Pega Infinity™ includes an industry-standard patch release process to simplify and maintain high-quality releases. Several cumulative patches are released a year for each release stream. The Resolved Issues page contains information about client-reported issues that have been addressed for the specific release.
For a formatted complete set of the Resolved Issues for this release, download the PDF attachment at the bottom of this page. (Note that you must be logged in to access the attachment.)
| Ticket # | Issue # | Title | Description | |
|---|---|---|---|---|
| SR-C41102 | 410424 | Resolved error in editing flow rule assignment shapes | The submit of an assignment properties modal window was not closing the modal, and instead opened a new window with an error page indicating a NullPointerException. This happened whether or not there were changes made to the assignment properties, and occurred for the assignment shapes on any flow. This was traced to a UI limitation in the backward compatibility of the "parent.parent.pyWorkClass" property which caused the property reference to not be resolved for hidden controls in client side template. To correct this, updates have been made to "pzAssignmentTab" (in Data-MO-Activity-Assignment class ) and "pzPropertyPanel" (in Data-MO-Activity-Assignment-WorkAction class). | |
| SR-C44836 | 410368 | Drag and drop fixed for report editor | An issue where it was not possible to drag properties to the report columns or filters in the report editor was traced to the HTML fragment RRBodyDiv outputting values from the paging settings in an unsafe manner, causing syntactically invalid JavaScript to be generated when the "Previous and next only" paging option was used. To correct this, RRBodyDiv has been updated to generate valid JavaScript even if some paging settings are blank. | |
| SR-C50355 | 407423 | Corrected applies-to class in Test case rule | On test case rule form with assertion type Result count Filter class MIDAS-SR and filter condition @endswith(.cid,".html"), opening the function builder using the gear icon and trying to use the Test feature of the Function builder resulted in an incorrect applies-to class. The result was for the Test case instead of showing correct Applies to class of MIDAS-SR. This was traced to the Filter class of the context page not being passed to the function builder in Test case rule, and has been fixed. | |
| SR-C51798 | 412382 | Modified pzPopulateTrainingandTestDetails to resolve input error in test model | After creating a text model, trying to update it resulted in errors. This was caused by the input bean for pzPopulateTrainingandTestDetails getting reset and causing a nullpointerexception while trying to load the spell checker because the taxonomy data was empty. This has been fixed by ensuring the input param object is created after setting the taxonomy data. A null check has also been added in the creation process of the input param object in pzPopulateTrainingandTestDetails. | |
| SR-C51848 | 406494 | Additional validation added for administrator role name in current application | When running an extract rule from designer studio, the when rule "hasadminrole" was getting called to check if the user has the "pegarules:sysadmin4" role. After upgrade, this when rule was returning false if "pegarules:sysadmin4" was defined in 'manage dependent roles', requiring the role to then be entered manually. This was traced to a condition where the BIX extract was being run from a CS application that did not have any access role with the current application name :Administrator, but instead used the access role name :SysAdm4, . To resolve this, an additional condition has been added to validate whether the current application has an access role with name "current application name:SysAdm4" in their access group to run BIX extract. | |
| SR-C53759 | 404974 | Corrected revalidate and save of Service REST rule after upgrade | After upgrade, opening an existing Service REST rule and trying to save it in a higher ruleset using the revalidate and save tool resulted in the error "pyResourcePathParams.pyParameterName : This field should not be blank". This was caused by the Service REST xml not getting updated and has been corrected by adding UpgradeOnOpen to the process. | |
| SR-C53866 | 402883 | Added handling for custom skin using the $ symbol for value | After upgrade, using a custom style sheet to display the $ symbol prior to values being shown or entered caused the application skin to not log and throw an error. This was caused by a missed use case for adding $ in css :before or :after content due to a dependency on this "$" notation for fetching right mixin color from skin color. The array index out of bounds exception was generated by the statemachine parser while handling special characters because it considered the quotes of csscontent to be corrupted. To resolve this, an extra check has been added for handling """ and "'" cases in cssCustomParser.java and junits have been updated. | |
| SR-C55082 | 402920 | Cassandra upgrade | Cassandra has been upgraded from 2.1.14 to 2.1.20. | |
