The diffing logic currently tries to load the owning class of the VI being diffed into memory (by getting the default value) before rendering the diff. Presumably this is necessary for some edge cases in the underlying diffing logic, as it was present when we inherited this code.
This change ignores error 1498, which occurs when the class is broken and can't be loaded. The diffs still render at a reasonable fidelity (considering the class is broken), and it's useful to have these posted.
Why should this Pull Request be merged?
Lower-fidelity visual diffs are more useful than "this VI could not be diffed".
What does this Pull Request accomplish?
The diffing logic currently tries to load the owning class of the VI being diffed into memory (by getting the default value) before rendering the diff. Presumably this is necessary for some edge cases in the underlying diffing logic, as it was present when we inherited this code.
This change ignores error 1498, which occurs when the class is broken and can't be loaded. The diffs still render at a reasonable fidelity (considering the class is broken), and it's useful to have these posted.
Why should this Pull Request be merged?
Lower-fidelity visual diffs are more useful than "this VI could not be diffed".
Fixes #44.
What testing has been done?
See the diff bot posts on this PR: https://github.com/ni/niveristand-xnet-system-tests/pull/11
The first is built against master, the second is built against the branch backing this PR.