eRAD
PACS Server
eRAD
PACS Server
The Series Number value was misinterpreted as a hexadecimal value and erroneously converted to a decimal value on the Tech View page. This has been resolved.
When locking a study, the worklist’s lock indicator displayed in red, indicating the study is locked by another user, rather than yellow, indicating you locked it. This has been resolved.
Duplicate names are rejected when creating and editing device categories.
The hyperlink used to download desktop apps did not apply custom port configuration settings, preventing users from installing these services. This has been resolved.
Displaying the Server Information page from a worklist server took a long time to collect the study volume information when it hadn’t been collected recently and when large log files existed. The process now collects baseline information during an off-hour cron job and the web page uses this information rather than collecting it on-demand from the log files. This has been resolved.
Patient matching criteria set to use subsets (e.g., the first three characters) from more than two name components failed because the resulting database query didn’t account for trailing data in the name values. This has been resolved.
When a web services command is using the default web services user account, the missing credentials was interpreted as an error, resulting in a failed login log entry for every command. This has been resolved.
Multiple failed login requests that occur in the same second resulted in a single error log file entry because the log entry ID value used seconds and would not be unique. This has been resolved.
A recent bug fix intending to verify a worklist filter satisfies the query qualifier’s unbounded query criteria failed to work on newly created filters because it attempted to use the filter by name before it was saved. This has been resolved.
When a delete action is defined to keep the report and execute the purge (as opposed to simply mark eligible studies, a feature used internally), the rule set is empty and the purge action would not run. This has been resolved.
The original, jetty-based streaming technology used by Evo viewers and the web viewer failed to release websockets when the study being downloaded changed and the underlying thread encountered an exception when terminating. This has been resolved.
Renaming old folders failed because the check for an existing folder with the same name referenced the wrong field. This has been resolved.
If a user with a custom default worklist filter created a folder, it disappeared from the folders and forward page after an apache restart because all folders with list description tags are moved to the list section of the user’s configuration file during restart. Note that the folders need to be recreated, but the content in those folders is unaffected.
When importing data from DICOM media, failure to create a DICOMDIR file due to invalid data on the media resulted in a silent failure. Now the user receives a notification indicating the media could not be imported.
Web services commands to edit a report issued to a worklist server in a multi-level dotcom failed because the RMI command sent to the child server did not use the RMI-specific session mechanism, resulting in permission failures. This has been resolved.
When a log or archive page query results are truncated because the number of total matches exceeds the applied limits, a partial results warning is displayed but only when there are more results to display. The warning was missing on the last page of the results. This has been resolved.
The viewer failed to display the notice informing a user that they lost the study lock because of a formatting error in the message the server sent to the viewer. This has been resolved.
If the user issues a quick search after editing the layout of the worklist, the results might appear using the original layout and not the modified layout settings. This has been resolved.
If in the process of configuring a compound worklist one of the source lists got deleted, the system failed to notice and saved the compound list, resulting in an exception when applying it. This has been resolved.
The rename and delete options on the Other Lists page are disabled for group lists unless the user has the required admin rights.
Non-urgent broadcast messages did not record the last update time and therefore would not appear in the message panel or dashlet. Additionally, when the only messages in the queue were non-urgent broadcast messages, the missing timestamp caused inconsistent list details, resulting in a logged exception. This has been resolved.
After creating and then editing an order via web services commands, some fields would not appear updated on the worklist because group order handling treated them as new orders (in the group) rather than modifications of the existing order. Edits that apply to an existing procedure step, based on the procedure step ID, are now treated as edits and not new orders in a group order.
The media creation server (MCS) always used the primary IP address of the server when constructing the download URL. If the certificate does not include this IP address, an error occurs when attempting to download the media file. A configurable server address and redirection rules have been added to accommodate environments running under these conditions. See Media_Creation_Server
Maintaining a list of studies open in web viewers was not thread safe. Simultaneous requests to load studies in a web viewer corrupted the list, causing an exception to subsequent web viewer open requests and a 104 error page displayed in the browser window. This has been resolved.
The error message displayed when attempting to log in with a non-existent user ID provided too much details about the login failure. The updated message is more generic and secure.
