nifi-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Matt Gilman (JIRA)" <>
Subject [jira] [Commented] (NIFI-339) When adding a template to the graph things can get very messy
Date Tue, 10 Feb 2015 16:17:13 GMT


Matt Gilman commented on NIFI-339:

We've discussed this in the past. Showing the total number of components, the number of components
in the root of the template (the things that will end up on the canvas), and the bounding
box of the template should help. Each of these options shouldn't be an issue.

I personally don't think we should prevent instantiation of a template (or creation of any
component for that matter) based on overlap. However, showing where the component(s) will
end up would be beneficial. 

FYI - when a template (or any component) is added to the graph the new item(s) are selected.
Based off this selection the user could group them using the Group icon in the toolbar. This
is by no means a solution to the problem, just a way work around the current behavior. Should
also keep in mind however, that we currently do not support deleting a non empty process group.

> When adding a template to the graph things can get very messy
> -------------------------------------------------------------
>                 Key: NIFI-339
>                 URL:
>             Project: Apache NiFi
>          Issue Type: Bug
>          Components: Core UI
>            Reporter: Joseph Witt
> When adding large/complex templates to the graph the system will let you unintentionally
drop a bunch of stuff on the graph.
> A great analogy provided for this was that this is like unpacking a tar bundle.  Unless
you look into the bundle first you don't know whether it will just dump all contents into
the current folder or whether it will put it all in a subfolder.
> In nifi the user cannot see the template ahead of time.  That is likely something which
should be fixed.  Further, when they place something on the graph if there will be overlap
with anything (visually) perhaps we can warn then ahead of time to verify their intent - or
simply say it isn't possible to do (because most of the time it would be really annoying).

This message was sent by Atlassian JIRA

View raw message