]> rtime.felk.cvut.cz Git - orte.git/log
orte.git
10 years agoJORTE: update header files
Martin Vajnar [Sun, 21 Jul 2013 09:04:13 +0000 (11:04 +0200)]
JORTE: update header files

Remove defines of UINT32__MIN and UINT32__MAX, because they are in the
<stdint.h> and UINT32__MAX is computed in a wrong way.

Make the HAVE_CONFIG_H hack Android specific and add extern "C" for C++
compatibility to onLoad.h

10 years agoANDROID: remove JORTEDomainStart.c from makefile
Martin Vajnar [Sun, 21 Jul 2013 08:35:25 +0000 (10:35 +0200)]
ANDROID: remove JORTEDomainStart.c from makefile

10 years agoJORTE: remove unused code
Martin Vajnar [Sun, 21 Jul 2013 08:25:50 +0000 (10:25 +0200)]
JORTE: remove unused code

Since the domain is started (receiving and sending threads) either by
calling ORTEDomainAppCreate() or ORTEDomainMgrCreate(), there is no need
to have DomainStart() on the Java side.

10 years agoJORTE: remove unused code
Martin Vajnar [Sat, 20 Jul 2013 22:11:40 +0000 (00:11 +0200)]
JORTE: remove unused code

10 years agoJORTE: incorporate new application domain destruction into example
Martin Vajnar [Sat, 20 Jul 2013 22:10:58 +0000 (00:10 +0200)]
JORTE: incorporate new application domain destruction into example

10 years agoJORTE: call correct JNI function to get values from NtpTime class
Martin Vajnar [Sat, 20 Jul 2013 22:09:14 +0000 (00:09 +0200)]
JORTE: call correct JNI function to get values from NtpTime class

Replace GetIntField() with GetLongField().

10 years agoJORTE: fix the '2^32' problem and formatting in debug printf()
Martin Vajnar [Sat, 20 Jul 2013 22:06:34 +0000 (00:06 +0200)]
JORTE: fix the '2^32' problem and formatting in debug printf()

This fixes problems with wrong calculation of (2^32) on many occasions
both on the Java and the C side.

Many format strings for printf() are fixed. Display Host ID
and App ID in hexadecimal (as the ortemanager does).

10 years agoJORTE: disable TEST_STAGE
Martin Vajnar [Fri, 19 Jul 2013 13:33:08 +0000 (15:33 +0200)]
JORTE: disable TEST_STAGE

10 years agoROBOT_DEMO: be less verbose
Martin Vajnar [Fri, 19 Jul 2013 13:32:09 +0000 (15:32 +0200)]
ROBOT_DEMO: be less verbose

10 years agoROBOT_DEMO: completely rewritten
Martin Vajnar [Fri, 19 Jul 2013 09:37:04 +0000 (11:37 +0200)]
ROBOT_DEMO: completely rewritten

This adds complete rewrite of the demo application. Buttons are now removed
from views and are instead placed to a menu. At the start a domain manager
and an application domain are created under which every subscriber
and publisher is launched. Support for hokuyo_scan is added.

10 years agoANDROID: incorporate DomainProp changes
Martin Vajnar [Fri, 19 Jul 2013 09:30:27 +0000 (11:30 +0200)]
ANDROID: incorporate DomainProp changes

10 years agoJORTE: simplify Manager destruction and remove unnecessary println()
Martin Vajnar [Fri, 19 Jul 2013 09:29:33 +0000 (11:29 +0200)]
JORTE: simplify Manager destruction and remove unnecessary println()

10 years agoJORTE: don't force application domain destruction
Martin Vajnar [Fri, 19 Jul 2013 09:25:15 +0000 (11:25 +0200)]
JORTE: don't force application domain destruction

This prevents appDomain being destroyed along with publisher or subscriber.

