<?xml version="1.0" encoding="utf-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Recent changes to bugs</title><link>https://sourceforge.net/p/ose/bugs/</link><description>Recent changes to bugs</description><atom:link href="https://sourceforge.net/p/ose/bugs/feed.rss" rel="self"/><language>en</language><lastBuildDate>Fri, 16 Nov 2007 16:32:55 -0000</lastBuildDate><atom:link href="https://sourceforge.net/p/ose/bugs/feed.rss" rel="self" type="application/rss+xml"/><item><title>OTC_ServiceBinding incorrectly detects IP vs. hostname</title><link>https://sourceforge.net/p/ose/bugs/81/</link><description>&lt;div class="markdown_content"&gt;&lt;p&gt;In the constructor of OTC_ServiceBinding, the receiver name is passed as a string containing colon-delimited fields. There is no way for OTC_ServiceBinding to know whether the host field is a text-encoded hex IP address or simply the string of the machine's configured hostname. So, it uses the istream "hex" operator and checks the fail bit.&lt;/p&gt;
&lt;p&gt;If the fail bit is set, then the assumption is that this was a hostname string, and not a hex value. Likewise, if the fail bit is not set, it is assumed to be parseable hex. However, because the colon character is parsed out by calling .ignore(1) on the stream, various likely hostnames can be misinterpreted.&lt;/p&gt;
&lt;p&gt;For instance, the hostname "bm32" is interpreted as the hex IP "b" because istream will parse as much as it can without setting the fail bit. Then the "m" is ignored, and "32" is interpreted to be the PID. This has the eventual effect that all generated events will be wrapped in an envelope that incorrectly identifies the source host and PID of the event.&lt;/p&gt;
&lt;p&gt;This will occur whenever the hostname can be interpreted as two text-encoded hex values separated by a non-hex character, and WinSock was not initialized before the first call to OTC_Program::uniqueId().&lt;/p&gt;&lt;/div&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">David Seaman</dc:creator><pubDate>Fri, 16 Nov 2007 16:32:55 -0000</pubDate><guid isPermaLink="false">https://sourceforge.net331e618aeece1104c59891083343c42a3cd88aac</guid></item><item><title>8.0b7 PyThreadState_Delete fatal error.</title><link>https://sourceforge.net/p/ose/bugs/80/</link><description>&lt;div class="markdown_content"&gt;&lt;p&gt;Changes in OSE 8.0b7 broke support for threading in Python.&lt;/p&gt;
&lt;p&gt;One of the errors pointing to problems is:&lt;/p&gt;
&lt;p&gt;Fatal Python error: PyThreadState_Delete: tstate is still current&lt;/p&gt;
&lt;p&gt;This can occur even if you aren't using threads.&lt;/p&gt;
&lt;p&gt;Quick solution is to use OSE 8.0b6 however OSE 8.0b8 will be &lt;br /&gt;
released soon and will fix the problem. Not possible to supply a &lt;br /&gt;
patch as changes had to be made across about 20+ files. :-(&lt;/p&gt;&lt;/div&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Graham Dumpleton</dc:creator><pubDate>Fri, 22 Apr 2005 01:49:02 -0000</pubDate><guid isPermaLink="false">https://sourceforge.netbce43d4a2a2524f8e38e1311642b256547574659</guid></item><item><title>XML-RPC C++ gateway wrongly converting structs.</title><link>https://sourceforge.net/p/ose/bugs/79/</link><description>&lt;div class="markdown_content"&gt;&lt;p&gt;In OSE 7.0 and OSE 8.0, the XML-RPC gateway which is &lt;br /&gt;
implemented in C++ is wrongly converting struct elements when a &lt;br /&gt;
key/member name has an embedded '.' in it.&lt;/p&gt;
&lt;p&gt;For example, for the struct { "a.b": 1 }, the result should be similar, &lt;br /&gt;
but the result is actually { "a": { "b": 1 } }. This is because the wrong &lt;br /&gt;
method was being called and the one which gives special meaning &lt;br /&gt;
to a '.' was being used by mistake.&lt;/p&gt;
&lt;p&gt;Attached patch is for OSE 8.0, but similar patch would be used for &lt;br /&gt;
