summaryrefslogtreecommitdiff
path: root/ChangeLog
diff options
context:
space:
mode:
authorschmidt <douglascraigschmidt@users.noreply.github.com>2001-11-23 20:27:57 +0000
committerschmidt <douglascraigschmidt@users.noreply.github.com>2001-11-23 20:27:57 +0000
commitd0a4235b6758b21421052fc34d70dba7f4a69630 (patch)
tree636c07fb1882d30f821019dbf692a5a05e83fb24 /ChangeLog
parent2e02e461ffcce244844da8bf4233022ddb89d1d7 (diff)
downloadATCD-d0a4235b6758b21421052fc34d70dba7f4a69630.tar.gz
ChangeLogTag:Fri Nov 23 08:40:33 2001 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu>
Diffstat (limited to 'ChangeLog')
-rw-r--r--ChangeLog57
1 files changed, 28 insertions, 29 deletions
diff --git a/ChangeLog b/ChangeLog
index 9706a4ed982..e0f61b3ba58 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -15,34 +15,34 @@ Thu Nov 22 05:24:51 2001 Craig Rodrigues <crodrigu@bbn.com>
Wed Nov 21 11:35:50 2001 Michael Kircher <Michael.Kircher@mchp.siemens.de>
- * ace/Bound_Ptr.h
- * ace/Bound_Ptr.i
- * ace/Future.cpp
- * ace/Future.h
- * ace/Refcounted_Auto_Ptr.h
- * ace/Refcounted_Auto_Ptr.i
-
- Made the memory allocation failure behavior consistent
- to the behavior prior to the change on Sun Oct 28 13:15:00 2001.
- Thanks to Christopher Kohlhoff <chris@kohlhoff.com> who prepared
- this change.
- The problem was that with the above mentioned change failed
- allocations would return 0 in either case, if exceptions were
- enabled by the platform or not. This was incosistent to the
- behavior before, as new threw bad_alloc on allocation failure
- when exceptions were supported by the platform.
-
- To fix this we faced the following forces:
- - Not to introduce yet another macro
- - Minimize impact of changes, localize the problem
- - Keep the new calls wrapped by a ACE_NEW macro to allow
- users to eventually redefine the macro (using e.g. new (nothrow)
-
- As the ACE_NEW macros are quite restrictive in their supported
- semantics, e.g. you can only return on error, we decided to introduce
- a new method (as inline) internal_create which does the allocation.
- internal_create is called by create which does the differentiation
- between the exception-supported and exception-not-supported case.
+ * ace/Bound_Ptr.h
+ * ace/Bound_Ptr.i
+ * ace/Future.cpp
+ * ace/Future.h
+ * ace/Refcounted_Auto_Ptr.h
+ * ace/Refcounted_Auto_Ptr.i
+
+ Made the memory allocation failure behavior consistent
+ to the behavior prior to the change on Sun Oct 28 13:15:00 2001.
+ Thanks to Christopher Kohlhoff <chris@kohlhoff.com> who prepared
+ this change.
+ The problem was that with the above mentioned change failed
+ allocations would return 0 in either case, if exceptions were
+ enabled by the platform or not. This was incosistent to the
+ behavior before, as new threw bad_alloc on allocation failure
+ when exceptions were supported by the platform.
+
+ To fix this we faced the following forces:
+ - Not to introduce yet another macro
+ - Minimize impact of changes, localize the problem
+ - Keep the new calls wrapped by a ACE_NEW macro to allow
+ users to eventually redefine the macro (using e.g. new (nothrow)
+
+ As the ACE_NEW macros are quite restrictive in their supported
+ semantics, e.g. you can only return on error, we decided to introduce
+ a new method (as inline) internal_create which does the allocation.
+ internal_create is called by create which does the differentiation
+ between the exception-supported and exception-not-supported case.
Wed Nov 21 09:07:37 2001 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu>
@@ -95,7 +95,6 @@ Fri Nov 16 10:39:52 2001 Douglas C. Schmidt <schmidt@macarena.cs.wustl.edu>
not stopped. Thanks to Kyle Brost <Kyle.Brost@quest.com> for
providing this fix.
->>>>>>> 4.2766
Wed Nov 21 10:00:20 2001 Chad Elliott <elliott_c@ociweb.com>
* ace/config-irix6.x-common.h: