Neil GM Richards [Wed, 27 May 2015 13:51:03 +0000 (09:51 -0400)]
Merge branch 'bidir-deadlock' into 'ibm-trunk'
Bidir deadlock
See merge request !49
Neil Richards [Wed, 27 May 2015 13:48:42 +0000 (14:48 +0100)]
Merge branch 'bidir-deadlock.2' into bidir-deadlock
Neil GM Richards [Wed, 27 May 2015 12:33:19 +0000 (08:33 -0400)]
Merge branch 'classbasedesc_doprivileged' into 'ibm-trunk'
Classbasedesc doprivileged
See merge request !48
Neil Richards [Wed, 27 May 2015 12:32:00 +0000 (13:32 +0100)]
Use doPrivileged in ClassBaseDescriptor.init()
Neil GM Richards [Wed, 27 May 2015 11:18:46 +0000 (07:18 -0400)]
Merge branch 'class_marshalling' into 'ibm-trunk'
Class marshalling
Sorts out conversion of Class to ClassDesc during marshalling and ClassDesc to Class during unmarshalling.
Also correct ArrayValueDescriptor so that ValueDescriptor.writeReplace is called for each element in the array.
See merge request !47
Neil Richards [Wed, 27 May 2015 11:04:03 +0000 (12:04 +0100)]
Correct Class marshalling using ClassDesc
David Jencks [Thu, 21 May 2015 18:35:40 +0000 (14:35 -0400)]
work on shutdown logic; wait to close down pool until GIOPConnectionThreaded shutdowns have been run
Neil Richards [Mon, 18 May 2015 23:01:46 +0000 (00:01 +0100)]
Use ReadWriteLock for receiver threads / shutdown
David Jencks [Sun, 17 May 2015 04:23:20 +0000 (00:23 -0400)]
Fix deadlock on some jdks between write and close by synchronizing shutdown using the sendMutex_ already used for send
David Jencks [Sat, 16 May 2015 22:43:14 +0000 (18:43 -0400)]
use internal ExecutorServices instead of thread groups in ORBInstance to remove client deadlock due to expecting threads ending to notify their thread group
Neil GM Richards [Mon, 25 May 2015 02:37:10 +0000 (22:37 -0400)]
Merge branch 'enum_type_fix' into 'ibm-trunk'
Enum type fix
See merge request !46
Neil Richards [Mon, 25 May 2015 02:36:10 +0000 (03:36 +0100)]
Ensure zero suid for java.lang.Enum
Neil GM Richards [Mon, 25 May 2015 02:26:08 +0000 (22:26 -0400)]
Merge branch 'enum_type_fix' into 'ibm-trunk'
Enum type fix
Find the correct Class in an enum's type hierarchy to use in the call to Enum.valueOf()
See merge request !45
Neil Richards [Mon, 25 May 2015 02:18:21 +0000 (03:18 +0100)]
Find correct enum type in class hierarchy
Neil GM Richards [Sun, 24 May 2015 20:42:13 +0000 (16:42 -0400)]
Merge branch 'unknown_exception_handling' into 'ibm-trunk'
Unknown exception info propagation
See merge request !44
Neil GM Richards [Sat, 23 May 2015 07:35:13 +0000 (03:35 -0400)]
Merge branch 're-add_idl_entity_check' into 'ibm-trunk'
Restore isIDLEntity check
I previously misread the isIDLEntity() check as being redundant compared to IDLEntity.class.isAssignableFrom(), so removed it.
However, isIDLEntity() only compares the interfaces directly declared on the class (not in its class hierarchy), and only compares those interface class for equality with IDLEntity (not merely if they're assignable to IDLEntity).
Thus, isIDLEntity() is a stronger check, which I now understand needs to be there.
