◢███◤      ◢██◤                            ◢██◤                            
     ◢██◤       ◢██◤                            ◢██◤                             
    ◢██◤       ◢██◤                            ◢██◤                              
   ◢██◤       ◢██◤                            ◢██◤                               
  ◢██◤       ◢██◤                            ◢██◤                                
◢███◤       ◢██◤                            ◢██◤                          ◥██◣   
◥███       ◢█████◣    ◢████████◤ ◢███████◤ ◢██◤ ◢██◤                 ◢██◤   ██◣  
 ███      ◢███████◣        ◢██◤ ◢██◤ ◢██◤ ◢███████◤                 ◢██◤    ███  
 ███     ◢██◤  ◢██◤ ◢████████◤ ◢██◤      ◢█████◣                   ◢██◤     ███  
 ███    ◢██◤  ◢██◤ ◢██◤  ███◤ ◢██◤ ◢██◤ ◢██◤◥███◣                 ◢██◤      ███  
 ◥██   ◢██◤  ◢██◤ ◢████████◤ ◢███████◤ ◢██◤  ◥███◣               ◢██◤       ███◣ 
  ◥██◣                                                          ◢██◤       ◢███◤ 
                                 ◢███◤ ◢███◤ ◢██◤  ◢██◤ ◢█████████◤       ◢██◤   
                                ◢█████████◤ ◢██◤  ◢██◤ ◢██◤  ████◤       ◢██◤    
                               ◢██◤◢█◤◢██◤ ◢██◤  ◢██◤ ◢██◤   ███◤       ◢██◤     
                              ◢██◤   ◢██◤ ◢████████◤ ◢█████████◤       ◢██◤      
                             ◢██◤   ◢██◤ ◢████████◤ ◢█████████◤      ◢███◤       

0001
(Idea/bug?) sys.xfer_upgrade_from takes GC on loaded upgrades
4ad7kd
At the moment, when attempting to xfer loaded upgrades with sys.xfer_upgrade_from, it still takes the 25BGC fee. This behaviour seems unintended (some other failures such as the target sys not being breached don't take GC) and it can be a pain to lose 25BGC while getting nothing from a small mistake.
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
I think this is so that you have to actually expose upgrades to take everything not loaded; if it just failed silently you could just try to steal 0, if it failed, steal 1, if it failed, steal 2 etc, until it works; then keep stealing the same index until it fails and then increase the index again. This way you have to actually expose or pay a lot extra. Then again, i think xfer_gc_from doesn't take the xfer cost if you ask for too much GC, so IDK.
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
yea, i deffo see a possible reason for it, but it's inconstant with other trust scripts, 25BGC is too much of a penalty for something like that anyway imo
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
I'd be satisfied with 'baseline expose_upgrades' cost at that tier. So there's still a penalty, but not that much. Problem is, this was not only a design choice for what dtr mentioned but also an implementation detail to improve performance & prevent race conditions, etc. I'll definitely think through all this stuff again when I come back to this issue.
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -