerLevel()
* new implementation of PowerMonitor (PowerTimer class now obsolete)
* clarification of CSThresh variable (see below)
* mac/channel.cc: for wirelessPhyExt only: correct calculation of distCST,
  using PowerMonitorThresh_ as the critical borderline power and including
  antenna gains.
* tools/ranvar.cc: - bugfix: resolved memory leak in Erlang distribution
* Bug fixes for 802.11 infrastructure code, posted by Ilango Purushothaman.
* Patches for better memory management - use standard C++ lists to maintain AP
  and client tables, memory leaks fixed in AP scheduling queue
* Triggering of handoff in Ad-hoc mode - Bug reported by Mayur - Resolved by
  adding an if condition for checking the 802.11 mode.
* Note:  The following traces changed as a result of this, but were verified
  by Ilango: ./test-all-wireless-infra ./test-all-wireless-infra-mobility
  ./test-all-wireless-shadowing ./test-all-wireless-lan-aodv
  ./test-all-wireless-gridkeeper ./test-all-wireless-lan-newnode
* Import Tmix synthetic Internet traffic generation tool
* Disable string literal warning for gcc version > 4.2

Log message:
Drop superfluous PKG_DESTDIR_SUPPORT, "user-destdir" is default these days.

Log message:
Restore patches to configure script lost in an update some time back.
Does not fix the build, but advances it; now it's having some kind
of bizarre problem with <sys/queue.h>.

While here, add three small patches to silence a large pile of const
warnings that made it impossible to see anything else that was going
on.

Log message:
WRKSRC should use PKGNAME_NOREV.