In the software factory I defined a report with two parameters. These parameters also exist in the Devexpress file uploaded to this report. After running the report in the Windows GUI the debug info clearly shows the report was fed with the correct parameter values. However apart from the static headers that exist in the design the resulting file remains empty. No data is retrieved from SQL Server. I’ve tried both PDF and XLSX as output with very similar result.When I open the file in Devexpress from the cache folder as mentioned in the debug info and then enter the same parameter values there the report does work. This suggests the GUI does not communicate the parameter values to the report.I haven’t seen a form or task to link a specific parameter in SF to its counterpart in the report file so I assume parameters in SF need to have the same name as those in the report file. Is that correct? If so what could I do to make the GUI produce a complete report?
Our users are logged in through various methods. Some have managed systems, others log in via RDP. Sometimes a user has their files synced with Onedrive, others have a profile folder on a shared drive via redirection. This leads to reports with export to pdf setting being saved in various locations.For a process flow we were able to program one of these file paths by putting a username variable as part of the path but this renders the flow unusable for quite a few of our users. The flow uses a move file action but it often can't find the exported pdf file. Therefor we need a more generic approach.What folder does the export to pdf option use internally to determine where to save a file? I've tried %USERPROFILE%\Documents but already in my own case this leads me to the wrong folder.
Let's say we have an order table containing orders numbered 1, 2, 3 etc. This order number is in the column ‘id’ which is the primary key of the table. These numbers are always integers but they are not an identity. When adding an order to the table a bit of logic needs to determine the next available number. In principle this can be done with max(id) + 1. This works fine in the Windows GUI by putting this logic in a default procedure. It stops working however when an order with the same id is added by someone or something else simultaneously.Say the highest order id is 10. Then I start creating a new order in the GUI which by the aforementioned logic gets order id 11.In the mean time another process inserts order 11 into the table. Then when I click the save button the GUI tries to save the new order with id 11 which is not available anymore and an error message follows.I've tried to get around this by using an insert handler with the same max(id) + 1 logic. This eliminates the error
It would be convenient to have a dropdown and autocomplete inside filter form fields when the field has an assigned lookup. This would make it easier to fill in a long value correctly and would make ‘equal’ searches more feasible.
We have lots of SQL sessions containing the tsf_user procedure taking up lots of CPU time. This only happens when tsf_user is called via Indicium. We can't really put the finger on why this is happening. What could be the cause? Today this even caused SQL server to lock up entirely.CPU time total and per session:
The Windows GUI produces empty lists in some screens when querying the database via Indicium Basic. In one of these cases the view being queried loads quite some time and then throws a 502 error (verified in Firefox). In another one loading is quick but the server gives a blank answer. In both cases the GUI translates this into an empty list. Both cases are views. Here’s a URL as an example:http://servername/indicium_windows/iam/184/vw_webshop_xml_07_persons?$filter=((bedrijf_nummer%20eq%201)%20and%20(klant_groep_nummer%20eq%2016))I don't see anything particularly weird about this URL.Any ideas what could be the cause?When connecting directly with SQL Server the GUI works fine. However we need to be able to use the end-product off-premise and in some cases the Windows GUI is the preferred method.
Often all we need for a badge is the number of rows in a grid. Instead of having to write a template for every new badge it would be a time saver to have a badge that does the same as a count aggregation underneath a grid.
When running the software factory in the Windows GUI via Indicium basic I get an error message saying "Object reference not set to an instance of an object” when going to the result tab of a control procedure. Then when clicking the generate definition tab a message saying "The provider 'Indicium' does not support tooling connections!” pops up.The alternative of course is a classic database connection but that generates quite a bit more traffic and is also slower. This is especially noticeable when connection latency is high due to long distance to the server. So what actually is the status of accessing the software factory via Indicium (basic or universal)?
After the platform upgrade to version 2023.1 I noticed that scheduled system flows in IAM were disabled. Until now I just assumed this was due to the platform upgrade. I re-enabled them and didn't give it a second look. But today reports came in that documents were missing in the end-product. It turned out that the flow schedules were disabled again. I synced to IAM this morning. Does that mean every sync to IAM will disable auto running system flows from now on?