10 years agoJORTE: use abstract methods
Martin Vajnar [Fri, 19 Jul 2013 09:22:22 +0000 (11:22 +0200)]
JORTE: use abstract methods

10 years agoJORTE: fix subscription destruction
Martin Vajnar [Fri, 19 Jul 2013 09:20:39 +0000 (11:20 +0200)]
JORTE: fix subscription destruction

Destroy subscription first, then destroy references to objects
in subscription callback. This prevents SEGFAULTs from the receiving
thread, that tries to save received instance to a non-existent object.
DeleteLocalReference was replaced with DeleteGlobalReference, since we are
destroying a global reference. This prevents dvmAbort() being called
on Android. This also fixes a memory leak (the ctx->msg reference was not
deleted).

10 years agoJORTE: pass instance of MessageData to Java's receive callback
Martin Vajnar [Fri, 19 Jul 2013 09:12:50 +0000 (11:12 +0200)]
JORTE: pass instance of MessageData to Java's receive callback

10 years agoJORTE: don't use finalizers in Java classes
Martin Vajnar [Tue, 16 Jul 2013 09:48:07 +0000 (11:48 +0200)]
JORTE: don't use finalizers in Java classes

This replaces finalizers with explicit destroy() functions, because the
order in which objects with no reference in the VM are garbage-collected
is not guaranteed. This caused trouble with finalization of DomainApp
and Publisher (or Subscriber). If the DomainApp was destroyed before
the Publisher (or Subscriber) the Publisher's finalizer would have been
stuck (eventually leading to TimeoutException to be thrown).

10 years agoJORTE: fix memory leak in jORTEDomainPropDefaultGet()
Martin Vajnar [Mon, 15 Jul 2013 13:02:28 +0000 (15:02 +0200)]
JORTE: fix memory leak in jORTEDomainPropDefaultGet()

This fixes memory leak in jORTEDomainPropDefaultGet and removes some
unused code.

10 years agoJORTE: switch to direct ByteBuffer
Martin Vajnar [Sun, 14 Jul 2013 18:08:07 +0000 (20:08 +0200)]
JORTE: switch to direct ByteBuffer

This commit significantly changes the way new issues are sent and received
both on the Java side and on the C side.

In the past there was one Java instance of HeapByteBuffer in the
MessageData class. ByteBuffer is used to perform serialization and
deserialization of received and sent issues in Java. It is compatible with
the CORBA standard that is used in the main ORTE library.

The HeapByteBuffer has one significant drawback. It allocates memory
on the heap managed by Java VM's garbage collector and when accessing
the backing array from C it has to copy the entire array to new space
in memory outside the VM. Added to that the backing array wasn't accessed
as an array object from the C code, but rather every byte from the Java
buffer was copied to a C byte array or vice versa (publisher x subscriber
part). This adds a lot of overhead (calling JNI functions to obtain new
values).

Instead the DirectByteBuffer is now used which allocates memory directly
(outside of Java VM) through malloc() and stores the pointer in such a way,
that it is easily accesible from C code using GetDirectBufferAddress,
that returns (void *). The Java part behaves much the same as normal
(Heap) ByteBuffer. The allocated byte array could be used like any other
C array and the VM ensures that the memory region allocated by direct
ByteBuffer is contiguous. This is very useful, because normal Java arrays
are not guaranteed to be contiguous which makes them difficult to handle,
because JNI functions must be called that make copies of them in a
contiguous space and that adds another overhead.

As a bonus the C code for both publisher and subscriber Java wrappers
is much shorter and simpler than it used to be.

10 years agoJORTE: fix endianness the proper way
Martin Vajnar [Sun, 14 Jul 2013 12:57:21 +0000 (14:57 +0200)]
JORTE: fix endianness the proper way

Use the GNU C compiler macro __BYTE_ORDER__ to set data_endian on the C
side and ByteOrder.nativeOrder() function to set ByteBuffer's order
in the MessageData class on the Java side.

10 years agoANDROID: correct return value
Martin Vajnar [Sat, 13 Jul 2013 17:15:34 +0000 (19:15 +0200)]
ANDROID: correct return value

10 years agoROBOT_DEMO: add proper destructor in C version
Martin Vajnar [Sat, 13 Jul 2013 17:11:51 +0000 (19:11 +0200)]
ROBOT_DEMO: add proper destructor in C version

10 years agoROBOT_DEMO: decrease CPU usage
Martin Vajnar [Sat, 13 Jul 2013 16:50:14 +0000 (18:50 +0200)]
ROBOT_DEMO: decrease CPU usage

Now there are 2 major threads on the Java side. The first one is main which
takes care of drawing GUI and receives callbacks from accelerometer.
The second one computes and sends values to the robot's motors.