See merge request !43
Neil Richards [Sat, 23 May 2015 07:28:56 +0000 (08:28 +0100)]
Restore isIDLEntity check
Neil GM Richards [Fri, 22 May 2015 17:07:50 +0000 (13:07 -0400)]
Merge branch 'TypeRepo-with-ClassValue' into 'ibm-trunk'
Type repo with ClassValue, search repIdDesciptors cache during meta()
See merge request !42
Neil Richards [Thu, 21 May 2015 18:22:03 +0000 (19:22 +0100)]
Search TypeDescriptor cache first in meta(repId)
Neil Richards [Thu, 21 May 2015 16:23:06 +0000 (17:23 +0100)]
Indenting
Neil Richards [Thu, 21 May 2015 13:08:15 +0000 (14:08 +0100)]
Merge branch 'ibm-trunk' into TypeRepo-with-ClassValue
Neil Richards [Wed, 20 May 2015 18:54:13 +0000 (19:54 +0100)]
Propagate unchecked exceptions in UnknownExceptionInfo service contexts
Neil Richards [Mon, 18 May 2015 23:01:46 +0000 (00:01 +0100)]
Use ReadWriteLock for receiver threads / shutdown
David Jencks [Sun, 17 May 2015 04:23:20 +0000 (00:23 -0400)]
Fix deadlock on some jdks between write and close by synchronizing shutdown using the sendMutex_ already used for send
David Jencks [Sat, 16 May 2015 22:43:14 +0000 (18:43 -0400)]
use internal ExecutorServices instead of thread groups in ORBInstance to remove client deadlock due to expecting threads ending to notify their thread group
Neil GM Richards [Thu, 14 May 2015 10:00:36 +0000 (06:00 -0400)]
Merge branch 'disable_idl_name_checking_tests' into 'ibm-trunk'
Disable idl name checking tests
See merge request !41
Neil Richards [Thu, 14 May 2015 09:58:20 +0000 (10:58 +0100)]
Disable TypeCodeFactory IDL name checking tests
Neil GM Richards [Thu, 14 May 2015 08:45:19 +0000 (04:45 -0400)]
Merge branch 'disable_idl_name_checking' into 'ibm-trunk'
Disable idl name checking
TypeCode creation tries to enforce all names are IDL names.
By comparison with the behaviour seen from other ORB implementations, this is not necessarily the case, for example when generating the FullValueDescription for Java (Serializable) valuetypes.
Other parts of yoko behave in a compatible manner to these other implementations, which currently causes the IDL name checking to fail during FVD generation.
This introduces minimal changes to disable IDL name checking.
See merge request !40
Neil Richards [Thu, 14 May 2015 08:37:42 +0000 (09:37 +0100)]
Disable IDL name checking in TypeCode construction
David Jencks [Tue, 12 May 2015 18:03:35 +0000 (14:03 -0400)]
Merge branch 'collocated-csiv2' into 'ibm-trunk'
Collocated csiv2
See merge request !39
Neil GM Richards [Tue, 12 May 2015 17:23:13 +0000 (13:23 -0400)]
Merge branch 'input_stream_idl_stub_loading' into 'ibm-trunk'
Input stream idl stub loading
See merge request !38
Neil Richards [Tue, 12 May 2015 17:20:09 +0000 (18:20 +0100)]
Use IDL Helper class to create stubs for IDLEntity
David Jencks [Tue, 12 May 2015 00:40:09 +0000 (20:40 -0400)]
Determine colocation for IORs with csiv2 transport addresses. Improve a couple error messages
Neil GM Richards [Thu, 7 May 2015 10:23:48 +0000 (06:23 -0400)]
Merge branch 'further_cmsfv2_tweaks' into 'ibm-trunk'
Further cmsfv2 tweaks
See merge request !37
Neil Richards [Thu, 7 May 2015 10:20:20 +0000 (11:20 +0100)]
Corrections to reading start of CMSFv2 valuetype
Neil GM Richards [Thu, 30 Apr 2015 21:42:43 +0000 (17:42 -0400)]
Merge branch 'cmsfv2_handle_objectreader_methods' into 'ibm-trunk'
Add custom marshal hooks for ObjectReader methods
See merge request !36
Neil Richards [Thu, 30 Apr 2015 21:36:50 +0000 (22:36 +0100)]
Add custom marshal hooks for ObjectReader methods
Neil GM Richards [Thu, 30 Apr 2015 12:58:41 +0000 (08:58 -0400)]
Merge branch 'support_idl_narrow' into 'ibm-trunk'
Support idl narrow
See merge request !35
Neil Richards [Thu, 30 Apr 2015 12:57:22 +0000 (13:57 +0100)]
Merge branch 'ibm-trunk' into support_idl_narrow
Neil GM Richards [Thu, 30 Apr 2015 10:24:01 +0000 (06:24 -0400)]
Merge branch 'improve_cmsfv2' into 'ibm-trunk'
Improve cmsfv2
Add additional guards to CmsfThreadLocal mechanis
See merge request !34
Neil Richards [Thu, 30 Apr 2015 10:20:09 +0000 (11:20 +0100)]
Additional guards to CmsfThreadLocal mechanism
Neil Richards [Thu, 30 Apr 2015 08:20:10 +0000 (09:20 +0100)]
Add support for IDL interfaces to PRO.narrow
Neil GM Richards [Mon, 27 Apr 2015 17:45:57 +0000 (13:45 -0400)]
Merge branch 'avoid_writing_zero_length_chunks' into 'ibm-trunk'
Avoid writing zero length chunks
It seems some other ORBs don't like reading in zero length chunks
(though I think the spec is okay with them being written).