| SR-C56648 | 405427 | Updated logic for imagepicker to resolve icon changes in containers | When attempting to change the expand or collapse icon, the error "pega.openUrilInModal.showModalDialog - error when calling callbackfunction : 'C.getElementById is not a function' in pega" was generated. Investigation showed that the browser-level JS file 'pega_ui_harness_deferred' was missing parameters needed to return the value from the showModalDialog method due to a hardcoded loop for the imagepicker only iterating through a set number of levels. This has been resolved by updating the logic to recursively find the document. | |
| SR-C56652 | 402945 | Struts upgraded | Apache Struts has been upgraded to version 2.5.17. | |
| SR-C56774 | 409916 | Removed double encoding on unicode local action parameter | After enabling Unicode based rulenames using DSS, a Repeating table layout event action could not call a Unicode based local action even when an event action on a Dynamic Layout Section could call it. This issue occurred when the name of the flowaction had Unicode characters like Japanese; the open local action was not able to open the modal when launched from a grid because the Location parameter used by the grid was being double encoded. To correct this, the system will avoid the double encoding by passing the true parameter to actionURL.toQueryString for subsequent encoding by asyncrequest's toURL. | |
| SR-C57130 | 402934 | TokenRangePartitioner modified to avoid Cassandra timeouts | The previous DSS settings caused the DDS dataset on Browse to create thousands of partitions. This caused thousands of files to be created for a file dataset which uses a DDS dataset as input. In order to reduce the number of partitions while still being able to take advantage of the changes to ensure Cassandra always reads from the local node and thus avoid the timeout, the TokenRangePartitioner will now create composite partitions with multiple token ranges (up to 60) per partition. This will reduce the number of files created by a factor of 60 reducing the number of files created from 60 - 100. Additionally, timestamp accuracy issue and batching for single case data flows have been fixed. | |
| SR-C57879 | 406005 | Updated parameter handling for SECU0001 alert for accuracy | In the SECU0001 alert, some fields contained incorrect values pertaining to a previous HTTP request, not the one that caused the alert, causing misleading information to be logged and sent to PDC. This was found to be an issue where when unexpected properties were found, the alert was being thrown from an early stage of processing, i.e., during inputProcessing where the current activity had not yet been set to the requestor state and therefore the current activity was not there in the stack frame. This has been addressed by getting the activity name from the HTTP request parameters and updating the diagnostic information with it so the alerts will be accurate. |
|
| SR-C58241 | 409217 | Agent Access Group corruption resolved | The Access Group used by agents was intermittently becoming corrupted. This was caused by a timing issue where sometimes the conclusion was built before the rulesets to the requestor were set up, resulting in the role being added to the RoleNotFound cache. To resolve this, the RoleNotFound cache has been removed so that every time a role lookup happens the system will check the database. | |
| SR-C59018 | 407438 | Resolved chart page lookup failure for Express Portal | In the Express portal, exceptions were generated when clicking on a chart in the case volume report. This was caused by a lookup failure: when the activity doUIAction was called (after drill down), the pyReportContentPage was getting deleted. This has been corrected by setting the property pyRetainPage to true within the displayReport activity to keep the pyReportContentPage from being deleted in frameless portals. | |
| SR-C60087 | 406620 | Failure status set correctly for WL to WL bulk WO transfer | When a work object that was part of a bulk transfer was being moved from one operator's work list to that of another operator, a blank status was appearing if the work object was resolved by the originating operator before it was submitted. This was caused by a missed use case for setting error messages and has been corrected by updating pzBulkProcessItemsInHarness. | |
| SR-C61154 | 411074 | Improved DB stability and retry for issues with RetryContext$ExceptionForDatabase | The agent was crashing when there was a report using a pr_read_from_stream for the 'where' clause combined with a PostgreSQL version with 'on pr_read_from_stream'. This has been addressed by improving the stability of the database and adding code to better handle reconnect and retry when there is a database problem. | |
| SR-C62241 | 405942 | Error handling corrected for Actions/Refresh menu actions | When using the ACTIONS / REFRESH menu action on a case that had a required property, the busy indicator appeared and remained on screen, blocking interaction with the portal until the browser was refreshed. The error "pzpega_ui_harness_deferred_ts_1342430166!!.js:1969 Uncaught TypeError: Cannot read property 'stopPropagation' of undefined at Object.stopPropagation" appeared. This was traced to a fault in the error handling and has been corrected by modifying "pzpega_control_actions_refresh.js" so the event will be passed as parameter in the "pega.u.d.submitWhenFail()" call and the "SubmitInProgress" value will be set in context in pzpega_ui_doc_submit.js. |