10 years agoJORTE: update required Java API version
Martin Vajnar [Thu, 11 Jul 2013 12:52:39 +0000 (14:52 +0200)]
JORTE: update required Java API version

10 years agoROBOT_DEMO: use Manager class instead of ortemanager binary
Martin Vajnar [Thu, 11 Jul 2013 11:46:23 +0000 (13:46 +0200)]
ROBOT_DEMO: use Manager class instead of ortemanager binary

10 years agoANDROID: replace native binary ortemanager with Java class
Martin Vajnar [Thu, 11 Jul 2013 11:44:44 +0000 (13:44 +0200)]
ANDROID: replace native binary ortemanager with Java class

This eliminates the need for binary ortemanager in assets/ folder
of projects that make use of this library.

NOTE: At current moment it is only possible to set IP addresses of fellow
managers.

10 years agoJORTE: add basic Java domain manager
Martin Vajnar [Thu, 11 Jul 2013 11:40:26 +0000 (13:40 +0200)]
JORTE: add basic Java domain manager

This adds basic Java manager. Currently it is only possible to set IP
addresses of fellow managers.

10 years agoMerge branch 'android' of ssh://git@rtime.felk.cvut.cz/orte into robot
Martin Vajnar [Sun, 7 Jul 2013 21:45:08 +0000 (23:45 +0200)]
Merge branch 'android' of ssh://git@rtime.felk.cvut.cz/orte into robot

10 years agoJORTE: use brackets around '^'
Martin Vajnar [Sun, 7 Jul 2013 21:40:44 +0000 (23:40 +0200)]
JORTE: use brackets around '^'

This avoids compiler warnings.

10 years agoMANAGER: add a variable for triggered signal
Martin Vajnar [Sun, 7 Jul 2013 21:39:25 +0000 (23:39 +0200)]
MANAGER: add a variable for triggered signal

This eliminates -Wnonnull warning during compilation.

10 years agoJORTE: use correct format specifiers
Martin Vajnar [Sun, 7 Jul 2013 21:38:13 +0000 (23:38 +0200)]
JORTE: use correct format specifiers

10 years agoROBOT_DEMO: add robot demo for Android
Martin Vajnar [Sun, 7 Jul 2013 18:16:52 +0000 (20:16 +0200)]
ROBOT_DEMO: add robot demo for Android

Initial commit of robot demo for Android.

10 years agoANDROID: add support for Android
Martin Vajnar [Sun, 7 Jul 2013 17:24:59 +0000 (19:24 +0200)]
ANDROID: add support for Android

Add preliminary Android library.

Note: Current Android NDK doesn't allow libraries to have assets folder,
but we place the ortemanager in there. At current point it is necessary
for applications using this library to either manually copy the compiled
ortemanager over to their own assets folder or to create a symlink.

10 years agoORTE: Android hack - define HAVE_CONFIG_H
Martin Vajnar [Sun, 7 Jul 2013 17:15:04 +0000 (19:15 +0200)]
ORTE: Android hack - define HAVE_CONFIG_H

10 years agoJORTE: alter sendCallBack
Martin Vajnar [Sun, 7 Jul 2013 17:13:41 +0000 (19:13 +0200)]
JORTE: alter sendCallBack

Alter the sendCallBack, so it doesn't scramble the sent instance.

10 years agoMANAGER: prepare for Android
Martin Vajnar [Sun, 7 Jul 2013 17:10:53 +0000 (19:10 +0200)]
MANAGER: prepare for Android

Android doesn't support sigwaitinfo, instead it offers only sigwait.
Since we don't use the information about which signal is being triggered
we can safely switch to sigwait.

10 years agoJORTE: prepare for Android (fix class loader problems + add logging)
Martin Vajnar [Sun, 7 Jul 2013 16:50:01 +0000 (18:50 +0200)]
JORTE: prepare for Android (fix class loader problems + add logging)

This fixes class loading problems on Android.