When upgrading from v7 to v8, worklist layouts that contain no customizations were inserted into the database as NULL entries, resulting in an exception when applied. Empty layouts are converted during upgrade and existing layouts are detected and handled correctly.
All logs pages consisting of more than one page of records indicated the results were “partial”, even when this wasn’t the case. This message should only appear on the logs page when collecting logs from other servers in a dotcom.
Web services calls were not updating the session, allowing the session to expire sooner than the inactivity timer specifies.
While taskd was stopped, order updates received via HL7 or web services were treated as new orders, resulting in duplicate orders in the PACS worklist. This has been resolved.
The object-matching settings for a Purge action did not list the configured attachment types, meaning users could not set up an action to purge select attachments. Additionally, the system failed to check the attachments type when matching attachments to the purge criteria. This has been resolved.
Filter macros using the format $$<COLID>$$ were not recognized, causing the filter criteria to return no results. The literal filter value defined appears in the filter criteria panel but the resolved macro value appears in the filter box at the top of the results. This has been resolved.
Web pages saved as Favorites by users with access to them could be viewed by other users who do not have the necessary permissions. Now, if a restricted page is requested, the user’s default start page appears. Note that data restrictions, if applicable, were enforced correctly.
Editing the system’s (user account) auto-unlock timeout setting attempted to update the value for each user account across all servers in a dotcom. The unlock time has changed so one value is applied to all user and group accounts regardless of the server from which the user logs on, allowing the system to propagate a single system setting instead.
Some user and group settings fields did not enforce value length limits, causing unintended truncation when saving. All input fields now enforce the field length limit.
Some system and prefetch sessions were checked for inactivity even though these sessions never time out. If a study matching an order arrived after the inactivity timer expired, the auto-correction failed to complete. This has been resolved.
The auto-correction log entry was missing a space, combining text and a study UID value. This has been resolved.
If the user edited the Radiologist field on the report panel and the value is empty or does not match a defined radiologist’s name, an unnecessary exception pops up when the user saves the report. This has been resolved.
If there is no license or there’s an error in the license, the server unintentionally initializes as a media creation server, whose built-in defaults might be different that those of a PACS server. This has been resolved.
After deleting a default worklist label from the list menu, the group or system default list should appear but did not because the list was not refreshed. This has been resolved.
The button controls on some curtain pages failed to close the popup panel, leaving the popup panel
visible after the operation was completed or canceled.
When replicating a study from the consolidation server, the target folder was created but the redirection request failed due to a call to a retired page. This has been resolved.
Login requests from web services calls or when reading location was specified failed to check the lock state of the user account.
Actions assigned to groups whose ID was 20 characters were ignored because the last character was truncated from the identifier saved in the action database table, meaning it didn’t match any registered group when actions ran. This has been resolved.
The recent fix to handle double quotes in patient matching criteria fields introduced a parsing bug when patient matching including initial matching parameters. This has been resolved.
Actions, with the exception of the Notify Action, cannot be assigned to group accounts, but the user interface failed to enforce this restriction. Existing actions assigned to group accounts disabled at first restart. Admins are notified if any Actions are disabled.
Spaces are not permitted in group IDs, but this was not enforced when creating new groups or when importing groups during upgrades. Existing group IDs containing spaces will be modified by replacing the space with an underscore.
Some remaining web pages were not defined to use UTF-8 by default. In this specific case, flat report text containing accented characters was not handled correctly when sent via HL7. This has been resolved.
After moving list management to the database, Action study lists were not sorted, which could result in processing a study multiple times and, less likely, missing a study completely. This has been resolved.
Displaying a list’s properties from the Other Lists page had some formatting issues that truncated the filter’s details. This has been resolved.
When deleting a folder, the system removed the container but failed to check for and remove study references assigned to the folder. If a folder is not empty when deleted, the user is prompted before the study reference is removed. This has been resolved.
When creating media with two or more studies that do not yet exist on the media creation server, the download of one might be delayed. For example, it might need to be retrieved from an archive, or it is still propagating up a dotcom from a child server. The system was not always waiting for all studies to arrive before building the media file. This has been resolved.
