<?xml version="1.0" encoding="utf-8"?>
<feed xml:lang="en" xmlns="http://www.w3.org/2005/Atom"><title>Recent changes to patches</title><link href="https://sourceforge.net/p/ose/patches/" rel="alternate"/><link href="https://sourceforge.net/p/ose/patches/feed.atom" rel="self"/><id>https://sourceforge.net/p/ose/patches/</id><updated>2003-02-25T10:51:04Z</updated><subtitle>Recent changes to patches</subtitle><entry><title>Ordered key/item pairs in OTC_TManyMap.</title><link href="https://sourceforge.net/p/ose/patches/10/" rel="alternate"/><published>2003-02-25T10:51:04Z</published><updated>2003-02-25T10:51:04Z</updated><author><name>Graham Dumpleton</name><uri>https://sourceforge.net/u/grahamd/</uri></author><id>https://sourceforge.netb94312142bb4e4b361d92a1fd4529879ad665c16</id><summary type="html">&lt;div class="markdown_content"&gt;&lt;p&gt;Attached replacement for &amp;amp;quot;include/OTC/collctn/tmanymap.c&amp;amp;quot; ensures that key/item pairs are kept ordered in underlying list. This provides guarantee that when iterating over key/item pairs in forward order they are visited in ascending order according to rank function in use. Previously order across key sets could be random.&lt;/p&gt;&lt;/div&gt;</summary></entry><entry><title>Shutdown on dispatcher restart.</title><link href="https://sourceforge.net/p/ose/patches/9/" rel="alternate"/><published>2002-11-25T11:19:44Z</published><updated>2002-11-25T11:19:44Z</updated><author><name>Graham Dumpleton</name><uri>https://sourceforge.net/u/grahamd/</uri></author><id>https://sourceforge.net34b27903e3219673f11fe8a8266d535e87b34be5</id><summary type="html">&lt;div class="markdown_content"&gt;&lt;p&gt;Attached patch allows the orderly shutdown mechanism&lt;br /&gt;
implemented by OTCEV_Shutdown to be used more than once&lt;br /&gt;
in the case where the dispatcher is shutdown but then&lt;br /&gt;
subsequently reset and restarted. Previously, using the&lt;br /&gt;
OTCEV_Shutdown class a second time would result in it&lt;br /&gt;
forcibly shutting down the dispatcher rather than doing&lt;br /&gt;
an orderly shutdown.&lt;/p&gt;&lt;/div&gt;</summary></entry><entry><title>Jar support for makeit java module.</title><link href="https://sourceforge.net/p/ose/patches/8/" rel="alternate"/><published>2002-11-25T10:21:54Z</published><updated>2002-11-25T10:21:54Z</updated><author><name>Graham Dumpleton</name><uri>https://sourceforge.net/u/grahamd/</uri></author><id>https://sourceforge.net78604e7130af1b158c832d83283889dfd1d73761</id><summary type="html">&lt;div class="markdown_content"&gt;&lt;p&gt;Java support in makeit is pretty bad, but here at least&lt;br /&gt;
is a little improvement. In the attached files you will&lt;br /&gt;
find jar support for Java. At the moment it will only&lt;br /&gt;
allow .class files to be put in the jar file. If you&lt;br /&gt;
want to give an entry point you will need to provide a&lt;br /&gt;
manifest additions file and list it in the MANIFEST&lt;br /&gt;
variable in your makefile. Oh, you also need to add&lt;br /&gt;
&amp;amp;quot;jar&amp;amp;quot; in the JAVA_OPTIONS variable.&lt;/p&gt;
&lt;p&gt;Going by old comments on a Java mailing lists elsewhere&lt;br /&gt;
on the net, at least one person has rewritten the java&lt;br /&gt;
module to work a lot better. It is a pity though that&lt;br /&gt;
they never chose to donate it back to be included in&lt;br /&gt;
OSE. :-(&lt;/p&gt;&lt;/div&gt;</summary></entry><entry><title>Intel C++ makefile changes.</title><link href="https://sourceforge.net/p/ose/patches/7/" rel="alternate"/><published>2002-10-04T05:21:57Z</published><updated>2002-10-04T05:21:57Z</updated><author><name>Graham Dumpleton</name><uri>https://sourceforge.net/u/grahamd/</uri></author><id>https://sourceforge.net2aa711e44625e13929b25302e71335f52cfbc9e9</id><summary type="html">&lt;div class="markdown_content"&gt;&lt;p&gt;Attached file includes changes for&lt;br /&gt;
include/makeit/compiler.mk and host config file for&lt;br /&gt;
using Intel C++ compiler.&lt;/p&gt;
&lt;p&gt;Although one can build the OSE C++ class libraries, the&lt;br /&gt;
compiler is unable to link a program which uses them.&lt;br /&gt;