As per the Android Documentation (http://developer.android.com/training/articles/perf-jni.html#faq_FindClass):

Android's Dalvik VM uses the loader associated with the method
at the top of the interpreted stack, or if there isn't one
(because the thread was just attached), it uses the "system" class
loader, for FindClass calls.

To overcome this the JNI_OnLoad function is used. Any FindClass calls made
from there will happen in the context of the class loader used to load
the shared library. So from there one class that will always be present
in the JORTE Java package is chosen (org.ocera.orte.JOrte) and it's
class loader is stored as a global reference inside the VM (thus shared
by all processes attached to the VM) and as a static variable (accesible
only by functions from the onLoad.c file) on the C side.Also the findClass
and findLoadedClass method IDs are cached. To simplify the class loading
process inside C functions the findClass functions is written.

Android also doesn't implement the printf function. However it does have
a logging capability, so a variadic macro replacing occurences of printf
with __android_log_print is used.

10 years agoJORTE: add 64-bit support
Martin Vajnar [Sun, 7 Jul 2013 15:12:27 +0000 (17:12 +0200)]
JORTE: add 64-bit support

Store pointers to C objects in jlong (always 64-bit)
rather than in jint (always 32-bit).

10 years agoJORTE: fix typo
Martin Vajnar [Sun, 7 Jul 2013 11:47:07 +0000 (13:47 +0200)]
JORTE: fix typo

Correct type conversions.

10 years agoJORTE: fix inverted logic and add brackets
Martin Vajnar [Sun, 7 Jul 2013 11:43:59 +0000 (13:43 +0200)]
JORTE: fix inverted logic and add brackets

Fix inverted logic bug, that prevented proper native
thread detachment from Java VM.

10 years agoJORTE: fix inverted logic
Martin Vajnar [Sun, 7 Jul 2013 11:29:44 +0000 (13:29 +0200)]
JORTE: fix inverted logic

10 years agoJORTE: attempt at fixing endianness problems
Martin Vajnar [Sun, 7 Jul 2013 11:27:22 +0000 (13:27 +0200)]
JORTE: attempt at fixing endianness problems

This is an attempt at fixing endianness problems
that occur when communicating with traditional ORTE
publishers/subscribers written in pure C.

10 years agoJORTE: fix typo
Martin Vajnar [Sun, 7 Jul 2013 11:19:57 +0000 (13:19 +0200)]
JORTE: fix typo

10 years agoJORTE: fix class name
Martin Vajnar [Sun, 7 Jul 2013 11:17:02 +0000 (13:17 +0200)]
JORTE: fix class name

10 years agoJORTE: add const qualifier to function declarations
Martin Vajnar [Sun, 7 Jul 2013 11:13:32 +0000 (13:13 +0200)]
JORTE: add const qualifier to function declarations

11 years agoAdd 0.3.3 release notes
Michal Sojka [Wed, 15 Aug 2012 12:05:48 +0000 (14:05 +0200)]
Add 0.3.3 release notes

11 years agoAdd .mailmap to generate nice shortlog
Michal Sojka [Wed, 15 Aug 2012 11:56:24 +0000 (13:56 +0200)]
Add .mailmap to generate nice shortlog

11 years agoPut compiled ORTE manual to the distribution tarballs v0.3.3
Michal Sojka [Wed, 15 Aug 2012 10:42:09 +0000 (12:42 +0200)]
Put compiled ORTE manual to the distribution tarballs

11 years agoUpdate Makefile.rules for OMK build to the actual version.
Pavel Pisa [Wed, 15 Aug 2012 10:42:56 +0000 (12:42 +0200)]
Update Makefile.rules for OMK build to the actual version.

Documentation for OMK Make-System can be found at

  http://rtime.felk.cvut.cz/omk/

Signed-off-by: Pavel Pisa <pisa@cmp.felk.cvut.cz>
11 years agoRemove pointer to OCERA_TOP_DIR, ORTE is used only as standalone in current projects.
Pavel Pisa [Wed, 15 Aug 2012 10:40:06 +0000 (12:40 +0200)]
Remove pointer to OCERA_TOP_DIR, ORTE is used only as standalone in current projects.

