This changes the buildlink3.mk files to use an include guard for the
recursive include. The use of BUILDLINK_DEPTH, BUILDLINK_DEPENDS,
BUILDLINK_PACKAGES and BUILDLINK_ORDER is handled by a single new
variable BUILDLINK_TREE. Each buildlink3.mk file adds a pair of
enter/exit marker, which can be used to reconstruct the tree and
to determine first level includes. Avoiding := for large variables
(BUILDLINK_ORDER) speeds up parse time as += has linear complexity.
The include guard reduces system time by avoiding reading files over and
over again. For complex packages this reduces both %user and %sys time to
half of the former time.
* From schedctl(8) of NetBSD-5-RC2, setting of priority for the process running
at SCHED_OTHER policy is not allowed, so change to skip if SCHED_OTHER is specified.
* Use sched_get_priority_min() instead of hard coded DEFAULT_SCHED_PARAM (= 0).
* Ajust return code as the "Returns:" section in this function's comment.
sched_setscheduler(3) return the previous policy if succeed, it may not be 0.
Bump PKGREVISION.
Changes:
*******************************************************************************
Version 1.6.6
*******************************************************************************
2008-04-24 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
* Added thread id's to the UpnpPrintf debug messages. Thanks to
Charles Nepveu for the idea.
2008-04-24 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
* SF Bug Tracker [ 1948586 ]
Uppercase U in in "xmlns:U" in Invoke Action causes seg. f.
Submitted By: Thomas Norheim - kjakan_no
Device no longer segfaults with the following malformed xml action:
<u:SetColor xmlns:U="urn:schemas-upnp-org:service:tvpicture:1">
<Color>2</Color>
</u:SetColor>
2008-04-23 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
* Use -O0 in debug builds so that variables do not get optimized out.
2008-04-10 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
* Apostolos Syropoulos changes for OpenSolaris x86.
2008-03-20 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
* Andre Sodermans (wienerschnitzel) patch for building libupnp under
windows systems with VC9.
2008-03-20 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
* Andre Sodermans (wienerschnitzel) patch for building libupnp under
windows systems with VC8.
2008-03-08 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
* Fixed a printf format problem on the upnp_tv_device.c from both
upnp/sample/tvdevie and upnp/sample/tvcombo directories. The variable
port was a short int instead of an unsigned short and it was beeing
print as a negative value.
2008-03-08 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
* SF Bug Tracker [ 1902668 ] Cannot compile on MSVC
Submitted By Luke Kim - nereusuj
Version 1.6.5 cannot be compiled because of some changes in 1.6.3.
MSVC does not support stdint.h, gettimeofday(), sys/param.h, const int
variables in array size and Windows does not define _WINDOWS_ but define
_WINDOWS.
* MSVC does not understand "const int"'s as declarators of array
dimensions, we must use #define'd constants.
* Use WIN32 instead of _WINDOWS_ or _WINDOWS.
2008-02-22 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
* No longer ignore "upnp:rootdevice" advertisement. Thanks to Bob Ciora.
2008-02-10 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
* Changed "sys_errlist[errno]", which is deprecated, by
"strerror_r()", which is thread safe.
2008-02-06 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
* Slightly improved error report by showing the sys_errlist string
corresponding to errno.
2008-02-06 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
* Got rid of two useless constants: UPNP_SOCKETERROR and
UPNP_INVALID_SOCKET. They both mean the same, that a network API
function has failed. -1 is the value to check, not an invented constant.
Changes:
*******************************************************************************
Version 1.6.5
*******************************************************************************
2008-02-02 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
* Peter Hartley's fix for wrong sized variable beeing passed to
http_MakeMessage() on 64 bit architectures.
Changes:
*******************************************************************************
Version 1.6.4
*******************************************************************************
2008-01-23 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
* Workaround for a problem with the new automake AM_CONDITIONAL macro
from autotools-1.10. Thanks to Ingo Hofmann for helping with debugging
this one.
2008-01-22 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
* Added quoting to macros AC_CONFIG_AUX_DIR, AC_CONFIG_MACRO_DIR and
AC_CONFIG_SRCDIR in configure.ac. Also changed the name of the
auxiliary directory in AC_CONFIG_AUX_DIR to build-aux.
2008-01-22 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
* Fix for setsockopt() in Threadpool.c to allow more than one process
to join the multicast-group on OSX. Thanks to Ingo Hofmann.
2008-01-22 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
* Using defined(__OSX__) || defined(__APPLE__) instead of just
defined(__OSX__) in the code. Thanks to Ingo Hofmann and Chris
Pickel.
2008-01-21 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
* Fix for isFileInVirtualDir. Thanks to Peter Hartley for the patch.
2008-01-07 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
* Putting back a "defined(__OSX__)" that has been removed in the
previous *BSD patch. Thanks to Chris Pickel for pointing it out.