In a new barcode scanning screen we’re constantly adding items to a list by scanning many barcodes in a barcode view. This barcode view receives updates at every scanned item which results in an insert statement to a table via an instead of update trigger on the view. The effect is that every action with the barcode scanner results in a new item in the list. New items are added from the top of the list.The problem is that as soon as the selected row reaches the bottom of the grid it stops going further down at consecutive scans. New items are still added but they aren’t visible without scrolling up. Because we want to minimize mouse actions during the scanning process to maintain workflow this effectively keeps new items invisible.This issue was already described in https://community.thinkwisesoftware.com/questions-conversations-78/barcode-input-2822 where an animated gif illustrating the issue can be found as well.The solution I would like to propose is an option in subjects to allow
We recently started using temporal tables for example to retrieve an address as it was at a certain time in the past. However none of the ‘_history’ tables underneath their regular respective temporal tables get any users or roles appointed to them. For example our ‘adres’ table contains 7 roles whereas the ‘adres_history’ table has none.This is a problem because we want to use this for invoices for example where the invoice retrieves the address as it was on the invoice date. This is vital for invoice retention as is dictated by law. Temporal tables are a great alternative to having to print out all documents to PDF and keeping them in a backup.I can of course apply these roles manually and that’s probably what I have to do if this is intended behavior. I would like to read some other thoughts on this though before I proceed.
Something seems to have gone wrong in a temporal table. There’s a gap between the current record and the previous record that’s now in the history table. This gap looks like this:Query:select *, tsf_valid_from, tsf_valid_tofrom adres for system_time allwhere adres_id = 51714It’s as if the history data is no longer complete but we never manually deleted this data. This is a problem because we use this data for historical records such as invoices.What could cause such a gap in temporal data?
Here’s a very simple question I somehow cannot figure out. How does one authenticate a connection to an off-premise SQL Server instance such as Azure in software factory? I’ve added a runtime configuration for the Azure database but there is no means here to store a username and password.I assume the current runtime configurations pick up the domain user credentials to connect to on-premise databases but this obviously doesn’t work with Azure databases.
Does Indicium run well in a Linux environment? I haven't tried Mono yet but apparently it is capable of running .net core applications. Is the current version of Indicium ready for it?
For one of our reports we prefill temp tables for performance. Without pre-generated temp data the report is unusable. As part of a larger process flow this is no issue. However because the Thinkwise platform doesn’t run stored procedures inside a REPX file this report cannot be implemented standalone.The solution for now is a new process flow for this one report. Drawback is that in the GUI the report button then becomes a task button. This is because a new table task to fill the temp tables needs to set the process flow in motion that then generates the report.Preceding process actions could be a more elegant solution here. I would like to be able to perform a GUI action (e.g. start a table report or table task) and have the process flow postpone the starting process action and perform another process action (e.g. a stored procedure) first. In this case the table report would be the starting process action inside the flow and the stored procedure to fill the temp tables would be the
I am trying to start a system flow via an API call to Indicium. The API URL works. When I visit it with a browser I get to log in and see the application version number. However the same URL followed with / and the system flow name gives a 404 Not Found in Insomnia. The system flow exists in IAM since I can find it there.Browser: http://[server]/indicium_universal_prod/iam/[alias]Insomnia: http://[server]/indicium_universal_prod/iam/[alias]/documenten_aanmakenWhy doesn’t the API call work?
When generating a report in a system flow and using the report data stream as input for a write file connector the resulting PDF file produces an invalid colorspace error when opening in Acrobat Reader.I’ve already tried several settings such as ANSI encoding in the write file connector without preamble. Notepad++ confirms that the file is indeed saved in ANSI. The file starts with a readable PDF header just like any other PDF file.The same report produces working PDF files in another regular (non-system) process flow.How can I produce working PDF files in a system flow?