Signed-off-by: Pavel Pisa <pisa@cmp.felk.cvut.cz>
11 years agodoc: Add bookinfo
Michal Sojka [Wed, 15 Aug 2012 10:30:31 +0000 (12:30 +0200)]
doc: Add bookinfo

11 years agoUpdate download instructions
Michal Sojka [Wed, 15 Aug 2012 10:21:46 +0000 (12:21 +0200)]
Update download instructions

11 years agodoc: Restructure - split it to multiple chapters
Michal Sojka [Wed, 15 Aug 2012 10:21:31 +0000 (12:21 +0200)]
doc: Restructure - split it to multiple chapters

11 years agodoc: Update sizes of figures to have readable text
Michal Sojka [Wed, 15 Aug 2012 09:50:26 +0000 (11:50 +0200)]
doc: Update sizes of figures to have readable text

11 years agodoc: Move API reference to the end
Michal Sojka [Wed, 15 Aug 2012 09:43:33 +0000 (11:43 +0200)]
doc: Move API reference to the end

11 years agoUpdate Makefile to properly generate documentation
Michal Sojka [Wed, 15 Aug 2012 09:35:40 +0000 (11:35 +0200)]
Update Makefile to properly generate documentation

11 years agoAdd sources of figures
Michal Sojka [Wed, 15 Aug 2012 09:33:41 +0000 (11:33 +0200)]
Add sources of figures

11 years agoAdd documentation from 0.3.2 release tarball
Michal Sojka [Wed, 15 Aug 2012 09:23:37 +0000 (11:23 +0200)]
Add documentation from  0.3.2 release tarball

11 years agoDo not distribute autom4te.cache
Michal Sojka [Wed, 15 Aug 2012 08:53:32 +0000 (10:53 +0200)]
Do not distribute autom4te.cache

11 years agoRemove automatically generated file orte_config.h.in
Michal Sojka [Wed, 15 Aug 2012 08:39:51 +0000 (10:39 +0200)]
Remove automatically generated file orte_config.h.in

11 years agoUpdate version to 0.3.3 even in header files for OMK build.
Pavel Pisa [Wed, 15 Aug 2012 08:33:16 +0000 (10:33 +0200)]
Update version to 0.3.3 even in header files for OMK build.

Signed-off-by: Pavel Pisa <pisa@cmp.felk.cvut.cz>
11 years agoPrepare for release
Michal Sojka [Tue, 14 Aug 2012 17:17:14 +0000 (19:17 +0200)]
Prepare for release

11 years agoBump version to 0.3.3
Michal Sojka [Tue, 14 Aug 2012 17:07:03 +0000 (19:07 +0200)]
Bump version to 0.3.3

11 years agoreadme: Put compile/install section in front of Git specific sections
Michal Sojka [Tue, 14 Aug 2012 16:26:26 +0000 (18:26 +0200)]
readme: Put compile/install section in front of Git specific sections

11 years agoUpdate readme
Michal Sojka [Tue, 14 Aug 2012 16:24:02 +0000 (18:24 +0200)]
Update readme

11 years agoRename .cvsignore to .gitignore
Michal Sojka [Tue, 14 Aug 2012 15:42:39 +0000 (17:42 +0200)]
Rename .cvsignore to .gitignore

13 years agoidl: Generate (de)serialization functions for enum types
Michal Sojka [Fri, 28 May 2010 11:28:11 +0000 (13:28 +0200)]
idl: Generate (de)serialization functions for enum types

13 years agoIDL: Fix *_(de)serialize macros for typedefed types
Michal Sojka [Fri, 21 May 2010 10:36:24 +0000 (12:36 +0200)]
IDL: Fix *_(de)serialize macros for typedefed types

These macros had a wrong number of parameters, which resulted in
compilation errors.

14 years agoZero mreq structure before using it
Michal Sojka [Wed, 31 Mar 2010 15:48:40 +0000 (17:48 +0200)]
Zero mreq structure before using it

There might be unset fields in mreq structure. This structure is
used to setup multicast group membership.

