Closed fozziethebeat closed 4 months ago
Hello! Thanks for taking the time to report this issue! I'm having some trouble reproducing the problem - a couple of questions:
I can't quite share the dataset I was using that triggered this but it uses a long list of conversation messages as input which I think is making the dropdown area too tall. Here's a screen grab of what I'm seeing.
And this is happening on Firefox v125.0.3 on MacOS Sonoma 14.4.1
https://github.com/UKGovernmentBEIS/inspect_ai/assets/398211/5a7625fb-c63d-4bf4-bc9b-241b8d6a49a6
Thanks for sharing this - I see what you're talking about! I'm taking a closer look at the log viewer this week and so will try to get to this soon!
great! other than this everything worked really smoothly!
Note that we are re-working the samples view to handle an arbitrarily large number of samples. Part of this work will resolve this issue (as there will be an entirely different expand implementation). So I'll close this issue (and you should see the fix land soon in 0.3.10 or 0.3.11).
Would it be possible to leave the issue open until the main branch has the fix commited? Otherwise anyone else who encounters this issue would be surprised to find that it's not actually fixed.
Typical practice is to only close an issue when uses/developers can access the update rather than when an update is planned to be worked on.
Yes, good call! Re-opening now.
Ah, thank you! I'll keep tabs on this and be sure to test things out once the fix is in~
Resolved with https://github.com/UKGovernmentBEIS/inspect_ai/commit/140c3d63ad42125213bffaa799147135c81857f3 (the sample view is now fullscreen)
When opening up
inspect view
and looking at the table of results for an eval, I find it difficult to expand individual rows that might have a large hidden area. For example, When I try to click row 1 the UI will expand row 2 (most likely because the hidden view for row1 covers row 2 and the click area gets confused). This gets worse as the hidden area gets longer and thus it makes it impossible to expand some random number of following rows.