|
| SR-C62434 | 410619 | Optimize Schema Index Creation updated to use Admin Datasource when available | When configured according to the Pega Installation Guide, the SQL Server base user did not have the ALTER privilege enabled on the data schema. The Optimize Schema Index Creation was using the base user datasource "jdbc/PegaRULES" instead of the admin datasource "jdbc/AdminPegaRULES" even though an admin connection was available. This was an oversight during development, and has been fixed by modifying NoDBAUtils.java to use an admin connection if one is available. In addition, a clarification has been added to the UI to alert the user that an optimization job has been scheduled rather than completed. | |
| SR-C62475 | 402873 | Improved security on attachments added in Case and Pulse | Data-Content-Image instances were getting created for attachments added in Case and Pulse. Since Data-Content-Image doesn't correspond to security once they are cached, they could be accessed by any authenticated user. To correct this, extensions have been added to stop generating Data-Content-Image while creating a pulse attachment post or case attachment. | |
| SR-C62835 | 408361 | pzReindex defaults to true during branch merge and initial save | When using Declare Indexes intended to be fired when a KYC rule type was saved, merging a KYC rule type from a branch to a ruleset which contained the rule resulted in the declare indexes not firing from Rule-pyMergeInstance. If the branch was merged into a ruleset which did not contain the KYC rule, the Rule-pyMergeInstance fired the declare index rules as expected. This was an issue with using Declare Indexes in an unexpected way: Declare Index works only when a property that is tracked is changed, and merging a branch rule to base ruleset does not change any tracked property. However, there was a Force Reindex happening when the saved instance had a different pzInskey than one that already existed in the db. This occurred when the rule was merged to new ruleset where the pzInskey in the DB would be null, and was expected behavior. In order to ensure Declare Indexes fire during a merge, pzReindex will be set to true in the Merge Wizard before saving the rule while doing branch merge. It will be reset/set to false after the rule has been merged and saved to avoid the additional overhead of triggering re-indexing every time the rule is saved afterwards. Further enhancements will be made in an upcoming release regarding parameter handling in templatized UIs. | |
| SR-C64243 | 412374 | Data flow region switching method added to prevent truncating event strategy data | An enhancement has been added to allow customers to switch data flow execution from one region to another without clearing event strategy data. This process looks like (1) Dataflow is started on one region. (2) Dataflow is paused in same region. (3) Dataflow is started in second region. Previously, step 3 would delete the existing event strategy data. To prevent that, a new run option has been added named 'pyDataFlowStorageTruncateOnStart'. By default, this is set to true and truncation will occur. If set to false, the event strategy data will not be deleted. | |
| SR-C64783 | 407089 | Corrected handling for SAML logoff | On SAML logoff, the error "There has been an issue; please consult your system administrator; Status:fail ... No certificate found in truststore" appeared. Investigation showed this was an issue with the aliasing of certificates and signing that led to the requestor not being terminated for that logoff response. To correct this, when Pega receives a logout request which is invalid, it will terminate the session instead of throwing a PRRunTimeException. | |
| SR-C65239 | 411326 | Check added to resolve nested grid component mismatch | When using a nested repeating grid with expand pane option and expand multiple row options, running a screen flow resulted in a component mismatch error when moving from one assignment to another. This was traced to a limitation in the implementation that sometimes caused the parameter of the outer and inner grid to become corrupted or affect the wrong grid. In this case, the parameters startIndex and recordsInCurrentPage were resulting in a numberformatexception which in turn caused the component stack mismatch error. To fix the issue, an empty check has been added for startIndex and recordsInCurrentPage in the grid generation code. | |
| SR-C65438 | 406711 | Added SAML parameter page generation for local use to bypass null auth object | After defining any step which used parameters in the Post authentication activity of a SAML authentication service, the authentication failed with a java.lang.NullPointerException at com.pega.pegarules.session.internal.mgmt.Executable.putParamValue(Executable.java:3030). This was traced to local logic inserted to iterate through the SAML attributes that used a null parameter page. To resolve this, code has been inserted that will generate a new parameter page for the iteration rather than getting it from the authentication object. |