We have problem with multicast on MPC5200-based boards and this was an
unsuccessful attempt to solve the problems.

14 years agoReport correct errno in case of multicast problem
Michal Sojka [Wed, 31 Mar 2010 11:27:01 +0000 (13:27 +0200)]
Report correct errno in case of multicast problem

14 years agoFix for gcc-4.4
wentasah [Thu, 5 Nov 2009 05:58:46 +0000 (05:58 +0000)]
Fix for gcc-4.4

14 years agoThe Shape Demo is now optional part of OMK build.
ppisa [Thu, 11 Jun 2009 20:19:24 +0000 (20:19 +0000)]
The Shape Demo is now optional part of OMK build.

14 years agoMakefile.rules updated to work correctly with QT build of shape demo.
ppisa [Thu, 11 Jun 2009 20:17:17 +0000 (20:17 +0000)]
Makefile.rules updated to work correctly with QT build of shape demo.

14 years agoAdded helper makefile to the contrib to allow direct ORTE demo build.
ppisa [Thu, 11 Jun 2009 19:59:15 +0000 (19:59 +0000)]
Added helper makefile to the contrib to allow direct ORTE demo build.

14 years agoThe ORTE demo can be compiled by OMK Makefiles now.
ppisa [Thu, 11 Jun 2009 19:55:47 +0000 (19:55 +0000)]
The ORTE demo can be compiled by OMK Makefiles now.

The directory orte/contrib/shape can be build by regular
make process if orte/contrib/shape/Makefile is replaced
by regular OMK Makefile helper - i.e. orte/Makefile
or omk build stript is run with USE_LEAF_MAKEFILES=n

  echo CONFIG_OC_ETH_ORTE_DEMO_SHAPE=y >>config.omk
  cd orte/contrib/shape
  omk USE_LEAF_MAKEFILES=n

14 years agofixed counting of get_max_size for structures
smolik [Wed, 20 May 2009 14:33:46 +0000 (14:33 +0000)]
fixed counting of get_max_size for structures

14 years agoCORBA_char redefined to char
wentasah [Wed, 20 May 2009 12:20:57 +0000 (12:20 +0000)]
CORBA_char redefined to char

This is to avoid signedness warnings when using str* functions.

14 years agoORTE IDL learned how to handle arrays
wentasah [Wed, 20 May 2009 12:20:50 +0000 (12:20 +0000)]
ORTE IDL learned how to handle arrays

14 years agoCorrected termination of ortemanager
wentasah [Wed, 20 May 2009 12:20:43 +0000 (12:20 +0000)]
Corrected termination of ortemanager

Sometimes I've realized that it was not possible to terminate ortemanager
by pressing Ctrl-C. It seems it was the case when I disconnecte from the
network and wanted to finish ortemanager later.

Since it is not safe to call pthread_mutex_lock() from signal handlers,
I've reworked termination to use sigwaitinfo() to wait for termination
signals.

14 years agoBetter fix of receive errors
wentasah [Wed, 20 May 2009 12:20:37 +0000 (12:20 +0000)]
Better fix of receive errors

Signed type is used as return value to catch -1 (error).
The receive thread exits on error in order to stop infinite loops.

14 years agoComplain verbosely if something goes wrong in receiver
wentasah [Wed, 20 May 2009 12:20:30 +0000 (12:20 +0000)]
Complain verbosely if something goes wrong in receiver

14 years agoRemoved circular make dependency in libddso
wentasah [Wed, 20 May 2009 12:20:23 +0000 (12:20 +0000)]
Removed circular make dependency in libddso

14 years agoortedemo checks for errors in publisher/subscriber creation
wentasah [Wed, 20 May 2009 12:20:17 +0000 (12:20 +0000)]
ortedemo checks for errors in publisher/subscriber creation

Signed-off-by: Michal Sojka <sojkam1@fel.cvut.cz>
14 years agoBetter error handling in ORTEDomainCreate()
wentasah [Wed, 20 May 2009 12:20:10 +0000 (12:20 +0000)]
Better error handling in ORTEDomainCreate()

Previous implementation did not check for errors in bind() and when
ortemanager was run two times, the second instance did not fail and
instead consumed 100% of CPU.