OSE 7.0.&lt;/p&gt;&lt;/div&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Graham Dumpleton</dc:creator><pubDate>Fri, 04 Mar 2005 10:35:49 -0000</pubDate><guid isPermaLink="false">https://sourceforge.netbe6c0bb821373e78ff798ff9977f662cdb2119fd</guid></item><item><title>undefined reference to `__gxx_personality_v0'</title><link>https://sourceforge.net/p/ose/bugs/78/</link><description>&lt;div class="markdown_content"&gt;&lt;p&gt;When compiling the OSE-8.0 Python libraries on Debian&lt;br /&gt;
(Stable) using python2.2 libraries=["stdc++"] must be&lt;br /&gt;
defined in the setup.py build file.&lt;/p&gt;
&lt;p&gt;If it is not, the module will not import and python&lt;br /&gt;
complains about "undefined reference to&lt;br /&gt;
`__gxx_personality_v0'"&lt;/p&gt;&lt;/div&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Richard Smith</dc:creator><pubDate>Wed, 23 Feb 2005 15:39:36 -0000</pubDate><guid isPermaLink="false">https://sourceforge.net1ddefa9a38dd95068614bfe553b96e43269ae880</guid></item><item><title>Win32 POSIX threads compilation problems.</title><link>https://sourceforge.net/p/ose/bugs/77/</link><description>&lt;div class="markdown_content"&gt;&lt;p&gt;More recent versions of POSIX thread library for Win32 have &lt;br /&gt;
changed the definition of pthread_t and has caused some dodgy &lt;br /&gt;
code in OSE to no longer compile.&lt;/p&gt;
&lt;p&gt;Two files are affected. These are:&lt;/p&gt;
&lt;p&gt;library/OTC/debug/otclogger.cc&lt;br /&gt;
library/OTC/thread/otcmutex.cc&lt;/p&gt;
&lt;p&gt;In "otclogger.cc", comment out the section of code reading:&lt;/p&gt;
&lt;p&gt;#if defined(OSE_WITH_THREADS)&lt;br /&gt;
pthread_t theThread;&lt;br /&gt;
theThread = pthread_self();&lt;br /&gt;
theThreadId = (int)theThread;&lt;br /&gt;
#endif&lt;/p&gt;
&lt;p&gt;A better fix for this will be supplied at a later time.&lt;/p&gt;
&lt;p&gt;For "otcmutex.cc", use the replacement file attached to this bug &lt;br /&gt;
report.&lt;/p&gt;
&lt;p&gt;This probably can feasibly affect any POSIX thread library.&lt;/p&gt;&lt;/div&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Graham Dumpleton</dc:creator><pubDate>Tue, 01 Feb 2005 22:45:01 -0000</pubDate><guid isPermaLink="false">https://sourceforge.neta4d552c58ba9cc8351caca2745a99a904b566e21</guid></item><item><title>Unique IDs not unique under threading.</title><link>https://sourceforge.net/p/ose/bugs/76/</link><description>&lt;div class="markdown_content"&gt;&lt;p&gt;If you are using OTC_Program::uniqueId() directly, or indirectly as &lt;br /&gt;
a consequence of using the messaging system, within a &lt;br /&gt;
multithreaded application, there is a risk that IDs being returned &lt;br /&gt;
are not unique. This is because a counter wasn't being correctly &lt;br /&gt;
incremented and cached within the context of the thread lock.&lt;/p&gt;
&lt;p&gt;A good example of the dangers of this are where you are using the &lt;br /&gt;
conversation ID for a message within the messaging system as a &lt;br /&gt;
key to some cached data which is later used when a response to a &lt;br /&gt;
message is received. If you have multiple threads within the same &lt;br /&gt;
context sending messages and caching the keys in the same table, &lt;br /&gt;
two different threads could get the same conversation ID for &lt;br /&gt;
different messages. This might result in the application erroring &lt;br /&gt;
because the conversation ID had already been used, or it might &lt;br /&gt;
result in overwriting of the data already stored under that key. &lt;br /&gt;
That might then result in matching up reponses with wrong data, or &lt;br /&gt;
of loosing a response.&lt;/p&gt;
&lt;p&gt;This could be quite a serious problem is particular applications, but &lt;br /&gt;
would require quite heavy use of threading under certain &lt;br /&gt;
cirucmstances to trigger it.&lt;/p&gt;
&lt;p&gt;This problem affects both OSE 7.0 and OSE 8.0.&lt;/p&gt;
&lt;p&gt;Files attached are replacements for OTC/system/otcprogram.cc in &lt;br /&gt;
library for respective versions.&lt;/p&gt;&lt;/div&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Graham Dumpleton</dc:creator><pubDate>Thu, 30 Dec 2004 23:59:27 -0000</pubDate><guid isPermaLink="false">https://sourceforge.net27c15f1bf2246dc8b0fbbdc8721f5a5c9e173693</guid></item><item><title>Installation failing for OSE 7.0pl10.</title><link>https://sourceforge.net/p/ose/bugs/75/</link><description>&lt;div class="markdown_content"&gt;&lt;p&gt;OSE 7.0pl10 will not be able to be installed on UNIX systems where &lt;br /&gt;