|
| SR-C65576 | 405497 | Data-WorkAttach-File made available for use of custom pxCreateSystemID | An enhancement has been added to allow storing the pxCreateSystemID from pyWorkPage to Data-WorkAttach-File and Link-Attachment instances which were created on attaching a file. Previously, the Data-WorkAttach-.Default Data transform and Work-.AttachFile both set the pxCreateSystemID property to pxProcess.pxSystemName for Data-WorkAttach-File and Link-Attachment instances respectively. While Data-WorkAttach-.Default is ‘available’ and can be overwritten to set pxCreateSystemID as desired, the Work-.AttachFile was ‘Final’ and it was not possible to overwrite that to set a pxCreateSystemID manually. In order to support this use, extension points have been provided for the pyDefault data transform in the activities Work-.AttachFile and Work-.WorkURL which can be overridden. | |
| SR-C65633 | 408044 | Improved handling for parsing multiple formats of DateTime | In order to further correct an issue with DateTime formats, the system will now use the parseAsDateTime and parseAsDate functions to properly handle the parsing of the date and date time values. The Pega format will be tried first, and if that doesn't work then the UTF formats will be used. If there is still an issue, the system will call back on the original parseAsXSDDateTime/Date which should catch all cases. An exception will no longer be thrown until after all parsing methods are tried. | |
| SR-C65757 | 404535 | Support added for POST binding in SAML logout | Enhancements have been added in code and rule changes to support POST binding for SAML logout. In addition, handling has been added to resolve an ArrayOutOfIndexException error that was appearing if IDP sent parameters without any value. | |
| SR-C66114 | 408082 | Resolved pg empty error after db.open and improved error message | When opening the data page D_RepositorySettings, the error "pg empty after db.open" appeared in the logs. This was due to the data page referring to the files webwb/zchecked.gif and webwb/zchecked.gif in many sections of the data page rule form despite these files having been removed as part of a recent cleanup meant to reduce storage requirements. In order to resolve this error, the zunchecked.gif and zchecked.gif images have been added back and the error message generated in the RUF pzGetbase64encodedImagehas been improved to add the target filename: "error occurred while opening the image:webwb/zunchecked.gif". | |
| SR-C66283 | 411634 | Report Usage updated to always fully populate statistics regardless of page source | After creating a report under the Log-ReportStatistics class, it was seen that reports under report usage were not reflecting the exact data. The inconsistency observed with embedded reports was due to grids being sourced by virtual RD, for which the report page had already been loaded elsewhere. In this case it was assumed that the statistics were already populated because the report page was already loaded. To ensure that reports are consistently accurate, pxRetrieveReportData has been updated to always call pzPopulateStatisticsInfo. | |
| SR-C66624 | 406226 | Case type validation localized | Localization has been added to the Case type validation description. | |
| SR-C66627 | 412779 | Reject step placeholder text localized | Localization has been added to the placeholder text of the approval reject step. | |
| SR-C66633 | 408774 | Corrected DADT mapping for PegaDATA tables during upgrade | After upgrading, old data tables being upgraded to new data tables (local storage) had the DADT table mapping changed from PegaDATA to CustomerDATA schema even though the tables themselves remained in PegaDATA format. This was traced to a missed use case in an enhancement that allowed creating a data type only in a CustomerData data base. To resolve this issue, the validation for restriction of PegaDATA has been removed and the param.nonPegatable will be set to false in cases where the database type is PegaDATA in pzCreateLocalDataStorage. | |
| SR-C66654 | 406218 | Pulse App Studio localization added | Localization has been added to the Developer collaboration text in Pulse App Studio. | |
| SR-C66998 | 404593 | Thread name implementation fixed for FCM use with micro DC | When attempting to invoke a remote case in the Interaction Manager portal via FCM with the Customer Service for Insurance application which uses micro DC, the case was not invoked and an empty screen was displayed. Invoking the case from the interaction portal resulted in a JavaScript error. Invoking the case from a WSS or directly from the Create button in Designer Studio worked correctly. This was traced to a recent change to the implementation for generating the thread name. Because of this, CRM apps could not load the Federated cases via micro DC. This has been fixed through changes in "pzFCMMashupGadget" to replace the slash with underscore for the FCM thread. In addition, a modification was made to use the pega:onlyonce tag so mashup scripts are not loaded multiple times and will maintain their unique data-pega-gadgetname attribute values. | |