If a web services API link could not complete, the response was hard coded to report an authorization failure. Now it prompts the user for credentials.
If prompted to enter user credentials after issuing a web services API link without a valid session ID, a failed first attempt lost the intended API target (page) and the user ended up on an unexpected web page. This has been resolved.
If a field used in the patient matching criteria contained double quotes, the patient folder panel would not open because the system failed to escape the met character. This has been resolved.
UPGRADE NOTICE: This fix does not remove existing system default worklists. If a system default worklist was created as a result of this bug, it must be deleted manually. The factory default worklist was incorrectly saved as the custom system-level default if no system default worklist existed.
Servers configured with a database set up for latin-2, or any character set other than latin-1, mishandled some collation details, resulting in a failure to manage some saved worklists, including deleting them. This has been resolved.
When moving worklists and actions to the database (medsrv-8.0.21), empty action study lists were interpreted as a disabled action. As a result, after the upgrade, the action was disabled. Note that this fix does not restore actions disabled by the bug. These actions must be re-enabled manually.
The database length of group list names failed to account for a character appended by the system software, causing errors when using a group list whose label was exactly 20 characters long. This has been resolved.
After moving saved worklists to the database, worklist actions failed to remove non-matching studies from the managed list of studies, causing the database table to grow large. This has been resolved.
Resetting the auto unlock time setting unintentionally changed the lock period for explicitly-locked accounts, allowing them to be reset by the user. This has been resolved.
Some dependencies were not cleared when a saved worklist filter was deleted or renamed, including removing Actions and updating compound lists. This has been resolved.
The saved worklist delete button was present in the list menu for users without the necessary (Create View) permission. This has been resolved.
The password characteristics were enforced when updating a password but not when creating the user account. This has been resolved.
The hierarchy used to select the data loaded into the patient folder’s view area by default was incorrect. Displaying a final report has precedence over an attachment. This has been resolved.
The application that draws annotations on key images on a report page could fail when applying the annotation. While the image and unaffected annotations appear in the report page, some annotations might be missing. This has been resolved.
Worklists including conditional row colors formed an invalid query request, causing non-matching entries to appear in the results. This has been resolved.
Some secondary capture object references, specifically for attachments, were hardcoded to suppress tagging, yet the thumbnail references followed the viewer’s preference. When the viewer requested tagged references, the inconsistency could cause the viewer to hang. This has been resolved.
An inherited page size limit used when collecting the objects for an ISO file truncated the list, resulting in missing objects on the media. This has been resolved.
An incorrect permission check permitted users with the right to approve a report to open it for editing even when they had no report editing permission, displaying an error warning in the process. This has been resolved.
If a retired database field, Scheduled Order Number, contained data, attachment uploaded to orders would fail. This has been resolved.
The database index has been restored to the study table’s FOLDER field, and a default statistics file has been added.
An issue with the way calls were made to a handler prevented some functions from completing, including the one displaying key images and another hiding the image size selector on a print-friendly report. This has been resolved.
Support for displaying content from third party SR objects on a report page or panel has been ported to v8.
Restrictions have been applied to disable LDAP’s support for password-less login attempts.
Studies with presentation states created by the eRAD viewer encountered an error when the MCS converted those objects to DICOM-compliant PS objects and copied them to the output folder. This has been resolved.
The image tagging solution the viewer uses to request the server reprocess an image object interfered with displaying annotations on key images and presentation states in some versions of the viewer. The solution was to disable tagging unless the viewer version indicates it supports the feature.
When a presentation state object was missing from the processed directory, the web viewer failed to initiate the just-in-time processing and then didn’t handle the missing file, preventing the whole study from loading in the web viewer. This has been resolved.
When saving a worklist filter, the query qualifier was not checked, allowing a user to save a poorly defined or unbounded query filter. This has been resolved.
The system failed to create a log entry when a non-image file (e.g., a document) was uploaded and attached to an order. This has been resolved.
A change to system user names got around a check meant to prevent system users from breaking locks. As a result, some system-initiated commands, such as study split, study merge, report saving and ownership passing, could be invoked multiple times, duplicating data. This has been resolved.
