Progress OpenEdge Release Notes Release: 11.3.3.045 Date: January 2018 -------------------------------------------------------------------------------- Copyright (c) 2016 Progress Software Corporation. All rights reserved. This Hotfix is fully supported by Progress Software. While it has been QA-tested to verify that it resolves the issue(s) that it is intended to address, OpenEdge developers are strongly advised to carry out their own application testing to verify functionality and to check for any unexpected behavior changes. Any issues with this Hotfix should be reported immediately to Progress Technical Support. An upgrade to the next Service Pack for the release associated with this Hotfix (if applicable) is recommended as soon as it is available. ======== CONTENTS ======== 1. Special Note 2. Installation Instructions 3. List of issues fixed in 11.3.3.045 4. List of issues fixed in previous 11.3.3.x releases ----------------------------------------------------- 1. Special Note ----------------------------------------------------- Issue:PSC00333559 There was a vulnerability in the AVM where the internal DTD of an XML document might point to a well known or common location of a sensitive file for extracting to the caller. For example, ]> &xxe90539; As of 11.3.3.015, 11.4.0.020 and 11.5.1.003, the AVM can now prohibit this extraction unless SCHEMA-PATH allows it. This applies to both the DOM and SAX parsers: :LOAD( ), :SAX-PARSE( ) and :SAX-PARSE-FIRST( ). In hotfix 11.3.3.016, this vulnerability is addressed in the WEB-CONTEXT:X-DOCUMENT attribute. Using the following DOM example, where the above XML document is xxe.xml: DEFINE VARIABLE hx AS HANDLE. CREATE X-DOCUMENT hx. hx:LOAD( "file", "xxe.xml", FALSE ). This will produce the following error: X-NODEREF or X-DOCUMENT LOAD got an error: FATAL ERROR: file 'C:\temp/xxe.xml', line '3', column '0', message 'unable to open external entity 'file:///c:/windows/win.ini''. (9082) This strict entity resolution behaviour is disabled by default. To turn on this behavior, you can specify -strictEntityResolution 1 on the command-line. In hotfix 11.3.3.016, this startup parameter also applies to WEB-CONTEXT:X-DOCUMENT. Alternatively, there are new attributes :STRICT-ENTITY-RESOLUTION and :STRICT-ENTITY-RESOLUTION. Setting the value of the attributes to TRUE before parsing the XML document will enable this behaviour. There are no attributes to govern the parsing of WEB-CONTEXT:X-DOCUMENT, so you must rely on -strictEntityResolution 1 for WebSpeed. There was also an issue with XML ENTITY expansion. As of 11.3.3.015, 11.4.0.020 and 11.5.1.003, the AVM now has a default limit of 50000 expansions for :LOAD(), :SAX-PARSE() and :SAX-PARSE-FIRST(). If you wish to set an alternative limit, you can set the :ENTITY-EXPANSION-LIMIT and :ENTITY-EXPANSION-LIMIT attributes before parsing the XML document. In hotfix 11.3.3.016, WEB-CONTEXT:X-DOCUMENT also has the default 50000 entity expansion limit. Hotfix 11.3.3.016 also introduces a startup parameter -entityExpansionLimit to change the default expansion limit from 50000. Issue PSC00322883 --------------------- This fix does contain the compile time changes hence You need to recompile your application after applying the patch. This exercise does not apply for those who does not hit this issue. Issue PSC00299051 --------------------- New Database parameter: -TXERetryLimit A new startup parameter to specify the number of TXE Lock conflict retries before queuing and waiting on the connection.s semaphore has been provided and is named TXERetryLimit. This database-wide parameter takes a positive value between 0 and 100,000. When a .TXE Lock conflict. is encountered, the connection will nap for 1 millisecond and re-try the TXE lock request for up to TXERetryLimit times. If after that many times the TXE lock request is still in conflict, the connection will queue, waiting on its semaphore, just as it did in the past. The default value for TXERetryLimit is 0 (retries disabled). A new message will be displayed in the database .lg file when the database is started in multi-user mode reporting the value of the new startup parameter: (17717) TXE Lock retry limit (-TXERetryLimit): 200. The promon utility provides the ability to monitor the affects of this new parameter as well as change the value while the database is online. To see the effect of the new value in promon you must: 1. R&D. Advanced options 2. debghb, 6 3. 