hmm, not clear to me where the failures can come from. Which version of sage do you use ?

ouch, are your really still using sage 9.7 ?

I not in favor of this change as implemented. Here are two reasons.

1. Currently, the LMFDB will run on Sage 9.3 (released 4 years ago). It's not uncommon at workshops for new LMFDB developers for someone to have an old version of Sage on their laptop; we're having such a workshop next month. If we can get LMFDB working on such a laptop now, I would rather not make a change that will force such a developer to upgrade their copy of Sage.
2. The Siegel section of the LMFDB is being rewritten (though I don't know how actively they're working on it; maybe @assaferan can say something). In particular, I don't think it's worth putting a lot of effort into improving the code in that part of the LMFDB.

I'm not opposed to requiring a newer version of Sage if we get substantial benefit from doing so, but I don't think removing the explicit precision cap is sufficient reason. I'll also note that the effort in #6386 to use 10.4 for CI is still worthwhile, since that's the version that we're using to run the actual webservers and we should have CI match that version.

ok, I understand. I need help in #6386.

As a side note, I think all these formulas for the dimensions of Siegel modular forms should rather be inside sage.

As a side note, I think all these formulas for the dimensions of Siegel modular forms should rather be inside sage.

I fully agree!

I don't have time today to look at #6386, but I'll try to help sometime this week.