As it's inefficient to write them out to the stream, and to maximise
interoperability, this change stops them from being produced.
See merge request !33
Neil Richards [Mon, 27 Apr 2015 17:41:33 +0000 (18:41 +0100)]
Avoid writing zero length chunks
Neil GM Richards [Mon, 27 Apr 2015 09:37:44 +0000 (05:37 -0400)]
Merge branch 'cmsfv2_setting_check' into 'ibm-trunk'
Cmsfv2 setting check
See merge request !32
Neil Richards [Mon, 27 Apr 2015 09:36:20 +0000 (10:36 +0100)]
Add guard code for unset CMSFv2 thread local
Neil GM Richards [Mon, 27 Apr 2015 09:08:34 +0000 (05:08 -0400)]
Merge branch 'yoko-util' into 'ibm-trunk'
Yoko util
See merge request !31
Neil GM Richards [Mon, 27 Apr 2015 09:03:27 +0000 (05:03 -0400)]
Merge branch 'distinguish_pad_bytes' into 'ibm-trunk'
Distinguish pad bytes
Make padding bytes in the stream easily distinguishable from zero bytes/ints or null objects.
See merge request !30
Neil Richards [Fri, 24 Apr 2015 14:50:08 +0000 (15:50 +0100)]
Move common utilities to yoko-util bundle
Neil Richards [Fri, 24 Apr 2015 12:20:50 +0000 (13:20 +0100)]
Write distinguishable bytes for padding
Neil GM Richards [Thu, 23 Apr 2015 20:55:35 +0000 (16:55 -0400)]
Merge branch 'cmsfv2' into 'ibm-trunk'
Add CMSFv2 support
Add support for Custom Marshaled Stream Format version 2
See merge request !29
Neil Richards [Thu, 23 Apr 2015 20:53:01 +0000 (21:53 +0100)]
Implement CMSFv2 reading support, last part
Neil Richards [Thu, 23 Apr 2015 14:08:55 +0000 (15:08 +0100)]
Merge branch 'ibm-trunk' into cmsfv2
Neil GM Richards [Thu, 23 Apr 2015 13:40:54 +0000 (09:40 -0400)]
Merge branch 'improve_logging' into 'ibm-trunk'
Improve logging
See merge request !28
Neil Richards [Thu, 23 Apr 2015 13:31:50 +0000 (14:31 +0100)]
Improve logging formatting for demarshalling
Neil Richards [Thu, 23 Apr 2015 10:42:59 +0000 (11:42 +0100)]
Don't override readDefaultObject delegation
Neil Richards [Wed, 22 Apr 2015 17:46:57 +0000 (18:46 +0100)]
Merge branch 'ibm-trunk' into cmsfv2
Neil GM Richards [Wed, 22 Apr 2015 17:46:04 +0000 (13:46 -0400)]
Merge branch 'improve_dumpData' into 'ibm-trunk'
Improve dump data
Improve the formatting of Buffer.dumpData and IORUtil.dump_octets, so that it's easier to deal with when debugging.