makeit is being used, due to a trailing space which has been &lt;br /&gt;
errornously added to the variable "OSE_MAINTENANCE_RELEASE" &lt;br /&gt;
in the file "include/makeit/version.mk". The error occurs at the &lt;br /&gt;
point that libraries are being installed into the final destination &lt;br /&gt;
directory.&lt;/p&gt;
&lt;p&gt;Make sure the setting of this variable is:&lt;/p&gt;
&lt;p&gt;OSE_MAINTENANCE_RELEASE := 10&lt;/p&gt;
&lt;p&gt;and that there is no space after the "10".&lt;/p&gt;&lt;/div&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Graham Dumpleton</dc:creator><pubDate>Thu, 30 Dec 2004 23:45:54 -0000</pubDate><guid isPermaLink="false">https://sourceforge.net9a2a9144c770fed6f93139f4b26ea0e139b72120</guid></item><item><title>python setup fails / gcc compiler error</title><link>https://sourceforge.net/p/ose/bugs/74/</link><description>&lt;div class="markdown_content"&gt;&lt;p&gt;HI all&lt;/p&gt;
&lt;p&gt;it seems that gcc is failing at &lt;/p&gt;
&lt;p&gt;OTC_InetAddress::hostname() &lt;/p&gt;
&lt;p&gt;due to a pointer conversion. &lt;/p&gt;
&lt;p&gt;Im using python2.3 and gcc3.3.4 under debian/sarge.&lt;/p&gt;
&lt;p&gt;Thanks&lt;/p&gt;
&lt;p&gt;Henrique &lt;br /&gt;
henrique@bucher.com&lt;/p&gt;
&lt;p&gt;gcc -pthread -fno-strict-aliasing -DNDEBUG -g -O3 -&lt;br /&gt;
Wall -Wstrict-prototypes -fPIC -&lt;br /&gt;
DOSE_WITH_THREADS=1 -Iinclude -&lt;br /&gt;
I/usr/include/python2.3 -c &lt;br /&gt;
library/OTC/message/otcinetaddr.cc -o build/temp.linux-&lt;br /&gt;
i686-2.3/library/OTC/message/otcinetaddr.o&lt;br /&gt;
In file included &lt;br /&gt;
from /usr/include/c++/3.3/backward/new.h:33,&lt;br /&gt;
from include/OTC/debug/newhndlr.hh:22,&lt;br /&gt;
from include/OTC/OTC.h:73,&lt;br /&gt;
from include/OTC/memory/pool.hh:11,&lt;br /&gt;
from include/OTC/memory/cmmnpool.hh:11,&lt;br /&gt;
from include/OTC/memory/mpobject.hh:11,&lt;br /&gt;
from include/OTC/text/rstring.hh:22,&lt;br /&gt;
from include/OTC/text/string.hh:22,&lt;br /&gt;
from include/OTC/message/inetaddr.hh:22,&lt;br /&gt;
from &lt;br /&gt;
library/OTC/message/otcinetaddr.cc:26:&lt;br /&gt;
/usr/include/c++/3.3/backward/backward_warning.h:32:2&lt;br /&gt;
: warning: #warning This file includes at least one &lt;br /&gt;
deprecated or antiquated header. Please consider using &lt;br /&gt;
one of the 32 headers found in section 17.4.1.2 of the &lt;br /&gt;
C++ standard. Examples include substituting the &amp;lt;X&amp;gt; &lt;br /&gt;
header for the &amp;lt;X.h&amp;gt; header for C++ includes, or &lt;br /&gt;
&amp;lt;sstream&amp;gt; instead of the deprecated header &lt;br /&gt;
&amp;lt;strstream.h&amp;gt;. To disable this warning use -Wno-&lt;br /&gt;
deprecated.&lt;br /&gt;
library/OTC/message/otcinetaddr.cc: In member function &lt;br /&gt;
`OTC_String &lt;br /&gt;
OTC_InetAddress::hostname() const':&lt;br /&gt;
library/OTC/message/otcinetaddr.cc:111: error: cannot &lt;br /&gt;
convert `int*' to `&lt;br /&gt;
hostent**' for argument `7' to `int gethostbyaddr_r&lt;br /&gt;
(const void*, unsigned &lt;br /&gt;
int, int, hostent*, char*, unsigned int, hostent**, &lt;br /&gt;
int*)'&lt;br /&gt;
library/OTC/message/otcinetaddr.cc: In member function &lt;br /&gt;
`void &lt;br /&gt;
OTC_InetAddress::initAddress(const char*, int)':&lt;br /&gt;
library/OTC/message/otcinetaddr.cc:223: error: cannot &lt;br /&gt;
convert `int*' to `&lt;br /&gt;
hostent**' for argument `5' to `int gethostbyname_r&lt;br /&gt;
(const char*, hostent*, &lt;br /&gt;
char*, unsigned int, hostent**, int*)'&lt;br /&gt;
/usr/include/netdb.h: In member function `void &lt;br /&gt;
OTC_InetAddress::initAddress(const char*, const &lt;br /&gt;
char*, const char*)':&lt;br /&gt;
/usr/include/netdb.h:311: error: too few arguments to &lt;br /&gt;
function `int &lt;br /&gt;
getservbyname_r(const char*, const char*, servent*, &lt;br /&gt;