Reports developed in Devexpress tend to run rather slow in the Windows GUI. Even when Devexpress itself generates a report within a second it can still take over 10 seconds for the GUI to do the same thing with the same report parameters. I’ve seen this happen in several occasions. Now it’s an issue because for a new module for our warehouse printing a specific report is rather time critical. A delayed print tends to interrupt the workflow too much.The desired behavior would be to have a printout of a report within a couple seconds instead of the current 10-15 seconds. A workaround for now is to make it possible to set the print command in motion earlier in the process but that is not a solution in all situations. It also makes the chances of human error a bit larger.Is there a way to find out why reports perform slow in the Windows GUI?
When a target table as detail has the same primary key as its source table the Windows GUI automatically hides the grid even when a grid is part of the screen design. The Universal GUI doesn't display the same behavior and always shows the grid even when the data model directly implies there will only ever be maximum one detail record.I would like to propose an optional setting for the Universal GUI to copy this behavior from its Windows sibling. This would take away the need to comb through all screens and possibly having to redesign many, and allow for an easier transition to the Universal GUI.
When a target table as detail has the same primary key as its source table the Windows GUI automatically hides the grid even when a grid is part of the screen design. The Universal GUI doesn't display the same behavior and always shows the grid even when the data model directly implies there will only ever be maximum one detail record.Is it possible to add a setting to the Universal GUI to copy this behavior from its Windows sibling? This would take away the need to redesign lots of screens.
Our Windows application has many carefully designed screens to optimize usability and screen estate usage. Our users are now very much used to the way the application looks and behaves. Because web apps seem to be the future I have tried the Universal GUI several times these past few years. Of course most development went into functionality which is understandable. But when will the interface itself get more options? It doesn’t have to look exactly the same as the Windows application but it would be nice if all our designed screens could at least be ported to some kind of Windows theme.It’s great to have a web app that’s able to adapt to smaller screens but for office use Universal still appears rather basic. Perhaps style sheet examples could be provided to tweak the Universal look and behavior towards a Windows look?
I am building a screen in which a barcode can be pasted and entered with a barcode scanner. For this purpose I’ve created a view with only its form visible and in it only a barcode field. This field should be editable multiple times without having to click any GUI buttons. According to the view itself these are updates of the barcode field. Behind the view is an instead of update trigger that adds the new data to a table. This works. As soon as a barcode is typed into the field and ctrl+enter’d a new row is added to the table.The problem is the form isn’t editable immediately despite having auto-edit enabled in its subject settings. I need to click in another tab in the gui and then return to the view to have its form editable. This seems to be the standard behavior of auto-edit in tables too.I’ve done some testing with combinations in a process flow to make the form editable without needing any user input between barcode updates but so far I’ve been unsuccessful to make this work. For
The filter bar searches for exact matches with substrings of searchable values in the table (where value like ‘%search%’). In most cases that is the desired behavior. However in one particular screen we would like to be able to insert the id of a record and only show that one record (where value = ‘search’). The reason for that is it greatly enhances performance. If a quick filter is applied also the same select statement goes down from 1529 ms to 3 ms. This means a refresh of the entire screen would take 1,5 seconds less than it does at the moment. Refreshes are frequent in this screen.May I suggest a new setting for a filter bar to make this possible? This way a search would perform just a well as a quick filter.
This is the first time I’ve tried a subflow. I made sure the flow parameters are fed with the values from the main process flow. However it doesn’t run.When I check debug info in the gui it says: “ExecuteSubFlow is not supported in the Windows or Web gui”Does this mean subflows don’t work at all? We’re on the latest 2022.1 version.
We recently upgraded the Thinkwise platform to the latest 2022.1 version. I just tried the very first creation process with it but it doesn’t seem to be doing a smart upgrade. As you can see in the screenshot I did choose the smart option:However this seems to produce code as if it was a full upgrade. The many code steps below indicate this as well. Before this list was often a lot shorter when doing smart upgrades: I don’t want to do a full upgrade especially when I repeat the process on the production database.I’ve already regenerated the definition of the previous version but that doesn’t seem to have an effect.What else can I try here?
Already have an account? Login
Enter your username or e-mail address. We'll send you an e-mail with instructions to reset your password.
Sorry, we're still checking this file's contents to make sure it's safe to download. Please try again in a few minutes.
Sorry, our virus scanner detected that this file isn't safe to download.