File uploads initiated at a worklist server and redirected to a hub server completed successfully but logged an error when it attempted to parse the command to see if further redirection was necessary. This has been resolved.
The delay and retry task information displayed on the task management page did not match the data in the database because some tasks weren’t removed from the Info Object tree. This has been resolved.
System and group default lists for the user accounts page and logs page were not being applied. This has been resolved.
The Author Observer (report author) name was saved using HTML encoding, which was inconsistent with all other names in the report. This has been resolved.
Access to system worklists are supposed to require Admin permissions but this restriction was not applied, making them visible to all users. This has been resolved.
Applying a saved, default Archive page filter containing more than one filter would not issue the query and no list was presented. Additionally, the system default Archive page filter contained a null filter value, resulting in an error and no list was presented. This has been resolved.
After saving a worklist filter with a text field using “is not empty”, applying the filter and then opening the filter criteria field to edit it, the contents contained an unwanted character that, if not cleared, would be applied when the query was run. This has been resolved.
Worklist filters created in previous versions and updated during the upgrade to PACS Evolution do not contain an explicit setting to show or hide the worklist header. They are supposed to show the header by default, but were not. This has been resolved.
Requests to create DICOM media using the MCS service failed if issued from a worklist consolidation server (a parent to a collection of worklist servers) because the software only collected data from neighboring hub devices. This has been resolved.
If a saved worklist filter results in a query qualifier warning, the resulting list layout ignores the saved filter’s layout and applies the default layout. This has been resolved.
The DICOM toolkit inherits character sets differently when editing and creating objects. When creating a new report object, the default (ASCII) character set was used, causing names such as the reading physician to ignore accented characters. This has been resolved.
The Prefetch and Stop Sending Updates action icons could not be filtered on from the Other Lists page. This has been resolved.
If the worklist filter configuration file was created as a favorite (starred) worklist, it would remain in the dashlet panel after it was purged from the system. This has been resolved.
A change in the way browsers handle iframes loaded from a different server ignored the cookie presented and reported a non-existent or expired session. As a result, loading a report in the web viewer’s report panel displayed an expired session error message instead. This has been resolved.
When attempting to update the study data from the document scanning page, the system attempted to save some retired scanning settings, causing the request to fail. This has been resolved.
Some browsers won’t pass along an access control parameter when redirecting web viewer open requests to a child server. As a result, the web viewer reported an error and would not load the data. This has been resolved.
Attempts to filter the Other Lists table returned no matches because the filter mechanism overhaul completed for v8 had not been applied to the action data. This has been resolved.
Outbound web service messages were sent twice because the event handler was registered twice during startup. This has been resolved.
When the viewer’s patient folder checked for study updates, the result was compared to the wrong data causing a refresh every time. This has been resolved.
The web service command to delete DICOM media files after they have been downloaded failed to pass along the media session ID, causing the request to fail and the data to remain until the system purged it when the timeout expired. This has been resolved.
User-scheduled tasks (from the GUI) resulted in a deadlock condition when checking for new tasks from the unprocessed queue occurred before first checking if the system can start new tasks. This has been resolved.
Support for customizing the default filter type for worklist fields has been added. See Customize Labels Settings
Some DICOM server and study operation settings appeared editable, but weren’t, for admin users who didn’t have the required support rights. This has been resolved.
When exporting a worklist, the system expanded the DICOM enumerated value for sex and then attempted to localize that word to the applied language. The translation effort failed, causing the export error. The DICOM value for sex is included in the exported data without expansion or translation.
When LDAP is not configured, the security settings page hid a field. When the user edited a security setting, the hidden field was checked and since it was empty, an error result was displayed, even though the edit was correctly applied. This has been resolved.
The frequency with which the server attempts to renew its license has changed to avoid creating a processing burden.
Attempting to save a worklist layout did not offer the current user as an option in the list owner menu if the currently applied layout is a system-owned (default) layout. This has been resolved.
The heading on the user account batch edit panel was incorrect. The panel was updated to display the list of all user accounts selected for the batch edit.