9. TXE Lock Activity The new line .TXE Lock Retries. in the output will display the total number of times a retry was attempted as well as the average number of times a retry was attempted per lock request. A lower average value indicates that more locks were granted without encountering a conflict. The ability to change the number of retries before queuing while the database is running is provided through the promon utility as well. To change the value you must: 1. R&D. Advanced options 2. debghb, 6 3. 10. Adjust TXE Options 4. 4. TXE lock retry limit Option 4 will request Enter TXE lock retry limit (0 to 100000): Tuning -TXERetryLimit The main thing that should be of concern to OLTP deployments is throughput of database activity. This is easily monitored through activity summary screen in promon. It will display the number and rate of data changes and transaction commits. To improve throughput at the database level (as opposed to making application changes) things that are obstructing throughput need to be identified and associated parameters adjusted. This is considered improving concurrency and throughput by avoiding conflict. Since this enhancement deals with improving conflict processing, any performance improvement gained should only be in a multi-user environment. Identifying if there is TXE lock conflict that is impeding throughput can be seen on the TXE Activity screen. If there are waits occurring, increasing the TXERetryLimit on the .Adjust TXE Options. screen can help avoid this. Realize that this will put additional pressure on the TXQ latch and a portion of the increased CPU utilization. Additionally there may be increased pressure on the MTX latch and recovery subsystem in general if more recovery notes are being recorded due to the increased throughput. However, the additional utilization is warranted if the change in the activity screen monitored after increasing the TXERetryLimit shows improved throughput (since that is after all what is being targeted here). If throughput is improved then the change was a good one and the increased contention for the TXQ latch and additional CPU utilization is (comparatively) warranted. However, if throughput is NOT improved (for the same workload), then increasing the TXERetryLimit does not warrant an increase in the TXQ latch contention or increased CPU utilization and should therefore be lowered. If changing the TXERetryLimit does not improve throughput as expected and there are still TXE waits, tuning the .grant. and .skip. limits should be investigated based on the activity being performed. Note that the TXERetryLimit only affects the TXE lock mechanism for database updates and will have no affect on improving the performance of .read. type database operations. Testing with 200 concurrent users performing insert and delete operations showed that on average there were less than one spin per TXE lock granted (0.8 spins per TXE lock) and the number of connection waiting in the queue was less than 1 per second when running with the setting of 200 retries. At lower numbers such as 4 concurrent insert operations and 4 concurrent delete operations, the TXERetryLimit 0 worked better than the TXERetryLimit of 200 on some platforms. Therefore this feature will not be enabled by default. ----------------------------------------------------- 2. Installation Instructions ----------------------------------------------------- 1. Make sure you shut down any instances of OpenEdge that you have running before installing this HOTFIX. If you do not, when you run the HOTFIX installation, you might receive messages indicating that some files could not be copied on top of other files (because they are locked/in use by another process). Check with your system administrator if you do not know how to shut down OpenEdge. 