See merge request !27
Neil Richards [Wed, 22 Apr 2015 17:43:14 +0000 (18:43 +0100)]
Improve format of dataDump and dump_octets
Neil Richards [Wed, 22 Apr 2015 17:36:10 +0000 (18:36 +0100)]
Enable Cmsfv2 reading support, part 1
Joe Chacko [Wed, 22 Apr 2015 09:48:38 +0000 (10:48 +0100)]
Generify offsetMap to use a value type of Object, not Serializable
Neil Richards [Tue, 21 Apr 2015 16:49:17 +0000 (17:49 +0100)]
Formatted source (automated)
Joe Chacko [Tue, 21 Apr 2015 18:04:00 +0000 (19:04 +0100)]
Refactor CorbaObjectReader (automated)
Joe Chacko [Tue, 21 Apr 2015 16:33:54 +0000 (17:33 +0100)]
Refactored ValueReader (automated)
Neil Richards [Mon, 20 Apr 2015 16:00:32 +0000 (17:00 +0100)]
Add test object for CMSFv2
Neil Richards [Mon, 20 Apr 2015 10:58:23 +0000 (11:58 +0100)]
Improve error handling; Clarify imports
Joe Chacko [Sat, 18 Apr 2015 00:43:32 +0000 (01:43 +0100)]
cleaned up ValueDescriptor
Joe Chacko [Fri, 17 Apr 2015 23:10:33 +0000 (00:10 +0100)]
sanitized ObjectWriter to avoid fully qualified classes and other style violations
Joe Chacko [Fri, 17 Apr 2015 22:04:07 +0000 (23:04 +0100)]
converted WriteObjectState to an enum
Neil Richards [Fri, 17 Apr 2015 16:23:41 +0000 (17:23 +0100)]
Fix custom marshal check on reading side
This encapsulates the check in a method called 'isCustomUnmarshalled()'
- this needs a better name.
Neil Richards [Fri, 17 Apr 2015 16:12:27 +0000 (17:12 +0100)]
Merge branch 'ibm-trunk' into cmsfv2
Neil GM Richards [Fri, 17 Apr 2015 15:14:26 +0000 (11:14 -0400)]
Merge branch 'perf_helper_lookup' into 'ibm-trunk'
Avoid BoxedValueHelper lookups where possible
See merge request !26
Neil Richards [Fri, 17 Apr 2015 15:09:22 +0000 (16:09 +0100)]
Merge branch 'improve_unknown_exception' into perf_helper_lookup
Neil Richards [Fri, 17 Apr 2015 10:48:05 +0000 (11:48 +0100)]
Improve information captured in UnknownException
Neil Richards [Wed, 15 Apr 2015 17:30:48 +0000 (18:30 +0100)]
Merge branch 'corbaname_parse_diagnostics' into perf_helper_lookup
Neil Richards [Wed, 15 Apr 2015 17:27:40 +0000 (18:27 +0100)]
Avoid BoxedValueHelper lookup where possible
Neil Richards [Wed, 15 Apr 2015 16:37:40 +0000 (17:37 +0100)]
Add temp. ability to disable CMSFv2 advertising
Neil Richards [Fri, 10 Apr 2015 18:14:31 +0000 (19:14 +0100)]
Add more info to corbaname parsing exceptions
Neil Richards [Fri, 10 Apr 2015 13:45:16 +0000 (14:45 +0100)]
Merge branch 'ibm-trunk' into cmsfv2
Neil GM Richards [Fri, 10 Apr 2015 13:37:17 +0000 (09:37 -0400)]
Merge branch 'enum_fix' into 'ibm-trunk'
Enum fix
Demarshalling of Java Enums need to be special-cased - the demarshalled value needs to be obtained via a call to
Enum.valueOf(type, name) as Enum values are singletons.
See merge request !25
Neil Richards [Fri, 10 Apr 2015 13:29:47 +0000 (14:29 +0100)]
Add Enum demarshalling handling
Neil Richards [Thu, 9 Apr 2015 21:36:00 +0000 (22:36 +0100)]
Correct writing of CMSF TC & SC data
Correct witing of CDR encapuslated data for Custom Marshal Stream Format
tagged component and service context.