| SR-C67376 | 406719 | SOAP Connector Wizard 'Create' button conditionally accessible | Previously, the 'create' button in the SOAP Connector Wizard was disabled if the application was password protected. As an enhancement, the create button is now accessible if the existing ruleset and version is selected. | |
| SR-C67870 | 407164 | Added evaluation client thread name check to Change Tracker | The Change Tracker was intermittently not working as expected between main DC and micro DC div when trying to use a change in microdc context (specifically an update to D_Interaction.InteractionHeaderRefreshDateTime) to refresh the document in the main DC. In order to ensure consistent updating, a check has been added for the current thread name while executing evaluate client conditions. | |
| SR-C68443 | 410237 | Changed French locale to use getshortmonths to discriminate between June and July in datepicker | When using the French locale, there was a problem selecting the month of July from the date picker because the three-letter abbreviations of June and July are the same in French. To resolve this, the system will now use the Java API getShortMonths in the French locale which gives 4 characters for the months. | |
| SR-C68500 SR-C70639 |
411294 409665 |
Node ID correctly transmitted for PEGA0066 Alerts | The PEGA0066 alert containing information about synchronization errors or connection lost did not contain the correct value in the Node ID field, causing PDC to not be able to handle the alerts on mobile. This has been corrected. | |
| SR-C68682 | 412925 | Layout Groups Accordion Styles modified to use refresh on change | After updating Layout Groups Accordion styles in a skin rule and saving and checking in the rule, re-opening the skin rule showed that the changes had not been saved. Investigation revealed that in accordion format, the font section is being rendered for the "header" and "tabs" of the of layout group but the third tab was not visible to accordion format. To resolve this, pyComponentDetails of Embed-skin-component-layoutGroup has been modified to use on onchange refresh section instead of "Run visibility condition on client". | |
| SR-C68750 | 407492 | PEGA0066 alert sends data needed for MyAlerts use | In order to facilitate the use of products like the MyAlerts application, the following data will be provided to PDC from PEGA0066 alerts to generate proper diagnostics information and propose solutions for issues with mobile client connections: pxClientErrorCode pxDeviceManufacturer pxDeviceModel pxOSName pxOSVersion pxHybridClientVersion pxInstallationId pxSynchronizationId pxSynchronizationTimestamp pxEnd The whitelisted parameters can be seen on the Alert Summary page in PDC. |
|
| SR-C69302 | 407178 | Support added for local datacenter only keyspaces for Cassandra | In order to support multi-datacenter system (one Pega system) that spans two data centers but avoids the latency of attempting to connect from a node in data center 1 to a Cassandra node in data center 2, an enhancement has been added by way of the system setting dnode/cassandra_local_datacenter_keyspaces which allows specifying only local datacenter keyspaces as comma separated string. The default behavior is not changed. | |
| SR-C69567 | 409192 | ICMP ping flag enabled and Hazelcast alerts/info messages added | Hazelcast Professional Services recommends enabling ICMP ping as a method to detect network failures more quickly. This can also help to determine whether a failure is due to a network outage: if the ICMP ping is successful but the heartbeat timed out, that may be an indication of a unresponsive (GC pause, long running process, etc.) node. To facilitate this use, the ICMP ping flag can now be used in conjunction with enabling ICMP ping on EC2 through the security group and enabling ICMP in the Hazelcast configuration. In order to facilitate determining root causes for Hazelcast errors in the logs, the following alerts and info messages have been added: Actionable Events (Alerts): Each of these events is worth noting an alert for. An action may be necessary. - [Partition] Lost - A partition was lost. Possible data loss has occurred - [Lifecycle] Merge Failed - A member’s attempt to merge with the cluster has failed - [Migration] Migration Failed - A partition migration has failed Non-Actionable Events (Info): These events are of interest but don’t require any actions. - [Membership] Added - A new member was added to the cluster - [Membership] Removed - An existing member has left the cluster - [Lifecycle] Merging - A member is merging with the cluster - [Migration] Started - A partition migration has started - [Migration] Completed - A partition migration has completed - [Lifecycle] Shutting Down - A member is shutting down |
|
| SR-C70146 | 407968 | Corrected SAML SSO logout error | When performing a SAML SSO Logout, an error appeared indicating some artifacts were missing. This was traced to an incorrect NameQualifier being generated with IDP in the logoff request, and has been fixed by modifying the code to include an SPNameQualifier attribute for the NameID element in the logout request. Namespace has been added for logout requests at the parent level instead adding it at each node element. | |