2. Backup the following files/directories in your current install before unzipping the patch. After you are satisfied with the patch installation, you may remove these backup files at any time. copy "bin" to "bin.bak" copy "lib" to "lib.bak" copy "oebuild" to "oebuild.bak" copy "java" to "java.bak" copy "odbc" to "odbc.bak" copy "esbadapter" to "esbadapter.bak" copy "servlets" to "servlets.bak" 3. Download the patch file and untar/extract it into the directory where you installed[RELEASE_VERSION][SP]. This directory defaults to $DLC. For Example: UNIX : [PATCHID].aix64.tar.Z Windows: [PATCHID].nt.zip This will unpack new versions of numerous files and install them into the appropriate locations. You will be prompted if you want to overwrite the files, so you may want to make sure step 2 above was followed before overwriting them. 4. Change the owner of executables in the bin directory as follow: su cd ./bin chown root _dbutil _mprosrv _mprshut prodb _progres _proutil _rfutil prodel _amspriv _dbagent _proapsv _probrkr xsdto4gl _ora* ora* chgrp _dbutil _mprosrv _mprshut prodb _progres _proutil _rfutil prodel _amspriv _dbagent _proapsv _probrkr xsdto4gl _ora* ora* chmod 755 _dbutil _mprosrv _mprshut prodb _progres _proutil _rfutil prodel _amspriv _dbagent _proapsv _probrkr xsdto4gl _ora* ora* chmod +s _dbutil _mprosrv _mprshut prodb _progres _proutil _rfutil prodel _amspriv _dbagent _proapsv _probrkr xsdto4gl _ora* ora* Please Note: If OpenEdge DB replication is installed run the above commands for the rpserver and rpagent executables in $DLC/bin also. 5. For AIX platforms, it is necessary to change the embedded library search path for some of the shared objects, only if these shared objects have been included in this hotfix. It is recommended to run slibclean on AIX both before unpacking the hotfix tar file and after running the rtl_enable commands. The rtl_enable command will change the file size of the library file from what was in the source tar file. File modified dates will also be changed. These indicate that the command was successful. Please use the following commands to do this: For 32-bit OpenEdge AIX: cd ${DLC}/lib /usr/bin/rtl_enable -R libproxml.so -blibpath:${DLC}/lib:/usr/lib:/lib /usr/bin/rtl_enable -R libicuuc-psc.so -blibpath:${DLC}/lib:/usr/lib:/lib /usr/bin/rtl_enable -R libicui18n-psc.so -blibpath:${DLC}/lib:/usr/lib:/lib /usr/bin/rtl_enable -R libicuuc48_psc.so -blibpath:${DLC}/lib:/usr/lib:/lib /usr/bin/rtl_enable -R libicui18n48_psc.so -blibpath:${DLC}/lib:/usr/lib:/lib /usr/bin/rtl_enable -R libxerces_psc.so -bbigtoc -blibpath:${DLC}/lib:/usr/lib:/lib /usr/bin/rtl_enable -R libprosoap.so -bbigtoc -blibpath:${DLC}/lib:/usr/lib:/lib For 64-bit OpenEdge AIX: cd ${DLC}/lib /usr/bin/rtl_enable -X64 -R libproxml.so -blibpath:${DLC}/lib:/usr/lib:/lib /usr/bin/rtl_enable -X64 -R libicuuc-psc.a -blibpath:${DLC}/lib:/usr/lib:/lib /usr/bin/rtl_enable -X64 -R libicui18n-psc.a -blibpath:${DLC}/lib:/usr/lib:/lib /usr/bin/rtl_enable -X64 -R libicuuc48_psc.a -blibpath:${DLC}/lib:/usr/lib:/lib /usr/bin/rtl_enable -X64 -R libicui18n48_psc.a -blibpath:${DLC}/lib:/usr/lib:/lib /usr/bin/rtl_enable -X64 -R libprosoap.so -bbigtoc -blibpath:${DLC}/lib:/usr/lib:/lib 6. Any projects/applications that were deployed using [RELEASE_VERSION] should be rebuilt and redeployed after the patch is installed. 7. After applying the patch for updating Progress Developer studio plug-ins, execute the following commands: a. {DLC}\oeide\integrateArchitect.bat -update b. {DLC}\oebpm\studio\integrateOEBPM.bat -update For existing workspace, delete following files & start PDSOE: ${WORKSPACE}\.com.savvion.studio\scripting\bizlogic\data_operations.palette ${WORKSPACE}\.com.savvion.studio\scripting\bizlogic\main.palette ${WORKSPACE}\.com.savvion.studio\scripting\bizsolo\main.palette ${WORKSPACE}\.com.savvion.studio\scripting\bizlogic\jst\jst_get.palette ${WORKSPACE}\.com.savvion.studio\scripting\bizlogic\jst\jst_set.palette ${WORKSPACE}\.com.savvion.studio\forms\widgets.xml 8. Before applying the patch for updating OpenEdge Management, backup