Neil Richards [Thu, 9 Apr 2015 11:12:15 +0000 (12:12 +0100)]
Enable outbound CMSFv2 support
Neil Richards [Tue, 7 Apr 2015 00:20:19 +0000 (01:20 +0100)]
Move CmsfThreadLocalStack to yoko-rmi-impl
Move CmsfThreadLocalStack to yoko-rmi-impl, as it's code there that
needs to find out what should be the Custom Marshal Stream Format
version to be used.
Neil Richards [Tue, 7 Apr 2015 00:03:01 +0000 (01:03 +0100)]
Add PI classes for CMSF handling
Add (IOR, Client, Server) Portable Interceptor classes for handling
negotiation of Custom Marshal Stream Formatting version support,
including a ThreadLocal for propagating the agreed version to write
operations.
Neil Richards [Sat, 4 Apr 2015 17:13:46 +0000 (18:13 +0100)]
Add missing API relating to CMSFv2
Add missing API relating to supporting Custom Marshal Stream Format
version 2.
Neil GM Richards [Tue, 31 Mar 2015 10:51:54 +0000 (06:51 -0400)]
Merge branch 'serializable_array_marshalling' into 'ibm-trunk'
Serializable array marshalling
As per the Java2IDL spec, the elements of a Serializable[] need to be marshalled as Anys.
See merge request !24
Neil Richards [Tue, 31 Mar 2015 10:48:18 +0000 (11:48 +0100)]
Send elements of Serializable[] as Anys
This conforms to the Java2IDL spec, 4.3.10.1
Neil GM Richards [Fri, 27 Mar 2015 15:22:37 +0000 (11:22 -0400)]
Merge branch 'stub.loader' into 'ibm-trunk'
Stub.loader
Add entry for ProviderLoader for javax.rmi.CORBA.Stub
See merge request !23
Neil Richards [Fri, 27 Mar 2015 15:20:37 +0000 (15:20 +0000)]
Add ProviderLoader for javax.rmi.CORBA.Stub
Neil GM Richards [Thu, 26 Mar 2015 00:54:00 +0000 (20:54 -0400)]
Merge branch 'remote.stub.loader' into 'ibm-trunk'
Remote.stub.loader
See merge request !22
Neil Richards [Thu, 26 Mar 2015 00:49:59 +0000 (00:49 +0000)]
Add _Remote_Stub to ProviderLoader list
Neil Richards [Wed, 11 Mar 2015 22:51:30 +0000 (22:51 +0000)]
Merge branch 'ibm-trunk' into TypeRepo-with-ClassValue
Joseph Chacko [Wed, 11 Mar 2015 01:32:04 +0000 (21:32 -0400)]
Merge branch 'avoid_RemoteException_initCause' into 'ibm-trunk'
Avoid remote exception init cause
One has to set the 'detail' public field of a RemoteException,
instead of trying to call initCause() (as that always throws IllegalStateException).
Some subclasses of RemoteException, such as NoSuchObjectException, provides no other
way of setting up the causal exception.
See merge request !21
Neil Richards [Tue, 10 Mar 2015 23:19:22 +0000 (23:19 +0000)]
Avoid calling RemoteException.initCause()
... as this will always throw IllegalStateException.
Instead set its public 'detail' field to the desired value.
Joseph Chacko [Tue, 10 Mar 2015 15:46:28 +0000 (11:46 -0400)]
Merge branch 'removeJTADependency' into 'ibm-trunk'
Remove jta dependency
See merge request !20
Joseph Chacko [Tue, 10 Mar 2015 15:40:00 +0000 (11:40 -0400)]
Merge branch 'RMIStub_toString_fix' into 'ibm-trunk'
Rmi stub to string fix
This addresses the BAD_OPERATION exception seen by Andy McCright when running with trace enabled.
See merge request !18
David Jencks [Tue, 10 Mar 2015 13:53:46 +0000 (09:53 -0400)]
Merge branch 'use-stringbuilder-1' into 'ibm-trunk'
Use stringbuilder 1
Use StringBuilder for lengthy string construction, avoid loop errors.
See merge request !19
Neil Richards [Tue, 10 Mar 2015 13:29:11 +0000 (13:29 +0000)]
Produce something sensible from RMIStub.toString()
... when its delegate is not hooked up.
Neil Richards [Tue, 10 Mar 2015 10:29:33 +0000 (10:29 +0000)]
remove hard dependency on JTA