| SR-C70539 | 412522 | Open API passes 'true' for aUnlockOnCommit parameter to minimize pr_sys_locks table filesize | The Archive/Purge Wizard was not removing the locks acquired during the execution from the pr_sys_locks table, resulting in locks remaining in the table for up to three days. This could potentially cause an enormous filesize and lead to other processes failing to acquire locks due to contention on the table. To resolve this, the system will pass 'true' for aUnlockOnCommit parameter on the open API so that the lock is released on commit of the deferredDelete on purged instances. | |
| SR-C70757 | 411814 | Added null check to resolve error when using datapage as dropdown source | The user application was hanging with the browser console error "Uncaught SyntaxError: Unexpected end of JSON input" when using Dropdown controls with Datapage as a source in conjunction with parameterized datapages to create a cascading second dropdown. This did not happen when using any other data sources for the same control dropdown. This has been resolved by adding a null check at the reloadcell response. | |
| SR-C71037 | 413294 | LegacyToReduxConverter modified to resolve readonly error for hybrid data pages | When a clipboard page or page list was added to data page backed by RD, the data page was becoming read only and not allowing pyRowsSelected to be modified. The issue was in redux design. If a datapage source was configured as a report definition, the resultant page was a flat data provider in redux design. Therefore the pyRowSelected setting was allowed even if the data page was readonly. But when in post load activity a normal page or page list was added to the data page, the structure of the data page did not remain as a flat data provider, it became a hybrid structure and the entire property became readonly and thereafter did not allow setting pyRowSelected to the data page. To resolve this, the class LegacyToReduxConverter has been modified to handle the case when a page or page list is added to flat data providers. | |
| SR-C71043 | 411255 | pzSendEmail returned to previous behavior of sending separate emails to work parties | After upgrade, a change in behavior was seen when using the pzSendEmail activity. When pzSendEmail was called to send emails to work parties, a single email is sent to all recipients instead of separate emails to each recipient as it did in earlier versions. The SendSimpleEmail activity accepts a SendDiscreteMails parameter, the pzSendEmail activity has 'Pass current parameter page' checked so the default value for the SendDiscreteMails parameter was not used. This was a backwards compatibility error, and has been corrected by setting the SendDiscreteMails parameter to true by default in the pzSendEmail activity. | |
| SR-C71253 | 409092 | Resolved issue with PRAuth handoff to SAMLAuth | When attempting to access a Portal with PRAuth, a null pointer exception was being generated instead of the system redirecting to SAMLAuth. This was traced to httpAPI.getAccessGroup() returning null when a DSS authentication/redirectguests/default was set to true, and has been resolved by fetching the access group from the thread instead of the HttpAPI. | |
| SR-C71290 | 410123 | Removed automated cleanup of old DSS entries to resolve memory issue | VBD was issuing an expensive query while attempting to clean up old DSS entries. The old entries were dynamically created with node ID as part of the setting key (vbd/pulse/{nodeid}, and vbd/cluster/{nodeid}). The cleanup logic attempts to get the list of all matching keys to delete by iterating the list of system settings and grabbing the ones that match these patterns. On some systems, there are enough system settings in the DB that it resulted in a memory issue. To avoid this issue, the automated cleanup process has been removed. | |
| SR-C71527 | 411036 | Added ignorecase to correct authorization issue for username in all capitals | It was not possible to delete a case note or attachment while logged in with a user ID containing all capital letters. This has been corrected by changing the equals comparison of .pxCreateOperator and pxRequestor.pyUserIdentifier to equalsIgnoreCase in CanDeleteAttachments access whens, HaveAttachmentDeletePrivileges when rule, pxHaveAttachmentSpecificAccess RUF, and HaveAttachmentSpecificAccess RUF. | |
| SR-C72114 SR-C74858 SR-C73608 |
413598 413604 413625 |
Updates to controls and labels | Issue 413604: When configuring a button with a disable condition, the condition was not returned in the readonly field. The PegaAPI has now been updated to support "disable". Issue 413597 / 413598: While calling the API to get the view of the screen as a JSON format, button type controls were not returning the label of the field using the OOB pxButton control. This has been fixed. Issue 413625: The PegaAPI section was missing when using clipboard page or data page with the rest method. To correct this, support for Embedded page has been added in the pyWorkPage for PegaAPI. |
|