The Storage AE and Archive Query AE fields on the settings page could be edited by Admins who didn’t have support rights.
A change in the way browsers handle iframes loaded from a different server ignored the cookie presented and reported a non-existent or expired session. As a result, loading a report from a child server when batch approving on a worklist server failed.
When launching the web viewer without an active user session, the user is prompted for credentials. While the login succeeded and returned a cookie with the necessary session ID, subsequent calls did not use it, causing them to do nothing or repeatedly prompt for credentials. This has been resolved.
If there is no number in the input string, the calculated field’s number() macro is supposed to return ‘0’. Instead, it returned the input string. This has been resolved.
Updates to report printing necessary for supporting wkhtmltopdf failed to include the title value when converting from PDF to postscript. This has been resolved.
Notification actions using the system’s built-in email template failed to generate email messages because two macros used by the template were removed in medsrv-8.0.10. This has been resolved.
The Safari browser does not support wildcard values in the access control origin header. The wildcard value has been replaced with the referrer host value.
Opening multiple studies in the web viewer when web sockets is unavailable, causing the web viewer to downgrade image download to use HTTP, texture blocks were resized when they shouldn’t be and data chunks could override each other causing missing data blocks. This has been resolved.
Some document type settings (default document type, modality and study date) were not applied to the scanning and upload user interfaces when the panel initialized. This has been resolved.
Log entries showing a change in plugin rights failed to identify the original permission setting. This has been resolved.
Some empty date fields, including Scheduled Date, were not assigned the default date, causing a java exception and an empty worklist. Additionally, searching on an empty date value returns all records with no date value rather than all records, as was the case before. This has been resolved.
A change in the WSDL moved the location of a file that was needed when sending orders and reports to web services devices. This has been resolved.
The filter type value in saved worklist filters was dropped when upgrading m8.0.8 systems, causing worklist filters to ignore the filter. This fix does not correct saved filters that have already been converted during an upgrade. This has been resolved.
A parsing error prevented the viewer from displaying the customized list of available speech recognition options if more than one option was included in the configuration file. This has been resolved.
Patient name and other field queries containing an apostrophe (') in the search criteria returned zero matches regardless of the data. This has been resolved.
The query qualification check was not activated when loading a saved worklist. This has been resolved.
The study date value stored in the database's log table and displayed on the log page did not include the time component. This has been resolved.
Unsuccessful requests to add or edit a device did not always notify the user the operation failed. This has been resolved.
Image-type attachments uploaded to an order from a worklist server are supposed to be stored on the hub server. It also changes the order to a study. All other attachment types are supposed to be stored on the worklist server. (They move to the hub server when the image objects arrive.) These attachments were mistakenly stored on a hub server causing a broken study if the images arrived at a different hub server. This has been resolved.
If a web services client creates a new user account set to change the password upon first login, the user was not required to change his/her password. This has been resolved.
When the same objects is resent multiple times and the queues are not empty, taskd might start multiple registration tasks for the same object at the same time. The object was successfully registered but errors were logged and one of the tasks might end up on the retry queue without the ability to complete when rescheduled.
After adding outbound notifications when modifying a study's instance UID (eg, after splitting or merging studies), error handling changed and releasing the lock upon error was unintentionally dropped. This has been resolved.
When exporting annotations saved as a presentation state, the server converts the internal representation of the data to DICOM Presentation State objects. Support for rotated ellipses and freehand ROI graphics was not included and these graphics would not be exported. This has been resolved.
Inserting a segment into a report template from the template editor resulted in a java crash. This has been resolved.
This bug prevented admins from customizing the viewer's patient folder template. This has been resolved.
System unable to support more than 32k users
File system limitations prevented the system from having more than 32K user accounts. This has been resolved.
When a series- or instance-level C-Move request was issued to a non-objectless consolidation (CWL) server, the resulting database query was invalid and failed. This has been resolved.
Web services calls to retrieve study information returned incorrect image and series counts if the number was greater than 15. This has been resolved.
The tool to assign filter criteria to a Q/R device was missing some features: warning about the effects of the filter; access to all database fields; and logging edits to a filter. Some presentation inconsistencies were also corrected.
