In chapter 2 of lishogi.org/study/lHxDLiCB/vmJjW4aC,
* I cannot promote the variation to the main line (reverts moments later)
* Engine analysis at the end of the variation is nonsense (illegal moves)
Hope this is the right place to post this bug!
Whoa the engine analysis bugging out is super weird. Never seen that before.
I will look into it, looks weird. Just to make sure - you wanted to promote the variation starting "73. +Px56" to mainline? I think there is an issue with handicap offset - for some reason the embed preview here starts with first move marked as number 2, but the studies start with move as 1...
Okay looked into it - studies from games have mainline fixed to match the game they came from. So promoting to mainline shouldn't work, therefore clicking it should correctly do nothing, but I will hide the option to make it clear.
But thanks to your report I found there is a bug sometimes when promoting variations inside variations with fixed game mainlines, which shifts the entire variation tree one move higher (look at the study the variation is listed under move 72, but it should be listed under move 73). This causes that the engine doesn't receive the N*7f, so the engine output is just nonsense, because the moves make no sense at that point. I will try to get a fix for the next update. For now, if you want to get correct engine analysis, you need to delete the variation and do it again - avoiding promoting subvariations.
You can't post in the forums yet. Play some games!