This appears to be a compiler problem.&lt;/p&gt;&lt;/div&gt;</summary></entry><entry><title>Class2man/Gawk 3.X Warnings.</title><link href="https://sourceforge.net/p/ose/patches/6/" rel="alternate"/><published>2002-10-04T05:17:10Z</published><updated>2002-10-04T05:17:10Z</updated><author><name>Graham Dumpleton</name><uri>https://sourceforge.net/u/grahamd/</uri></author><id>https://sourceforge.net59c27d57051cd2b9fed41a6034f8a30d54dce386</id><summary type="html">&lt;div class="markdown_content"&gt;&lt;p&gt;Attached files include patches to class2man awk scripts&lt;br /&gt;
to eliminate warnings when using GAWK 3.1.&lt;/p&gt;&lt;/div&gt;</summary></entry><entry><title>Service Reports and Failures</title><link href="https://sourceforge.net/p/ose/patches/5/" rel="alternate"/><published>2002-09-10T01:12:58Z</published><updated>2002-09-10T01:12:58Z</updated><author><name>Graham Dumpleton</name><uri>https://sourceforge.net/u/grahamd/</uri></author><id>https://sourceforge.net38cbc73e0ea555a27b12aa89e9d5de9bdded03a3</id><summary type="html">&lt;div class="markdown_content"&gt;&lt;p&gt;Following is a bunch of patches will be going into next&lt;br /&gt;
version. These relate to unsubscribing to all reports&lt;br /&gt;
from a particular service in one go and changes to the&lt;br /&gt;
failure event object to allow generation of these&lt;br /&gt;
events independent of actual message system messages.&lt;/p&gt;
&lt;p&gt;This is all obscure stuff and is only going up here as&lt;br /&gt;
it was generated away from where OSE stuff is kept and&lt;br /&gt;
I needed a way to transfer it back. There is no need&lt;br /&gt;
for anyone to actually apply this. The changes are more&lt;br /&gt;
to do with laying a foundation for more significant&lt;br /&gt;
feature additions later on.&lt;/p&gt;&lt;/div&gt;</summary></entry><entry><title>Changes for GNU C++ 3.X.</title><link href="https://sourceforge.net/p/ose/patches/4/" rel="alternate"/><published>2002-08-07T04:32:11Z</published><updated>2002-08-07T04:32:11Z</updated><author><name>Graham Dumpleton</name><uri>https://sourceforge.net/u/grahamd/</uri></author><id>https://sourceforge.net10e07dbf1615cf26f6a4bc8d5ea5aee4845eb15a</id><summary type="html">&lt;div class="markdown_content"&gt;&lt;p&gt;Following needs to be inserted into&lt;br /&gt;
&amp;amp;quot;include/makeit/compiler.mk&amp;amp;quot; in order to be able to use&lt;br /&gt;
GNU C++ 3.X.&lt;/p&gt;
&lt;p&gt;# GNU C++ Version 3.X.&lt;/p&gt;
&lt;p&gt;ifneq &amp;amp;quot;$(filter GNU3.%,$(C++COMPILER))&amp;amp;quot; &amp;amp;quot;&amp;amp;quot;&lt;/p&gt;
&lt;p&gt;_makeit_cxx_valid_compiler_tag := YES&lt;/p&gt;
&lt;p&gt;_makeit_cxx_supports_o_for_objects := YES&lt;br /&gt;
_makeit_cxx_supports_template_file := YES&lt;br /&gt;
_makeit_cxx_supports_ansi_exceptions := YES&lt;/p&gt;
&lt;p&gt;_makeit_cxx_supports_precompiled_templates := YES&lt;br /&gt;
override INTERNAL_C++FLAGS += -DEXPAND_TEMPLATES&lt;/p&gt;
&lt;p&gt;ifneq &amp;amp;quot;$(filter %MACOSX,$(OSE_HOST))&amp;amp;quot; &amp;amp;quot;&amp;amp;quot;&lt;br /&gt;
_makeit_cxx_supports_precompiled_templates := NO&lt;br /&gt;
endif&lt;/p&gt;
&lt;p&gt;override INTERNAL_C++FLAGS += -DENV_CXXSTDLIB&lt;/p&gt;
&lt;p&gt;override INTERNAL_C++FLAGS += -DCXX_GNU&lt;/p&gt;
&lt;p&gt;override INTERNAL_C++FLAGS += -DCXX_$(subst&lt;br /&gt;
.,_,$(C++COMPILER))&lt;/p&gt;
&lt;p&gt;ifeq &amp;amp;quot;$(filter no_compiler_warnings,$(MAKEIT_OPTIONS))&amp;amp;quot; &amp;amp;quot;&amp;amp;quot;&lt;br /&gt;
override INTERNAL_C++FLAGS += -Wall&lt;br /&gt;
endif&lt;/p&gt;
&lt;p&gt;endif&lt;/p&gt;
&lt;p&gt;Note that you probably also want to include&lt;br /&gt;
&amp;amp;quot;-Wno-deprecated&amp;amp;quot; as a C++ compiler flags as well. Ie.,&lt;/p&gt;
&lt;p&gt;GNU3.1_C++FLAGS = -pipe -Winline -Wno-deprecated&lt;/p&gt;
&lt;p&gt;The compiler will still complain a lot about the&lt;br /&gt;
#pragma's related to template stuff, but need to change&lt;br /&gt;
header files to get rid of that. Intention is to&lt;br /&gt;
conditionally exclude that for versions of GNU C++ from&lt;br /&gt;