| SR-C72225 | 409998 | Improvements to Services Tracer | When clicking on tracer to get details of a page, a pop-up window opened up with a highlighted link (in blue color) to open the parameter page or the primary page, but clicking on either of these page names opened a blank window without any properties mapping. This was due to the underlying object missing required data because the Services related tracer events did not always have PRThread objects associated with them. The Tracer infrastructure expected the proper PRThread to be in the "ThreadContainer" but offered no option for providing the thread, and the ThreadContainer was not always populated when Services ran. To resolve this, the Tracer Event infrastructure has been refactored to allow specifying a PRThread in RULE-HTML-FRAGMENT TRACERTRACEEVENTSSCRIPTS. The constructor logic in TraceEventBaseImpl has also been modified so child classes can pass a thread instead of relying on ThreadContainer. The Node ID for tracer events will be set based on PREngineProvider instead of ThreadContainer, and the ServiceEvent Interface's methods have been updated to take pointers to PRThreads so that the tracer knows who/what is being traced. In addition, NPE protection has been added to RemoteTracerStub so Tracing Services will not crash the Tracer UI when running from eclipse / prwebpacked. | |
| SR-C72568 | 410838 | Corrected underscores removed from attached filenames | As an inadvertent side effect of work done to handle thumbnail image generation for file attachments with special characters, underscores were being removed from the names of all files attached to work objects. To correct this, the code has been refined to remove special characters from the parameter passed to generate thumbnail activity. Special characters will not be removed from the filename. | |
| SR-C73100 | 411034 | Synchronized block method added to load keystore set with RUF | A recent enhancement allows for the SAML auth service to use a reference to an external file for the signing certificate keystore. When the keystore had the password set using a Rule-Utility-Function, rather than being static, there were intermittent login fails with the error "unable to process SAML WebSSO request1" on the screen, and the RULES logs contained thousands of entries of an error indicating "Password is wrong". It was sometimes possible to log in by starting a new browser sessions and trying again. This was an issue with the PasswordHash property being changed from a static to a non-static field: it was not thread safe, though each object got its own copy of the instance, because if two or more threads call the setPasswordHash() method on the same object, all of these threads tried to simultaneously update the passwordHash instance variable and incorrect results were seen. To correct this, the system will use a synchronized block when loading the entries into cache in the getKey() method - Caller function of KeystoreCacheImpl.java. |
|
| SR-C73131 | 411758 | SAML login rule form documentation updated to reflect usage with IDP | SAML authentication was not working properly during login. This was traced to the assertion not being signed when SAML responses received from IDP require this, and was due to the check box under SP settings not being applied for POST binding. As a result, a local IDP configuration sent a signed response but not a signed assertion. To clarify this, the label message in the SAML rule form has been modified to read "Reject unsigned response and assertion" and the help documentation has been updated to reflect this usage. | |
| SR-C74201 | 412426 | Fixed pyMemo overwriting set values with null | The titles for correspondence and URL attachments were empty due to pyDefault setting pyMemo to "", overwriting the value that was previously set. This was caused by the Link-Object java generation code first setting the properties and later applying the DT. To fix this, the pyDefault DT will be applied for the Link-Attachment instance in activities Work-.AttachURL amd Work-.CorrAttach. | |
| SR-C74954 | 413140 | Fixed column search function for Report Browser Data Explorer | Searching/filtering for columns in the Data Explorer in the Report Browser was not filtering the tabs by the search string. This was an error caused by the unique ID changes for controls, and has been fixed by updating the getSearchTextElement() function in pega_report_events.js to use the name attribute rather than the ID attribute to identify the search bar. The tabClickCallback() has also been updated to not rely on window.event as it was causing discrepancies in Safari. | |
| SR-C75689 | 413136 | restored to correct "pg empty after db.open" errors | When opening the data page D_RepositorySettings, the error "pg empty after db.open" appeared in the logs. This was due to the data page referring to the files webwb/zchecked.gif and webwb/zchecked.gif in many sections of the data page rule form despite these files having been removed as part of a recent cleanup meant to reduce storage requirements. In order to resolve this error, the zunchecked.gif and zchecked.gif images have been added back and the error message generated in the RUF pzGetbase64encodedImagehas been improved to add the target filename: "error occurred while opening the image:webwb/zunchecked.gif" |