2008-01-07 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
* SF Patches Tracker [ 1865812 ] typo in docs comment
Submitted By: Hartmut Holzgraefe - hholzgra
typo in docs comment ACCAPTED instead of ACCEPTED in
@name UPNP_E_UNSUBSCRIBE_UNACCAPTED [-302]
Also, the documentation file name was mispelled and was corrected in
the Makefile.am.
*******************************************************************************
Version 1.6.3
*******************************************************************************
2007-12-25 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
* Using pthread flags for the whole project, not just at the places
individually indicated by several Makefile.am files spread all over
the directories. That was too much error prone.
2007-12-24 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
* Added a configure test to check if pthread_rwlock_t is available.
Define _GNU_SOURCE if needed. The fallback behaviou will only be
implemented if _GNU_SOURCE prooves to be insufficient on some
platforms. Thanks to Jonathan Casiot (no_dice) and Robert Gingher
(robsbox).
2007-12-17 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
* Removed unused iasnprintf.{c,h} files.
2007-12-17 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
* Removed STATSONLY() macro from ThreadPool.{c,h}.
* Removed time() usage from ThreadPool.c.
* Fixed STATS = 0 compilation.
2007-12-16 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
* Library was not compiling on FreeBSD 7. Code now no longer uses
ftime(), using gettimeofday() instead. Thanks to Josh Carroll.
Pkgsrc changes:
Added test target.
Changes:
*******************************************************************************
Version 1.6.2
*******************************************************************************
2007-12-10 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
* Fixed a compilation error due to a missing #ifdef in
upnp/src/genlib/miniserver/miniserver.c. Thanks to Eugene Christensen.
2007-11-12 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
* "make check" was failing because ixml/test/test_document.sh did not
have the executable flag set. Thanks to Steve Bresson.
2007-11-12 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
* Fixed a memory leak in upnpapi.c to delete gMiniServerThreadPool in
the call to UpnpFinish(). Thanks to Fabrice Fontaine.
2007-11-09 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
* Added a isleep() call to the error handler of select() in
RunMiniServer(), so that it does not take 100% cpu in case select()
fails repeatedly.
While here, add DESTDIR support.
Changes:
*******************************************************************************
Version 1.6.1
*******************************************************************************
2007-11-07 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
* SF Bug Tracker [ 1825278 ] AdvertiseAndReply sleeps with handle lock out
Applied patch from Alex (afaucher) to change some write locks to read
locks.
2007-11-06 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
* Adjusting libtool library numbers to reflect the last changes.
2007-11-06 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
* SF Bug Tracker [ 1825278 ] AdvertiseAndReply sleeps with handle lock out
GlobalHndMutex, which was a mutex is now GlobalHndRWLock, which is a
rwlock. HandleLock() is mapped to HandleWriteLock() while all other
instances have not been checked. One instance in AdvertiseAndReply()
has been changed to HandleReadLock(). Thanks to Alex (afaucher) for the
bug report and suggestions.
2007-11-06 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
* Added support for rwlocks.
2007-11-05 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
* SF Bug Tracker [ 1825929 ] woker thread still alive after UpnpFinish()
Submitted By: Luke Kim - nereusuj
Worker thread still alive after calling UpnpFinish() because
ThreadPoolShutdown() is in the #ifdef DEBUG block.
421
422 #ifdef DEBUG
423 ThreadPoolShutdown( &gSendThreadPool );
424 ThreadPoolShutdown( &gRecvThreadPool );
2007-08-28 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
* Changed the calls to virtualDirCallback.open(filename, UPNP_WRITE)
to (virtualDirCallback.open)(filename, UPNP_WRITE) (notice the
parenthesis) due to a change in glibc that produces compilation
errors.
2007-08-28 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
* Initialization of the "randomness" struct so that valgrind does not
complain.
2007-08-06 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
* Merge of patch submitted By Keith Brindley - brindlk
SF Bug Tracker [ 1762758 ] Seek not working for large files
Problem:
Requests from the uPnP client to seek to a position beyond 2GB in a large
file are handled as a request to see from the 2GB point.
Impact:
Varies depending on client. The Xbox 360 kills the connection when it
realises.
Solution:
GetNextRange function (webserver.c) is updated to handle large file sizes.
Fix should also recognise when built on a 32bit platform rather than 64 and
handle accordingly.
2007-08-05 Marcelo Jimenez <mroberto(at)users.sourceforge.net>
* Merge of Mac OS X patch from Stephane Corthesy (davelopper),
SF Bug Tracker [ 1686420 ] Modifications for MacOSX.
Some of the proposed changes were already done by Rene Hexel's patch.
bugfixes, including a couple of serious bugs that prevented the library
from working on NetBSD and Darwin (for anything other than trivial tests).
Add a buildlink3.mk while at it.
