[saffron] less punishing ci

[martyall]

Simple fix for this https://github.com/o1-labs/proof-systems/actions/runs/13273640255/job/37058646902?pr=2969

The problem comes from the time spent computing commitments for test iterations generating "large" user data. Using the DataSize::Medium and reducing to 10 test cases still gives good coverage. Medium allows for larger than 1 chunk which is the most important thing here. On my machine this went from 1min to 11s, CI went back to the normal 60min range.

While working on this I realized that while correct, the test was written awkwardly. I reordered the commitment checks to more closely match the lifecycle of the update

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 76.89%. Comparing base (dca3c7f) to head (1569ce6).
Report is 4 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #3023      +/-   ##
==========================================
+ Coverage   76.87%   76.89%   +0.01%     
==========================================
  Files         262      262              
  Lines       62114    62118       +4     
==========================================
+ Hits        47752    47763      +11     
+ Misses      14362    14355       -7     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[dannywillems]

Ok for the changes in the user input. I hope it unblocks the CI of other people.