fusion_schematic_run rejects schematic groups over 400 commands

Overview Files Skills Discover Packages Releases Dependencies Issues Breadcrumbs Pull Requests Activity

Closed bug report

Andy Engevold · 8d ago ·closed by Colby Knox

When a schematic group is large, the bridge rejects it before drawing anything: [invalid_args] commands must be a list of 1..400 schematic commands. AI Flow sends about 7 commands per part, so a group of about 80 parts comes to about 560 commands.

Expected: the bridge batches long command lists itself, or reports its limit so AI Flow can split the group.

Workaround: split the block in the netlist. Groups of 40 parts or fewer draw fine.

Versions: Fusion Bridge 1.13.59, adom-bridge 2.1.185, AI Flow 0.2.43, Hydrogen 1.0.701, Fusion on Windows 11.

2 Replies

Colby Knox marked this in progress · 8d ago

Colby Knox · 8d ago

Thanks for this report, Andy, and for retesting #136 and #135. Clear reports with the exact error and versions make these much faster to fix, and your retest is what let us close those two.

The fix for this one is written but not released yet. fusion_schematic_run will take a list of any length: the bridge checks the whole list first, so a bad command still means nothing runs, then runs it in batches of 300 and answers as one call.

I still need to run your case on Fusion: 80 parts at 7 commands each. I'll post here when it's out. Until then, splitting the block as you're doing is the right workaround.

Colby Knox · 8d ago

Fixed in fusion-bridge 1.13.64, which is public. Update the bridge to get it. Fusion doesn't need a restart.

fusion_schematic_run now takes a list of any length. The bridge checks the whole list first, so a bad command still means nothing runs, then runs it in batches of 300.

We tested your case on Fusion: 80 parts at 7 commands each, 560 commands. All 80 parts landed in about 2 minutes. A bad command placed late in the list was refused with nothing drawn.

One thing is new for a list that long: if it isn't finished after about 100 seconds, the call answers stillRunning: true with a runId, not a timeout. Poll fusion_schematic_run {runId} until stillRunning is false, and that answer is the whole reply. Shorter lists answer as before.

AI Flow 0.2.46 also splits big groups for you (adom-aiflow#133), so you shouldn't normally reach this limit.

Thanks again, Andy. Closing this one; reopen it if a long list misbehaves.

Colby Knox closed this · 8d ago