When a spatial registration object is registered for a study currently open in an active viewer session, the server notifies the viewer so it can retrieve the new object.
Internet Explorer issues an incorrect event when dragging a scroll bar in a dropdown box (in this case, the canned report selection list), causing the dropdown list to close before selecting an item. This has been resolved.
Scheduled Procedure Step ID values defined in an order creation message failed to show up in the database (and worklist). This has been resolved.
The study edit page failed to list values for fields in the DICOM object but not stored in the database. This has been resolved.
The study forward page would take a long time to load when the system contained many (tens of thousands) user accounts. This has been resolved.
The edit report page, launched from the patient folder, would take a long time to load when the system contained many (tens of thousands) user accounts. This has been resolved.
The report approval page would take a long time to load when the system contained many (tens of thousands) user accounts. This has been resolved.
The Message dashlet's recipient list and the dashlet's configuration page would take a long time to load when the system contained many (tens of thousands) user accounts. This has been resolved.
Auto-retrieving priors from the archive to the hub server failed if the study was registered on a child server. This has been resolved.
Locked studies would not unlock after the lock period expired if the user's session had timed out first. This has been resolved.
New user accounts using IDs of deleted user accounts would be notified of multi-level authentication requirements even when the feature was disabled. This has been resolved.
Displaying the report approval web page was very slow on systems with a large number of user accounts due to an inefficient database query. This has been resolved.
The option to set a report to Final was unavailable if the user had permission (the Approve right) to approve a report. This has been resolved.
Objects that arrive at a server after the post process or compress data actions have been applied to the study are automatically processed.
If an archive was an independent eRAD PACS system, the retrieve and forward tools on the Archive page failed because the archive expected a user session ID from it's own dotcom. This has been resolved.
Web services requests to create DICOM media performed an invalid check of the requested compression and media type parameters, thereby ignoring the requested settings and always uncompressing the data. This has been resolved.
Displaying a report of external origin using a template that does not contain an areagroup for handling external reports resulted in a crash. This has been resolved.
Secondary capture and presentation state objects sent to the worklist server after being created in a viewer session were propagated to their owner device but remained on the worklist server. This has been resolved.
When the viewer sends new objects to the server when it also sends a report, the study's State Changed value was updated to indicate the study changed. This has been resolved.
The viewer's patient folder study list was misaligned due to a position calculation error when collecting the data, and neglecting to account for the separator spaces in the connector bar. This has been resolved.
If the patient folder is configured to automatically load data into the view area, and the area is already populated when the patient folder is opened, the background refresh changes the data already displayed in the view area. This has been resolved.
Studies that are no longer included in the filtered worklist were checked for updates when the page refreshed, causing excessive log entries because no lock state existed.
When using Internet Explorer 11, the Rights section of the user account edit page extended outside the boundaries of the browser window, with no scroll bar to reposition the options. This has been resolved.
Uploading a non-image attachment to an order from a worklist server didn't push the data to the selected hub server. As a result, the data was not visible on the Tech View page or in the patient folder.
Web services commands issued to a server and forwarded to an owner device for execution failed the session validation if the web service user was not the registered default user for the device issuing the call. This has been resolved.
Correction messages that propagate up a dotcom to the server that is configured to send a notice to a web services client (eg, to a RIS) before the image objects arrive and get registered will fail, but the task did not get added to the retry queue. As a result, the client never received the correction notification. This has been resolved.
Workstation prefetch requests resulting from a Forward Action didn't include the instruction to expand the list and therefore failed to include relevant prior studies. This has been resolved.
A recent change to handling date and time filters affected the birth date field, causing the exact match needed to grant access to restricted studies to fail. This has been resolved.
When opening a PDF attachment from the Technologist View page in a multilevel dotcom, a java dump appeared on the screen because the redirection failed to include the session ID. This has been resolved.
When a presentation state object specified an annotation on an image that was not present in the data set, the web viewer would fail to load the study or report an error. This has been resolved.
The units for the media export download timeout setting (on the Server Settings/Study Operation page) have changed from days to hours. Existing settings are automatically adjusted.