the following files/directories in your OpenEdge Management installation directory. copy "web" to "web.bak" copy "config" to "config.bak" copy "jars" to "jars.bak" After extracting the contents of the HOTFIX for OpenEdge Management you will need to copy the contents of the OEMgmt directory from the archive to your OpenEdge Management installation directory. After copying the contents of this directory you will need to restore the web/WEB-INF/web.xml file from your backup copy. copy "web.bak/WEB-INF/web.xml" to "web/WEB-INF/web.xml" Once you are satisfied with the patch installation, you may remove these backup files at any time. 9. If OE BPM is being used, execute scripts in {DLC}\oebpm\server\dbscripts\openedge\sdmscreate.sql on OEBPM db if the table 'DOCUMENT_STORE' doesn't exist. This can be done using sqlexplorer or db navigator. 10. If OEBPM is being used, execute the DB migration tool bundled as PROGRESS_OEBPM_DATAMIGRATION_11.3_TO_11.3.X.zip under {DLC}/oebpm/server/migration folder. 11. To enable the multi-tab support for Business Process Portal, add the following line of code to the bpmportal.conf file at {DLC}/oebpm/server/conf/ location: bpmportal.relogin.or.continue = true 12. If OEBPM is being used, add the following lines of code to the bpmportal.conf file at {DLC}/oebpm/server/conf/ location: # # BPM Portal limited Longchar display dataslot length # false # The max. length of Longchar dataslot shown in BPM portal pages. # ANY # bpmportal.longchar.display.dataslot.length=25 13. If OEBPM is being used, edit the value of the property "bpmworkflow.root" in the file bpmworkflow.conf at {DLC}/oebpm/server/conf/ location: Old entry - bpmworkflow.root = @@PATH_TAILORED_BY_INSTALL@@/BPMWorkFlow New entry - bpmworkflow.root = @@PTH_TAILORED_BY_INSTALL@@/BPMWorkflow Note : Only the upper case letter "F" in the word "BPMWorkFlow" need to be changed to lower case. Please note that steps 11,12 and 13 needs OEBPM servers to be restarted for the changes to take effect. 14. When running PDSOE or .NET code in the procedure editor, if you encounter the following errors: Java was started but returned exit code=1 C:\Progress\DeveloperStudio#.#\jre\bin\javaw.exe -Xms40m -Xmx384m -Dosgi.framework-extensions=com.openedge.keybiding.hook -XX:MaxPermSize=256m -jar ... you must unblock these .NET dlls: $DLC\bin\dotnetmsgs\Progress.Messages.dll $DLC\bin\Progress.clrbridge.dll $DLC\bin\Progress.NetUI.dll Go to the dll's file properties, in the security section under the General tab there will be an Unblock Button. Select this for each dll. After the changes have been applied the Progress Development Studio will launch succesfully. ----------------------------------------------------- 3. List of issues fixed in 11.3.3.045 ----------------------------------------------------- Issue ID: OCTA-7749 --------------------- There is an issue with indexes that span the 32\/64-bit boundary. You are vulnerable if the High Water Mark of the area containing the index is above 2 gig. To determine your HWM, use prostrct statistics on areas containing indexes. If the number of active blocks listed by the prostrct statistics for an index area is greater than (2^31 \/ area records per block) the area is susceptible to this defect. Error messages that indicate that you may have the issue include: Getting error Attempt to read block 0 which does not exist in area , database . (210) (14684) or (12011) Could not find record. Recid 0, error number 4, return -1220 After the hotfix is applied a full index rebuild of effected indices (if known) or all indices (if specific indices which have this problem are unknown) is required to eliminate index corruption created prior to the applicaiton of the fix. The fix is intended to prevent future occurrences of this defect but cannot fix existing instances of the index corruption, only the idxbuild can cleanup existing problems after the fix is applied. ----------------------------------------------------- 4. List of issues fixed in previous 11.3.3.x releases ----------------------------------------------------- 4.1 - List of issues fixed in 11.3.3.001 