More interesting issues, and a potential fix...
Found a particularly long code block in one of the private rooms I'm in, and it actually managed to (repeatedly) crash WindowServer just by scrolling past it. Trying to figure out what was wrong, I eventually decided to try to blow away all of my preferences (following the instructions https://confluence.atlassian.com/hipchatkb/mac-desktop-app-clearing-preferences-751436224.html) and this seems to have resolved the slow scroll issues. Specifically, I only followed Steps 1 (Close Hipchat) and 7 in the above, since we use SSO and I didn't think KeyChain would be involved:
At the command prompt, enter the following commands:
defaults delete com.hipchat.HipChat
rm ~/Library/Preferences/com.hipchat.HipChat.plist
rm -rf ~/Library/HipChat/
rm -rf ~/Library/Application\ Support/hipchat/
rm -rf ~/Library/Application\ Support/com.hipchat.HipChat
rm -rf ~/Library/Caches/HipChat
rm -rf ~/Library/Caches/com.hipchat.HipChat/
rm -rf ~/Library/Saved\ Application\ State/com.hipchat.HipChat.savedState
rm -rf ~/Library/WebKit/com.hipchat.HipChat
There's still a small amount of slowdown when scrolling past code blocks but it's orders of magnitude better, and totally livable, vs. what was happening before.
Note to anyone following this issue - it'll make you log in to HipChat again, so make sure you know or can reset your password before you proceed.
We recently announced that we would be discontinuing our team messaging tools, Stride and Hipchat, and will therefore be closing out all related feature requests and open issues. The end of life date for Stride and Hipchat Cloud is February 15, 2019. Until that time, we will only issue product updates for security issues or critical bugs.
If you are an existing Hipchat Server or Data Center customer and feel like this issue has been closed in error, please contact support to let us know.
Our recommended migration path is to Slack - you can read more about this and other options at Stride and Hipchat Cloud End of Life: Just over 90 days
Thank you for your input on this ticket and for being a valued Atlassian customer.