* The ErrorBoundary is super powerful and useful. There is a default implementation that seems to be recommended now, so dropping our homegrown one for the standard one.
We can now render custom components when an error happens within an error boundary, and we can use error boundaries within sub-components as we now do in the reactFormBuilder which will capture form rendering errors, and allow you to fix the error and retry.
The more global ErrorBoundary set in the "ContainerForExtensions" now users a the ErrorBoundaryFallack to render the error - which looks a little cleaner, and tries to offer a little more information about what went wrong.
* Detect focus loss/return in the Diagram Editor - so that we can reload the process model and assure that we know about any changed files.
Allow for looking specifically for json SCHEMA files (those files that are named -schema.json or .schema.json (as is the convention). Only show these in the dropdown for the form.
* * Run descriptions through the markdown processor so you can use bold/italic etc... in your description fields within a form.
* Move ExampleTable into it's own view component to keep the size of the form builder sane.
* Assure markdown within jrsf forms have reasonable styling that follows the containers style, rather than setting to some other default.
* Add a couple of example forms so people can get a sense of what is possible.
* Connect up the new Json Schema Editor Component to the process model edit diagram.
* Just select the schema file - not the ui file when selecting the form for a component - we may revert this to just a text box.
*
* Cleanup the formatting of arrays, so that they are sligtly intended, do not contain an awkward unneeded heading, and have some tighter css.
* Connect the form editing in the modal back to the BPMN-JS editor
Auto-Save edits in the Form Builder
Lots and lots of tweaks to the react form builder ui
* various fixes.
* test for prepare_schema
* minor fix for run_pyl
* css cleanup
less issues with reloading and jumping about when in the editor
Don't sort keys when returning the json.
More intelligent "ready"
* bump package to point to branch of bpmn-js-spiffworkflow so others can check it out.
* Assure that json keys are not sorted during serialization by default.
Allow adding example fields to an existing schema
Create a set of examples.
* db complaints in migration change.
* removed items from interface file that had been moved elsewhere w/ burnettk
* rename prepare_form to prepare-form
* rename prepare_form to prepare-form
* Remove commented out code.
* typo
* add a comment about the empty column
* move back to the main branch
---------
Co-authored-by: jasquat <jasquat@users.noreply.github.com>
Co-authored-by: burnettk <burnettk@users.noreply.github.com>
* upgrade to correct multi-instance error bug in BPMN-JS-Spiffworkflow
* update the git service to re-build cache after doing a git pull when a web-hook is called.
* update to SpiffWorkflow - which includes 'replace root with spec start'
* updating the lock file.
* bump spiff to change task internal data, add new method and new exceptions
* bump bpmn-js-spiffworkflow for MI updates
---------
Co-authored-by: burnettk <burnettk@users.noreply.github.com>
Co-authored-by: jasquat <jasquat@users.noreply.github.com>
* Revert "Revert "Feature/event payloads (#393)""
This reverts commit 95fafb7af1.
* Revert "Revert "poet not available in container""
This reverts commit 140220498c.
* Revert "Revert "Run event payloads data migration from background processor (#399)""
This reverts commit 2afced3a51.
* Revert "Revert "using new spiff api to get info about events. w/ elizabeth""
This reverts commit af857fee22.
* Revert "Revert "fix tests for waiting_event_can_be_skipped""
This reverts commit 886e6bd42a.
* push image for preview env
* default scripts to localhost w/ burnettk
* use the bugfix/update-split-task-inputs spiffworkflow branch w/ burnettk
* removed debug json files
* use main for spiffworkflow
* do not attempt to highlight non-diagram boundary items w/ burnettk
* updated SpiffWorkflow to fix multiple signal event issue w/ burnettk
---------
Co-authored-by: burnettk <burnettk@users.noreply.github.com>
Co-authored-by: jasquat <jasquat@users.noreply.github.com>
* move data migration code out of bin so it can be reused in background processor
* sleep for 5 minutes and update bpmn js to pull in some fixes from elizabeth
* update spiff to pull in parser update to make it act like before
---------
Co-authored-by: burnettk <burnettk@users.noreply.github.com>
* use carbon components when adding secrets
* docker compose pull before running up w/ burnettk
* use all carbon on script task unit test modal w/ burnettk
* removed remaining references to bootstrap from frontend w/ burnettk
* updated usage of hidden to use the class instead of the bootstrap attribute w/ burnettk
* print out docker version for debugging
* docker pull with docker instead of compose
* added comment for docker pull in check docker start script
---------
Co-authored-by: jasquat <jasquat@users.noreply.github.com>
* This alters how the send_event endpoint responds - it originally responded with a process instance, it now responds with the next task, in keeping with how other task completion endpoints behave.
* I was forced to upgrade some of the bpmn-js libraries which fixes some of the linting errors on the front end.
* The "Return to home" button isn't always displayed. It will not display when it is redirecting, or when the current task is running.
.
Show the "next" task if no task is provided on the task_show api endpoint
Adding interstitial endpoint
Rename run to run_and_save
Remove repeated code from execution strategy
Adding interstital frontend page
* Link between message instance and correlations is now a link table and many-to-many relationships as recommended by SQLAlchemy
* Use the correlation keys, not the process id when accepting api messages.
The same issue was happening on the ProcessInstanceListTable, and there it was being managed by a "SafelySetErrorMessage" function in one case,
but would not be addressed in all possible cases.
Reworked error handling into a context provider (APIErrorProvider) and hook (UseApiError) and removed the "(useContext as any)(ErrorContext)[1];" that felt a little off but that never was an actual problem.