version 3 onwards.&lt;/p&gt;&lt;/div&gt;</summary></entry><entry><title>LD_RUN_PATH and Solaris.</title><link href="https://sourceforge.net/p/ose/patches/3/" rel="alternate"/><published>2002-08-05T07:35:53Z</published><updated>2002-08-05T07:35:53Z</updated><author><name>Graham Dumpleton</name><uri>https://sourceforge.net/u/grahamd/</uri></author><id>https://sourceforge.net49d1ff1dcb5e954405c4561651ba7d0113525d0e</id><summary type="html">&lt;div class="markdown_content"&gt;&lt;p&gt;In the very obscure situation whereby GNU make might be&lt;br /&gt;
used to trigger execution of &amp;amp;quot;makeit&amp;amp;quot; in a&lt;br /&gt;
subdirectory, LD_RUN_PATH doesn't get set properly or&lt;br /&gt;
exported into the environment by makeit.&lt;/p&gt;
&lt;p&gt;A consequence of this is that on the Solaris platform&lt;br /&gt;
when a shared library is created, the first object file&lt;br /&gt;
in the directory can be left out of the shared library&lt;br /&gt;
and the directory search path encoded into the shared&lt;br /&gt;
library is actually the name of that first object file.&lt;/p&gt;
&lt;p&gt;Attached is a replacement for the file&lt;br /&gt;
&amp;amp;quot;include/makeit/init.mk&amp;amp;quot; which fixes the problem.&lt;/p&gt;&lt;/div&gt;</summary></entry><entry><title>Subscribing to specific service agent.</title><link href="https://sourceforge.net/p/ose/patches/2/" rel="alternate"/><published>2002-08-02T04:28:06Z</published><updated>2002-08-02T04:28:06Z</updated><author><name>Graham Dumpleton</name><uri>https://sourceforge.net/u/grahamd/</uri></author><id>https://sourceforge.nete217e7483aa2e49a049feed84be4437cb0a5452a</id><summary type="html">&lt;div class="markdown_content"&gt;&lt;p&gt;Attached files change the OTC_SVBroker class to allow&lt;br /&gt;
reports published by a specific service agent to be&lt;br /&gt;
subscribed to. Up till now you could only subscribe&lt;br /&gt;
based on service name, for which there might be&lt;br /&gt;
multiple corresponding service agent instances.&lt;/p&gt;
&lt;p&gt;These changes should appear in OSE 7.0pl4, although&lt;br /&gt;
they need to be tested further at this point to make&lt;br /&gt;
sure they don't break existing functionality.&lt;/p&gt;
&lt;p&gt;Note that changes would still need to be made to the&lt;br /&gt;
corresponding Python interfaces to expose the same&lt;br /&gt;
functionality.&lt;/p&gt;&lt;/div&gt;</summary></entry><entry><title>Root installation directory.</title><link href="https://sourceforge.net/p/ose/patches/1/" rel="alternate"/><published>2002-08-02T04:23:23Z</published><updated>2002-08-02T04:23:23Z</updated><author><name>Graham Dumpleton</name><uri>https://sourceforge.net/u/grahamd/</uri></author><id>https://sourceforge.net7dae9492a747fb54633d6f8c9908b549950f79eb</id><summary type="html">&lt;div class="markdown_content"&gt;&lt;p&gt;Changes in attached files allow OSE to be installed&lt;br /&gt;
into a directory different to where it will finally be&lt;br /&gt;
used out of. This is useful where you need to install&lt;br /&gt;
it somewhere first for the purposes of archiving it&lt;br /&gt;
into a tar ball or some other archive used by a&lt;br /&gt;
particular packaging system. The package would then be&lt;br /&gt;
distributed for installation in the real location from&lt;br /&gt;
which it will be used.&lt;/p&gt;
&lt;p&gt;To make use of these changes, set OSE_BUILD_ROOT when&lt;br /&gt;
running make to install OSE. The pathname defined in&lt;br /&gt;
OSE_BUILD_ROOT must be an absolute path. Ie.,&lt;/p&gt;
&lt;p&gt;gnumake OSE_BUILD_ROOT=/some/dir install&lt;/p&gt;
&lt;p&gt;When this is done, the value of OSE_ROOT_DIRECTORY&lt;br /&gt;
defined in the host config file, or the default of&lt;br /&gt;
/usr/local/ose, will still be the path which is encoded&lt;br /&gt;
into scripts, but OSE itself will instead be installed&lt;br /&gt;
under the path defined in OSE_BUILD_ROOT.&lt;/p&gt;
&lt;p&gt;The files, relative to the source code root, which are&lt;br /&gt;
changed are:&lt;/p&gt;
&lt;p&gt;include/makeit/version.mk&lt;br /&gt;
Makefile&lt;/p&gt;
&lt;p&gt;This feature will be added in OSE 7.0pl4.&lt;/p&gt;&lt;/div&gt;</summary></entry></feed>