14 years agoFixed error handling in ORTEDomainCreate()
wentasah [Wed, 20 May 2009 12:20:01 +0000 (12:20 +0000)]
Fixed error handling in ORTEDomainCreate()

Signed-off-by: Michal Sojka <sojkam1@fel.cvut.cz>
14 years agoIDL compiler is compiled before anything else
wentasah [Wed, 20 May 2009 12:19:35 +0000 (12:19 +0000)]
IDL compiler is compiled before anything else

15 years agoorte 0.3.2 release v0.3.2
smolik [Tue, 7 Oct 2008 21:19:01 +0000 (21:19 +0000)]
orte 0.3.2 release

15 years agoThe uLUt files included in ORTE updated to actual version from uLan repository.
ppisa [Tue, 7 Oct 2008 08:59:07 +0000 (08:59 +0000)]
The uLUt files included in ORTE updated to actual version from uLan repository.

15 years agoInitialization of the first_changed field of timer queue added to make Valgrind happy.
ppisa [Tue, 7 Oct 2008 08:08:38 +0000 (08:08 +0000)]
Initialization of the first_changed field of timer queue added to make Valgrind happy.

15 years agoCorrected seeking for MAKERULES_DIR in leaf-node Makefile redirector.
ppisa [Mon, 7 Jul 2008 07:56:26 +0000 (07:56 +0000)]
Corrected seeking for MAKERULES_DIR in leaf-node Makefile redirector.

The operator "==" is not defined for some non-bash species of shells,
the use of "=" operator corrects the problem.

16 years agoFixed incorrect behaviour when is received ACK on missing GAP - note:GAP dynamic...
smolik [Mon, 3 Dec 2007 11:15:58 +0000 (11:15 +0000)]
Fixed incorrect behaviour when is received ACK on missing GAP - note:GAP dynamic changeing size.

17 years agoadded new contributor
smolik [Mon, 16 Apr 2007 20:21:55 +0000 (20:21 +0000)]
added new contributor

17 years agoVirgil Ansems:
smolik [Mon, 16 Apr 2007 20:18:29 +0000 (20:18 +0000)]
Virgil Ansems:
If the call to inet_addr fails, the value INADDR_NONE is returned, which usually is -1. Therefore the call to ntohl will return a large number (the return type is unsigned) instead of 0.

17 years agoFixed a bug in the previous ORTE commit.
wentasah [Thu, 22 Mar 2007 20:57:05 +0000 (20:57 +0000)]
Fixed a bug in the previous ORTE commit.

17 years agoAdded an option to allow ortemanager listen on a specific interface (-L).
wentasah [Wed, 21 Mar 2007 20:00:52 +0000 (20:00 +0000)]
Added an option to allow ortemanager listen on a specific interface (-L).

17 years agoEndian is correctly detected if compiled by OMK.
wentasah [Wed, 21 Mar 2007 19:57:19 +0000 (19:57 +0000)]
Endian is correctly detected if compiled by OMK.

17 years agoVirgil Ansems patch to correct state handling with commStateSend in CSTWriterAddCSCha...
smolik [Tue, 20 Feb 2007 10:39:06 +0000 (10:39 +0000)]
Virgil Ansems patch to correct state handling with commStateSend in CSTWriterAddCSChange -> MUSTSENDDATA in any case of new csChange

17 years agochanged name to Open Real-time Ethernet, some source header arranging
smolik [Mon, 25 Sep 2006 20:07:43 +0000 (20:07 +0000)]
changed name to Open Real-time Ethernet, some source header arranging

17 years agofixed message lost for case no memory space at sending message queue
smolik [Wed, 13 Sep 2006 21:28:11 +0000 (21:28 +0000)]
fixed message lost for case no memory space at sending message queue

17 years agoa small bug reported by Carl Wolff, on fast cpu new incomming message is faster then...
smolik [Tue, 22 Aug 2006 05:53:54 +0000 (05:53 +0000)]
a small bug reported by Carl Wolff, on fast cpu new incomming message is faster then persistence timer. Some throubles occure with small times - zero.