Add exception type to error output

[dyackzan]

Knowing the type of exception that occurred provides more context and makes launch exceptions easier to track down.

[dyackzan]

@tylerjw I just tested locally, but I'll work on getting a test added to this PR

[dyackzan]

@tylerjw I don't think a backport will work here...It looks like the functionality to report multiple exceptions was added into rolling and not backported to humble and iron (from this PR). So this change won't be 1-to-1 and the test I added for the multiple exception case wouldn't pass. What's the recommended course of action here? Attempt to backport those changes first and then these? Or should I just open separate PRs

[clalancette]

What's the recommended course of action here? Attempt to backport those changes first and then these? Or should I just open separate PRs

Generally, we get the fixes into rolling first, and then open backports. So even though this is going to have to be a "manual" backport (rather than using the bot), we should get this landed first.

[clalancette]

CI:

• Linux
• Linux-aarch64
• Windows

[dyackzan]

CI:

• Linux
• Linux-aarch64
• Windows

@clalancette in the last section of the logs it looks like the test_xmllint.py script failed. I see

14:07:00 test/test_xmllint.py .                                                   [ 38%]
14:07:03 FATAL: command execution failed

but I don't think this would be caused by this PR since I didn't commit any xml changes. Let me know what you think

[clalancette]

but I don't think this would be caused by this PR since I didn't commit any xml changes. Let me know what you think

Yeah, that was just an infrastructure problem. I've kicked off the job again.