Issue ID: PSC00312419 --------------------- When using a ProDataSet containing tables in a parent / child relationship with the Combit List & Label .NET reporting control bound to a BindingSource using the constructor with the reportMode (previously called batchMode) parameter set to True, the wrong child data is displayed as follow: * First parent * Second parent ---- First Child of FIRST parent ---- Second Child of FIRST parent ---- * Third parent ---- First Child of SECOND parent ---- Second Child of SECOND parent ---- 4.2 - List of issues fixed in 11.3.3.002 Issue ID: PSC00313366 --------------------- Appserver agent can get stuck in SENDING state after adminserver have performed warm restart Issue ID: PSC00316197 --------------------- If a window is maximized and then minimized, the children of the parent window will not be visible on-screen when the parent window is restored. 4.3 - List of issues fixed in 11.3.3.003 Issue ID: PSC00315657 --------------------- If you have a custom codepage defined in convmap.cp, non-ASCII characters might not be correctly handled when using XML. This can occur for reading or writing documents, with either DOM or SAX APIs. 4.4 - List of issues fixed in 11.3.3.004 Issue ID: PSC00315557 --------------------- Report Engine (PRORE32.EXE) may crash when printing a single page from the print preview window using the "Print Current Page" button. Issue ID: PSC00321263 --------------------- DateTimeTZ (datetime dataslot) not persisted for locales other than English - exception thrown in bpmportal.log 4.5 - List of issues fixed in 11.3.3.005 Issue ID: PSC00319199 --------------------- Broker detects death of a _sqlsrv2 server when the ports are being scanned by a 3rd party port scanning software. Issue ID: PSC00320781 --------------------- The following error will trigger followed by the abnormal termination of the AppServer server process: SYSTEM ERROR: fmget: bad record or unknown value. (13) 4.6 - List of issues fixed in 11.3.3.006 Issue ID: PSC00288880 --------------------- Executing a FOR EACH ..., OF, query where the WHERE clause on the second table contains an OR operator does not all results. Issue ID: PSC00322751 --------------------- Executing dynamic query, referencing 2 separate tables, via DataServer for Oracle returns the wrong results due to incorrect ORDER BY clause 4.7 - List of issues fixed in 11.3.3.007 Issue ID: PSC00322883 --------------------- Foreign owner in query's inner select is different from Oracle user name used in SchemaHolder connection string. 4.8 - List of issues fixed in 11.3.3.008 Issue ID: PSC00326979 --------------------- The Developer Studio Database connections preferences blanks database connections parameters when there are many connections defined. Issue ID: PSC00328077 --------------------- Unable to generate a Java Open Client proxy for a ProDataSet that contains more than 46 data-relation objects. Issue ID: PSC00329866 --------------------- If a .NET object gets a property from an ABL hybrid, and the ABL hybrid has a get accessor implementation for the property, the AVM leaks a small amount of memory. Calling this property repeatedly will eventually exceed the -s threshold of the AVM, resulting in error messages such as the following: SYSTEM ERROR: -s exceeded. Raising STOP condition and attempting to write stack trace to file 'procore'. Consider increasing -s startup parameter. (5635) 4.9 - List of issues fixed in 11.3.3.009 Issue ID: PSC00328910 --------------------- Oracle data server QUERY-TUNING phrase HINT causes memory leak. Issue ID: PSC00329820 --------------------- Failure in large object operation error when reading no-lock and writing CLOB field of a table at about the same time 4.10 - List of issues fixed in 11.3.3.010 Issue ID: PSC00328106 --------------------- Client may raise error "Array subscript 0 is out of range" when executing the BUFFER-COPY statement or function and the table has an assign trigger for one or more fields. Issue ID: PSC00328910 --------------------- Oracle data server QUERY-TUNING phrase HINT causes memory leak. 4.11 - List of issues fixed in 11.3.3.011 Issue ID: PSC00313083 --------------------- Runtime exception occurs when BizPulse rule makes use of custom Java class on Windows platforms. 