char*, unsigned int, &lt;br /&gt;
servent**)'&lt;br /&gt;
library/OTC/message/otcinetaddr.cc:248: error: at this &lt;br /&gt;
point in file&lt;br /&gt;
library/OTC/message/otcinetaddr.cc:248: error: invalid &lt;br /&gt;
conversion from `int' to &lt;br /&gt;
`servent*'&lt;br /&gt;
error: command 'gcc' failed with exit status 1&lt;/p&gt;&lt;/div&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Anonymous</dc:creator><pubDate>Sun, 10 Oct 2004 12:51:48 -0000</pubDate><guid isPermaLink="false">https://sourceforge.net9a9f70f5464baac0600182a06686356296494be0</guid></item><item><title>Python FileServer directory redirection.</title><link>https://sourceforge.net/p/ose/bugs/73/</link><description>&lt;div class="markdown_content"&gt;&lt;p&gt;When access is made to an address such as:&lt;/p&gt;
&lt;p&gt;&lt;a href="http://www.example.com/sample" rel="nofollow"&gt;http://www.example.com/sample&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;and "sample" refers to a directory, the web client should have been &lt;br /&gt;
redirected by Python FileServer instance to:&lt;/p&gt;
&lt;p&gt;&lt;a href="http://www.example.com/sample/" rel="nofollow"&gt;http://www.example.com/sample/&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;This should be done to ensure that if an index page is attached to &lt;br /&gt;
the URL that the web client correctly resolves relative URLs for &lt;br /&gt;
images and style sheets.&lt;/p&gt;
&lt;p&gt;For example, if the HTML page returned referred to a style sheet &lt;br /&gt;
through the relative path "example.css", the web client needs to &lt;br /&gt;
access it as:&lt;/p&gt;
&lt;p&gt;&lt;a href="http://www.example.com/sample/example.css" rel="nofollow"&gt;http://www.example.com/sample/example.css&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;and not:&lt;/p&gt;
&lt;p&gt;&lt;a href="http://www.example.com/example.css" rel="nofollow"&gt;http://www.example.com/example.css&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;The bug was that for any directory which wasn't the root of the &lt;br /&gt;
web site, this path redirection wasn't occuring correctly and thus &lt;br /&gt;
any relative file loaded immediately into the page was being &lt;br /&gt;
sourced from the parent directory.&lt;/p&gt;
&lt;p&gt;This problem would affect OSE 8.0b3, OSE 7.0pl9 and earlier &lt;br /&gt;
versions. The attached diff is generated from working version of &lt;br /&gt;
OSE 8.0b4, and the line numbers are probably wrong for all prior &lt;br /&gt;
versions. Apply changes visually.&lt;/p&gt;&lt;/div&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Graham Dumpleton</dc:creator><pubDate>Sun, 18 Jul 2004 04:56:40 -0000</pubDate><guid isPermaLink="false">https://sourceforge.net57e8ec50520dbc9004842ea2b3f9c5ede7135c71</guid></item><item><title>Event dispatcher prematurely exiting.</title><link>https://sourceforge.net/p/ose/bugs/72/</link><description>&lt;div class="markdown_content"&gt;&lt;p&gt;When thread support is built into the C++ class library the event &lt;br /&gt;
dispatcher can prematurely exit for no apparent reason resulting in &lt;br /&gt;
the application terminating.&lt;/p&gt;
&lt;p&gt;This was being caused by a thread specific key context not being &lt;br /&gt;
initialised properly prior to use. On some platforms this would &lt;br /&gt;
result in some random piece of memory being returned instead of &lt;br /&gt;
a null pointer. When that piece of memory was eventually set to a &lt;br /&gt;
particular value, the dispatcher would think that the thread for &lt;br /&gt;
dispatcher had been shutdown and thus exit.&lt;/p&gt;
&lt;p&gt;Replacement for the file "library/OTC/thread/otctask.cc" is &lt;br /&gt;
attached.&lt;/p&gt;
&lt;p&gt;Note this is for OSE 8.0 only. There is no comparable file in prior &lt;br /&gt;
versions of OSE.&lt;/p&gt;&lt;/div&gt;</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Graham Dumpleton</dc:creator><pubDate>Sat, 17 Jul 2004 10:20:24 -0000</pubDate><guid isPermaLink="false">https://sourceforge.net535792f698884220c784c7b25c4aba32bb59205d</guid></item></channel></rss>