4.12 - List of issues fixed in 11.3.3.012 Issue ID: PSC00325661 --------------------- If a GUI client (prowin32.exe/prowin.exe) process is started from (thus parented to) a Windows Service, on Windows 2012 server this causes a system-wide delay in applying Windows Group Policies when a user logs in. As the GUI client is a GUI process, it interacts with the UI through Windows messages. Newer versions of Windows prevent services from interacting with the UI, with Session 0 Isolation. The OpenEdge GUI clients should not be used for running Windows Services, even when using -b. Instead, use the OpenEdge character client (_progres.exe). 4.13 - List of issues fixed in 11.3.3.013 Issue ID: PSC00332409 --------------------- HotFix for SSLv3/TLS. 4.14 - List of issues fixed in 11.3.3.014 Issue ID: PSC00331692 --------------------- Adding LONGCHAR to ParameterList object isn't passing in the value correctly, causing subsequent Class:Invoke() to fail. Either Invoke() crashes the session or the LONGCHAR value in the invoked method is left blank where it should contain text. Issue ID: PSC00333127 --------------------- Error 1436 when compilation foreign owner is different than run time foreign owner. 4.15 - List of issues fixed in 11.3.3.015 Issue ID: PSC00333499 --------------------- The X-DOCUMENT and SAX-READER handles are vulnerable to an XML Entity Expansion Attack. Issue ID: PSC00333559 --------------------- X-DOCUMENT and SAX-READER are susceptible to XML External Entity Processing vulnerability - can expose contents of targeted files. 4.16 - List of issues fixed in 11.3.3.016 Issue ID: PSC00334314 --------------------- Value of a shared variable changes unexpectedly The value of a shared variable may be changed when the SET option of a .NET WAIT-FOR statement is used. In this case, the value which is supposed to be stored in the variable specified by the SET option is instead stored in an unrelated shared variable. 4.17 - List of issues fixed in 11.3.3.017 Issue ID: PSC00338175 --------------------- 1436 error is still experienced even with hotfix 11.3.3.014 4.18 - List of issues fixed in 11.3.3.018 Issue ID: PSC00317319 --------------------- Possible crash when running automatic transaction procedure on AppServer when triggers are involved Issue ID: PSC00335770 --------------------- The following error is returned when trying to connect to multiple Web Services via HTTPS from the same OpenEdge client session: Error loading WSDL document : Internal error: Receive status (11748) 4.19 - List of issues fixed in 11.3.3.019 Issue ID: PSC00248677 --------------------- Dynamics treeview doesn't show folder page details **************************************** In some circumstances a dependent window attached to a Dynamics treeview that contains a folder, doesn't show data on page one unless the page is refreshed by clicking on another page or Treeview node. Issue ID: PSC00338175 --------------------- 1436 error is still experienced even with hotfix 11.3.3.014 Issue ID: PSC00339302 --------------------- A query with a complex CONTAINS expression can hang when a client is connected client/server to a database. A complex CONTAINS expression would contain more than 5 words with '&' and '|' operators. 4.20 - List of issues fixed in 11.3.3.020 Issue ID: PSC00331601 --------------------- Getting error 455 when a browse on a query for a table that has a database FIND trigger that returns error. Unable to read record buffer because it is empty. (455) When browsers are not involved, the program gets an error condition raised (as if the RETURN ERROR was still in effect). 4.21 - List of issues fixed in 11.3.3.021 Issue ID: PSC00338995 --------------------- Using OpenEdge DataServer for Oracle when running a specific program that executes the REPOSITION TO ROWID, it crashes. 4.22 - List of issues fixed in 11.3.3.022 Issue ID: PSC00288458 --------------------- Intermittent java.lang.NullPointerException errors when stopping and starting an appserver under OpenEdge 11.3, 11.4 or 11.5. Issue ID: PSC00327816 --------------------- The client may crash trying to compile a EXPORT statement where a field has a compile error in the VIEW-AS phrase. We were trying to display information for an error message that couldn't actually be captured in this particular case. We updated this to make sure that we were only trying to display information that we had stored as part of the error handling process. Issue ID: PSC00339873 --------------------- The Web Service client declares an invalid namespace in the SOAP Header. For example: ns0:ns1="http://MyNamespace.xsd" Issue ID: PSC00341428 --------------------- ORACLE error -907 see "ORACLE Error Messages and Codes Manual". (1252) -- ** missing right parenthesis Issue ID: PSC00341715 --------------------- ORACLE errors due to extra characters being generated in SQL query by the DataServer. 4.23 - List of issues fixed in 11.3.3.023 Issue ID: PSC00339435 --------------------- DB cannot get through crash recovery after normal shutdown with open live JTA transactions. Fails with: (-----) dbxaUndoLogical: after dbxaGlobalTxnReserve call Issue ID: PSC00340735 --------------------- DB cannot get through crash recovery after normal shutdown with open live JTA transactions. 4.24 - List of issues fixed in 11.3.3.024 Issue ID: PSC00317491 --------------------- In some specific cases, when we undo sub-transactions where db triggers are involved, the client may crash or exhibit incorrect behavior if update are done in different sub-transactions and inside the trigger. Issue ID: PSC00342642 --------------------- Application server broker's SSL connections slow down over time and cause client timeout 4.25 - List of issues fixed in 11.3.3.025 Issue ID: PSC00336832 --------------------- Network error: \"Handshake Failure. No supported cipher suites (9996) (10770)\" occurs when connecting to remote secure wsa. Issue ID: PSC00341777 --------------------- Oracle single-shot server-side query very slow compared to the same query executed with server-side joins disabled. 4.26 - List of issues fixed in 11.3.3.026 Issue ID: PSC00334000 --------------------- The POODLE fix does not work for Webclient. When a TLS connection attempt is made, the following error is returned: SSL Client handshake failure (-54) certificate signature failure: for xxxxxxxx.0 in C:\Program Files (x86)\Progress Software\WebClient\certs 4.27 - List of issues fixed in 11.3.3.027 Issue ID: PSC00257044 --------------------- Additional cleanup of non async-signal safe localtime and ctime functions. Issue ID: PSC00260459 --------------------- The AVM leaks memory when an instance of a class that subclasses the JsonObject, JsonArray or ObjectModelParser class is deleted or garbage collected. Issue ID: PSC00308471 --------------------- The LOAD-IMAGE() method returns yes even if the image file is not found. LOAD-IMAGE should return no if it can't find the image file. Issue ID: PSC00342586 --------------------- Session crashes on DELETE PROCEDURE THIS-PROCEDURE when the procedure contains temp-tables. Issue ID: PSC00343353 --------------------- Error 26 and 142 when starting the Data Administration tool while connected to more than 32 databases. 4.28 - List of issues fixed in 11.3.3.028 Issue ID: PSC00342428 --------------------- The Webspeed agent crashes with a Memory Violation (49) error when uploading a unicode XML file. Issue ID: PSC00343318 --------------------- The message displayed by the MESSAGE UPDATE statement may be incorrectly truncated if its length in bytes is longer than its length in columns and the length in bytes exceeds the number of columns available to display the message in the message area. 4.29 - List of issues fixed in 11.3.3.029 Issue ID: PSC00344157 --------------------- Client logging can cause client to deadlock itself due to async signal handling 4.30 - List of issues fixed in 11.3.3.030 Issue ID: PSC00343489 --------------------- Inconsistency assigning datetime-tz to datetime/date when SESSION:TIMEZONE set. 4.31 - List of issues fixed in 11.3.3.031 Issue ID: PSC00344804 --------------------- wtbman -q -all command to list status of resources failed with error. Error Message: No entries for personality WS (8300) 4.32 - List of issues fixed in 11.3.3.032 Issue ID: PSC00345114 --------------------- The AVM crashes when executing ABL code that access large CHARACTER fields. 4.33 - List of issues fixed in 11.3.3.033 Issue ID: PSC00204238 --------------------- The SET-READ-RESPONSE-PROCEDURE method may cause a memory leak if you call it more than once for a given socket object that has not been deleted (such as in a loop that connects, process and disconnects the socket). Issue ID: PSC00344222 --------------------- Webspeed agent may experience a memory violation crash if a cgi form variable name is greater than 32K. This can happen if input is improperly labeled as cgi form data when it is something else, e.g. and XML document. Issue ID: PSC00345639 --------------------- Dynamic query running against DataServer for MS SQL Server returns the wrong results when joining on an array element. 4.34 - List of issues fixed in 11.3.3.034 Issue ID: PSC00255733 --------------------- Dialogs are not centered over floating child forms **************************************** ABL dialog boxes are not centered over floating (undocked) .NET child forms. They are parented and centered over the parent .NET form. 4.35 - List of issues fixed in 11.3.3.035 Issue ID: PSC00347545 --------------------- On a replication-enabled database, executing prostrct builddb, followed by an index build, the index build will crash with a memory violation error (49). 4.36 - List of issues fixed in 11.3.3.036 Issue ID: PSC00347434 --------------------- Calling LOCALTIME function causes db crash Issue ID: PSC00348399 --------------------- Attempting to connect from an ABL client to a new Web Service that the government in Czechoslovakia has mandated all businesses use for collecting electronic receipts generates the following error: Error loading WSDL document: 'mixed" attribute must not be used on the parent of 'xs:simpleContent' declaration (11748) 4.37 - List of issues fixed in 11.3.3.037 Issue ID: PSC00348894 --------------------- Slow network performance when making appserver calls in all unix machines. 4.38 - List of issues fixed in 11.3.3.038 Issue ID: PSC00348894 --------------------- Slow network performance when making appserver calls in all unix machines. 4.39 - List of issues fixed in 11.3.3.039 Issue ID: PSC00342409 --------------------- CHUI client(shared memory connection or client\/server) crashing after enable auditing on a utf8 database under 64bit OE. works fine for 32bit OE Issue ID: PSC00349765 --------------------- It may occur that a JTA enabled database, that has pending transactions when it crashes, is not able to be restart, showing the message: BROKER: rlLockApply: failed to acquire lock ret -1217 1 4.40 - List of issues fixed in 11.3.3.040 Issue ID: PSC00349740 --------------------- Shutdown of AdminServer can hang if remote Web Services Adapter ping requests initiated from OpenEdge Management fail to terminate. Issue ID: PSC00349746 --------------------- OEM pings to remote WSA appear to have a very long timeout 4.41 - List of issues fixed in 11.3.3.041 Issue ID: PSC00354138 --------------------- Client hung with asynchronous call to malloc on Linux. This can be triggered by sending multiple signals to the process in a very short space of time. 4.42 - List of issues fixed in 11.3.3.042 Issue ID: PSC00325183 --------------------- WebSpeed agent randomly choose the log file sequence after restart of the broker. The last written log file is skipped and writes to much lower sequence number associated wit the log file. 4.43 - List of issues fixed in 11.3.3.043 Issue ID: PSC00356583 --------------------- Error 12668 when running clientConnect to connect directly to a SonicMQ broker after Windows 10 Creators update. 4.44 - List of issues fixed in 11.3.3.044 Issue ID: PSC00347994 --------------------- Corrupt index entries for _aud-audit-data._event-context if -cpinternal is different from database codepage and the data contain non-ascii chars.